GUEST EDITORIAL Congenital disease

Pediatric Anesthesia is the only anesthesia journal ded- who developed hypoglycemia were infants. (9). Steven icated exclusively to perioperative issues in children and Nicolson take the opposite approach of ‘first do undergoing procedures under anesthesia and sedation. no harm’ (10). If we do not want ‘tight glycemic con- It is a privilege to be the guest editor of this special trol’ because of concern about hypoglycemic brain issue dedicated to the care of children with heart dis- injury, when should we start treating blood sugars? ease. The target audience is anesthetists who care for There are no clear answers based on neurological out- children with heart disease both during cardiac and comes in children. non-cardiac procedures. The latter takes on increasing Williams and Cohen (11) discuss the care of low importance as children with heart disease undergoing birth weight (LBW) infants and their outcomes. Pre- non-cardiac procedures appear to be at a higher risk maturity and LBW are independent risk factors for for cardiac arrest under anesthesia than those without adverse outcomes after . Do the anes- heart disease (1). We hope the articles in this special thetics we use add to this insult? If prolonged exposure issue will provide guidelines for management and to volatile anesthetics is bad for the developing neona- spark discussions leading to the production of new tal brain, would avoiding them make for improved guidelines. outcomes? Wise-Faberowski and Loepke (12) review Over a decade ago Austin et al. (2) demonstrated the current research in search of a clear answer and the benefits of neurological monitoring during heart conclude that there isn’t one. So we muddle along with surgery in reducing adverse neurological outcomes in a mix of anesthetics not sure whether we are helping children. It is not yet standard of care and the current improve outcomes or not. level of evidence supporting its routine use is level 2 For the purist, Wong and Morton (13) present ideas (non-randomized controlled trials). However Kasman on total intravenous anesthetics and target controlled and Brady (3) wonder if the burden of proof we infusions (TCI). Since ‘early extubation’ and early dis- impose on the cerebral oximeter to improve neurologi- charge from the intensive care unit and length of stay cal outcomes is reasonable. The article on transesopha- are closely tracked by insurance payors, TCI has consid- geal (TEE) by Kamra et al. (4), erable appeal. Once the patient is in intensive care, man- coupled with the video clips available online, provides aging their pain and sedation becomes important (14). a comprehensive overview of TEE in children. Moni- Children with heart failure are a growing segment of toring and treating coagulation disorders after heart patients for whom pediatric anesthesiologists must surgery (5) and the risks of transfusion are reviewed care. These children are often subjected to multiple and provide guidelines for management provided (6). diagnostic tests and interventions for which anesthesia With improved survival after heart surgery, the is required. Rosenthal and Hammer (15) discuss the number of children with cardiac rhythm management current etiologies and pathogenesis of heart failure and devices (CRMD) has substantially increased. Navarat- offer ideas for safe perioperative care. Once vasodila- nam and Dubin (7) present a practical approach on tor and other medical therapy fails, these children are how to care for patients with these devices in the peri- then placed on mechanical assist devices (16). Finally, operative period. The information on CRMD, along some of these children undergo with White’s (8) practical approach to children with and their perioperative management and outcomes are heart disease for non-cardiac surgery, provides a addressed (17). framework for managing these patients. Traditionally, mortality has been the outcome mea- Of the many controversies in the perioperative man- sure of success or failure of a surgical program. Thia- agement of children during heart surgery, glucose con- garajan and Laussen (18) offer a thoughtful expose´on trol or lack thereof, is high on the list. A recent why low mortality should no longer be the yardstick placebo controlled, large, prospective study of patients with which to measure success. Other quality indica- in a pediatric intensive care unit reported that ‘tight tors are urgently needed if we are to improve the peri- glycemic control’ resulted in shorter intensive care unit operative care of children with heart disease after length of stay, reduced lactate levels, decreased C-reac- cardiac surgery. tive protein and reduced incidence of infections. How- Doherty and Holtby (19) address the issue of the ever the incidence of hypoglycemia (blood glucose unavailability of pediatric subspecialists for the care of <40 mg)1 dl) was higher in the intensive insulin treat- children with heart disease in developing countries. ment group (25%) and the majority (80%) of patients Finally, as adults with congenital heart disease

Pediatric Anesthesia 21 (2011) 471–472 ª 2011 Blackwell Publishing Ltd 471 Editorial

(ACHD) now outnumber the children with heart dis- ease, where should the adults be cared for (20)? Will Chandra Ramamoorthy Department of Anesthesiology, care givers be from the adult world or will they be Stanford University Medical Center, pediatric subspecialists? Seal (21) reviews current data Stanford, CA, and discusses their care model in Alberta, Canada, USA Guest Editor where pediatric cardiac anesthesiologists also manage ACHD. Are we pediatric anesthesiologists or cardiac doi:10.1111/j.1460-9592.2011.03592.x anesthesiologists specializing in congenital heart dis- ease? Time to ponder.

References 1 Ramamoorthy C, Haberkern CM, erative care. Pediatr Anesth 2011; 21: 512– 15 Rosenthal DN, Hammer GB. Cardiomyopa- Bhananker S et al. Anesthesia-related car- 521. thy and heart failure in children: anesthetic diac arrest in children with heart disease: 8 White MC. Approach to managing children implications. Pediatr Anesth 2011; 21: 577– data from the Pediatric Perioperative Car- with heart disease for noncardiac surgery. 584. diac Arrest (POCA) registry. Anesth Analg Pediatr Anesth 2011; 21: 522–529. 16 Mossad EB, Motta P, Rossano J et al. Peri- 2010; 110: 1376–1382. 9 Vlasselaers D, Milants I, Desmets L. et al. operative management of pediatric patients 2 Austin EH, Edmonds HL, Auden SM et al. Intensive insulin therapy for patients in on mechanical cardiac support. Pediatr Benefit of neurolophysiologic monitoring pediatric intensive care: a prospective ran- Anesth 2011; 21: 585–593. for pediatric cardiac surgery. J Thorac Car- domized controlled study. Lancet 2009; 373: 17 Schure AY, Kussman BD. Pediatric heart diovasc Surg 1997; 114: 707–715. 717. 547–556. transplantation: demographics, outcomes, 3 Kasman N, Brady K. Cerebral oximetry for 10 Steven J, Nicholson S. Perioperative man- and anesthetic implications. Pediatr Anesth pediatric anesthesia: why do intelligent clini- agement of blood glucose during open heart 2011; 21: 594–603. cians disagree? Pediatr Anesth 2011; 21: surgery in infants and children. Pediatr 18 Thiagarajan RR, Laussen PC. Mortality as 473–478. Anesth 2011; 21: 530–537. an outcome measure following cardiac sur- 4 Kamra K, Russell I, Miller-Hance WC. 11 Williams GD, Cohen RS. Perioperative gery for congenital heart disease in the Role of transesophageal echocardiography management of low birth weight infants for current era. Pediatr Anesth 2011; 21: 604– in the management of pediatric patients open-heart surgery. Pediatr Anesth 2011; 21: 608. with congenital heart disease. Pediatr Anesth 538–553. 19 Doherty C, Holtby H. Pediatric cardiac 2011; 21: 479–493. 12 Wise-Faberowski L, Loepke A. Anesthesia anesthesia in the developing world. Pediatr 5 Arnold P. Treatment and monitoring of during surgical repair for congenital heart Anesth 2011; 21: 609–614. coagulation abnormalities in children under- disease and the developing brain: neurotoxic 20 Marelli A J., Therrien J, Mackie A S. et al. going heart surgery. Pediatr Anesth 2011; or neuroprotective? Pediatr Anesth 2011; 21: Planning the specialized care of adult con- 21: 494–503. 554–559. genital heart disease patients: from numbers 6 Guzzetta NA. Benefits and risks of red 13 Wong GLS, Morton NS. Total intravenous to guidelines; an epidemiologic approach. blood cell transfusion in pediatric patients anesthesia (TIVA) in pediatric cardiac anes- Am Heart J 2009; 157: 1–8. undergoing cardiac surgery Pediatr Anesth thesia. Pediatr Anesth 2011; 21: 560–566. 21 Seal R. Adult congenital heart disease. Pedi- 2011; 21: 504–511. 14 Wolf AR, Jackman L. Analgesia and seda- atr Anesth 2011; 21: 615–622. 7 Navaratnam M, Dubin A. Pediatric pace- tion after pediatric cardiac surgery. Pediatr makers and ICDs: how to optimize periop- Anesth 2011; 21: 567–576.

472 Pediatric Anesthesia 21 (2011) 471–472 ª 2011 Blackwell Publishing Ltd Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Adult congenital heart disease Robert Seal

Department of Anesthesia and Pain Medicine, University of Alberta and Stollery Children’s Hospital, Edmonton AB Canada

Keywords Summary congenital heart disease; general anesthesia; cardiac surgery For a decade now, it has been recognized that optimal management of adult congenital heart disease (ACHD) requires a skilled multidisciplinary Correspondence team. The size and complexity of the population of adults with congenital Robert Seal, heart disease (CHD) are increasing. This article reviews the general consid- Department of Anesthesia and Pain erations for giving an anesthetic to an adult with CHD for cardiac or non- Medicine, University of Alberta and Stollery cardiac surgery and provides further elaboration for a variety of complex Children’s Hospital, 1C1.04 WMC, 8440 112 Street, Edmonton AB T6G 2B7, Canada patient types. Lastly, the advantages of an organized multidisciplinary Email: [email protected] approach to patients with ACHD are discussed.

Accepted 17 March 2011 doi:10.1111/j.1460-9592.2011.03591.x

living adults (9%) have complex CHD (9). When Introduction ACHD is stratified as simple, moderate, or complex, In October 2000, the American College of Cardiology there are roughly the same numbers of adults in the hosted the 32nd Bethesda Conference on ‘Care of the simple group as in the other two groups combined, Adult with Congenital Heart Disease’. Recommenda- and all three groups combined are a smaller popula- tions were generated pertaining to the organization of tion of patients than those having a bicuspid aortic health care delivery for adult congenital heart disease valve (12). The untreated survival of approximately (ACHD) including the required health human 40% in both the simple and moderate groups improves resources and the educational needs of healthcare pro- to 75–80% with treatment. The adult prevalence of viders, access to and continuity of care, as well as severe CHD is increasing as improvements in surgical meeting the special needs of this population (1–5). Spe- and medical management result in more than 90% of cialized ACHD regional centers with multidisciplinary the children with complex CHD surviving into adult- expertise serving catchment populations of 5–10 mil- hood compared with the only 15–25% survival for lion each were proposed. Further, it was advised that untreated patients, which was seen five decades ago these centers should provide a smooth transition from (9,13). In the ACHD population, nearly 50% of the care previously provided to these patients by pedi- patients require ongoing follow-up and further proce- atric hospitals. Almost a decade later, concerns have dures. This matches a predicted need for follow-up in been raised about the incomplete realization of these ACHD of >2 per 1000 live births. Following a previ- recommendations (6–8). ous corrective or palliative surgical or interventional procedures, half of patients will have two or more surgeries and one-quarter will have Prevalence and changing profile three or more surgeries in their lifetime. Congenital heart disease (CHD) in children has a pre- valence of between 5 and 10 per 1000 live births of Transition from pediatric to adult medical care which approximately 1.5 per 1000 have complex CHD (1,9–11). In adults, the prevalence of CHD is approxi- Although up to 10% of first presentations may occur mately 4 per 1000 living adults of which 0.38 per 1000 in adulthood, most patients with ACHD begin their

Pediatric Anesthesia 21 (2011) 615–622 ª 2011 Blackwell Publishing Ltd 615 Adult congenital heart disease R. Seal cardiac medical care as infants or children. Affiliation (i.e., diabetes) and/or complications of ACHD such as of pediatric cardiology programs with an adult ACHD erythrocytosis, developmental and central nervous sys- program facilitates the patient’s transition from the tem disorders (seizures, previous ischemic/embolic pediatric to the adult program and ensures continuity events, and intracerebral abscesses), chronic lung dis- of patient care (14,15). This can be especially challeng- ease, cholelithiasis, renal dysfunction, and nephrolithi- ing in the subset of patients with ACHD who have asis (25–28). multiple physical and/or mental disabilities. It is not uncommon for adults with CHD to be living with their Investigations parents or to have complex psychosocial needs. Conti- nuity of care may also be compromised by noncompli- In addition to assessing the anatomy, investigations ance in the presence of minimal or no symptoms or are helpful in the evaluation of the patient’s functional with the denial phase, which is common during adoles- class and ventricular performance. Patients may cence. Finally, patients may be lost to follow-up when undergo exercise testing, Holter monitoring, and/or they move for education or work-related reasons. electrophysiologic studies. Both transthoracic and/or Other factors such as health care infrastructure and transesophageal echocardiography (TEE) and cardiac economics in any given country can influence the magnetic resonance imaging (MRI) are relatively non- ACHD patient’s access to care. In the United States, invasive and are well suited to serial follow-up assess- of an approximate burden of a half million adults with ments. MRI may be the preferred technique for the complex CHD, it has been estimated that only 30 000 assessing right ventricular (RV) volumes and the receive follow-up care in a ACHD center (6). degree of pulmonary regurgitation (PR) as well as for viewing both the right ventricular outflow tract and the proximal pulmonary arteries (23,29). In coarctation Anesthetic considerations of the , MRI provides optimal visualization of More detailed information pertaining to the anesthetic the entire aorta to enable the identification of residual considerations in ACHD is available in several review or recurrent coarctation as well as the formation of articles (16–18). Detailed information pertaining to the aneurysms of the aortic root or the site of pervious overall assessment and ongoing management of the surgical repair (13). MRI research tools such as the patient with ACHD may be found on the Internet site assessment of ventricular mass, vector change, and of the Canadian Adult Congenital Heart Network, contractile geometry are likely to find clinical utility in http://www.cachnet.org/index.cgi, as well as in Ameri- ACHD (23). can, British, Canadian, and European Guidelines Cardiac catheterization is helpful: (i) in the investi- (15,19–23). The paragraphs that follow explore a num- gation into pulmonary vascular disease and the deter- ber of important general considerations for giving an mination of its responsiveness to oxygen and anesthetic to an adult with CHD for cardiac or non- vasodilators, (ii) in the assessment of functional suit- cardiac surgery. ability for Fontan completion, (iii) as a supplement to echocardiography and MRI in the delineation of com- plex congenital lesions, (iv) in the assessment of the Preoperative evaluation coronary arteries, and (v) for interventional Obtaining an accurate history can be problematic as procedures. only one-half to three-quarters of adults with CHD can correctly name or describe their diagnosis (24). By Arrhythmias limiting their exercise, patients with ACHD may be relatively asymptomatic or have grown accustomed to In ACHD, the congenital anomalies themselves as well their condition. Assessment should look for the evi- as changes attributable to chamber dilation, fibrosis, dence of long-term cardiac complications and risk fac- surgical incisions, and hemodynamic deterioration tors such as pulmonary hypertension (PHT; threefold contribute to the development of atrial and ventricular risk increment), cyanosis (higher mortality and post- arrhythmias (23,30). The occurrence of these arrhyth- operative complications), ventricular dysfunction, con- mias is a poor prognostic sign. Even with a simple duction defects and arrhythmias, residual shunts, atrial atrial septal defect (ASD), there is an increased regurgitant or stenotic valves, and aneurysms (2). As risk of the late onset of atrial tachyarrhythmias and poor general health is also a risk factor, it is important atrial fibrillation with closure after the age of 40 (21). to evaluate chronic cardiac (i.e., hypertension and cor- Sinus node dysfunction, poor ventricular function, and onary artery disease) or noncardiac coexisting diseases the desire for pregnancy may complicate pharmacologic

616 Pediatric Anesthesia 21 (2011) 615–622 ª 2011 Blackwell Publishing Ltd R. Seal Adult congenital heart disease management of arrhythmias. Thus, patients with distal arterial pressures and waveforms. In reoperative ACHD often present for procedures such as catheter open heart surgery, a femoral artery may be unavailable ablation, open ablation (Cox Maze procedure), as well for monitoring as the surgeons may require it for femo- as the insertion of automated implantable cardioverter ral access. In addition to inter- defibrillators and anti-tachycardia pacemakers. Epicar- mittent arterial blood gas analysis, intra-arterial dial lead placement may be required for AICDs and monitoring facilitates a continuous assessment of hemo- pacemakers as transvenous placement may be compli- dynamics and the alterations that result from changes cated by difficult venous access, unacceptable thrombo- in anesthetic depth and/or intravascular volume status. sis risk, or the presence of intracardiac shunts. A previous Fontan repair precludes central venous access to the systemic and . Two- dimensional and most recently three-dimensional Anesthetic plan ultrasound visualization is valuable for preinsertion The optimal anesthetic plan relies upon the anesthesi- evaluation of central veins and guidance during inser- ologist’s ability to correctly understand the patient’s tion (31). Pulmonary arterial are rarely used pathophysiology. Premedication should be selected because of (i) difficult or impossible placement, (ii) risk carefully, and its use particularly with opioids may be of arrhythmias, and (iii) measurements confounded by risky in patients in who further hypoventilation and the presence of shunts or valve regurgitation. With hypoxemia could lead to decompensation. In patients open heart surgery, it is common to place intrathoracic with erythrocytosis, it is advisable to maintain preop- catheters into otherwise inaccessible atria (left atrial or erative hydration (minimizing preoperative clear fluid systemic atrial lines). Intraoperative TEE (including fasting interval or administering preoperative intrave- three-dimensional TEE) can provide a continuous nous fluids). Intravenous fluid administration requires assessment of ventricular function, preload, and vigilance to avoid systemic embolization of air bub- changes in valve function. In addition to its ability to bles. Previous cardiac and noncardiac procedures and guide and subsequently evaluate the surgical repair, hospital admissions may limit the availability of cen- TEE has shown utility by uncovering and assessing tral and peripheral vascular access sites. Because of unexpected presurgical findings in the OR. Even rou- the variety of CHD lesions and the variations within a tinely used monitors may have additional benefits, given type of CHD, there is no ‘one-size-fits-all’ selec- such as the ability of pulse oximetry to track changes tion of anesthetic medications. Rather, the choices in pulmonary blood flow and R fi L shunting in must be tailored to unique features of each case. patients with cyanotic lesions, or limitations, such as

Although of considerable utility in pediatric CHD, it underestimation of PaCO2 by end-tidal CO2 in patients is important to remember that ketamine may depress with R fi L shunts (32). ventricular function in patients with maximal sympa- thetic activation. Antibiotic prophylaxis In contrast to the more widespread use of antibiotic Monitoring prophylaxis with many dental and surgical procedures Selection of monitoring must take into account the in the past, current American Heart Association rec- patient’s cardiac defect and overall health status, the ommendations limit use to a much narrower spectrum planned surgical or cardiac catheterization procedures, of procedures in patients with previous endocarditis, and consideration of the pros and cons of invasive unrepaired cyanotic CHD including those with pallia- monitoring. Combining this information with a preop- tive shunts and conduits, the presence of prosthetic erative examination of the patient’s pulses and four material, or device during the first 6 months after the limb blood pressure measurements helps guide the procedure, and residual defects at the site or adjacent placement of both noninvasive and invasive monitors. to the site of a prosthetic patch or device (33). In car- Left upper limb blood pressures are rendered unreli- diac surgery, anti-staphylococcal (typically cefazolin) able by previous subclavian flap repair of coarctation of surgical wound antibiotic prophylaxis remains stan- the aorta while a Blalock Taussig shunt renders the dard practice. invasive and noninvasive blood pressure diminished or unobtainable in the corresponding upper limb. Previous Reoperation arterial cut down or percutaneous cannulation may result in stenosis, collateral vessels, or scar tissue forma- With reoperation, reopening the sternum can be dan- tion that may impair successful access as well as reliable gerous because of the possibility of adhesions between

Pediatric Anesthesia 21 (2011) 615–622 ª 2011 Blackwell Publishing Ltd 617 Adult congenital heart disease R. Seal the heart, great vessels (or conduit), and the sternum microvascular perfusion as a result of erythrocytosis and the lack of a retrosternal space. Further, the adhe- and hyperviscosity. Chronic hyperventilation and sions obscure anatomic landmarks including the route increased hematocrit improve oxygen transport in cya- of the coronary arteries. Collateral vessels may compli- nosis. Unfortunately, the increased mass and stiffness cate dissection and increase the risk of hemorrhage. of the red blood cells (RBC) result in elevated viscosity Previous cardiac surgeries may also have contributed that further increases cardiac work in already to long-term myocardial damage (1). compromised by chronic hypoxia-induced systolic and diastolic dysfunction. Erythrocytosis in the presence of iron deficiency results in poorly deformable RBCs that Considerations for patients with cyanotic lesions raise the risk of thrombosis (34). In patients with cyanotic lesions, anesthesia manage- Coagulation anomalies found in patients with cya- ment is aimed at promoting pulmonary blood flow notic lesions include abnormal platelet function, dimin- through maintaining a favorable pulmonary vascular ished platelet survival because of peripheral resistance (PVR), ensuring optimal driving pressure consumption, reduced levels of vitamin K-dependent for pulmonary perfusion (systemic arterial or venous coagulation factors as well as factor V and von Wille- pressure depending upon the physiology), decreasing brand factor (34). Erythrocytosis not only elevates R fi L shunting through maintaining normovolemia INR and PTT but may interfere with the standard lab- and systemic vascular resistance), optimizing hemato- oratory tests used to measure these indices. crit, and modulating dynamic right ventricular out- flow tract obstruction (RVOTO), if present, with Considerations for patients with pulmonary b-blockers. Altering the inspired concentration of oxy- hypertension gen (FiO2) will have little effect on peripheral oxygen saturation (SpO2) while systemic vasoconstriction In PHT, PVR increases as a result of changes in the (phenylephrine, norepinephrine) will increase SpO2 pulmonary vascular bed as a consequence of being either by decreasing R fi L shunting or by improving subjected to increased pulmonary blood flow and near the flow through an aortopulmonary shunt or aorto- systemic blood pressure. Initially, reactive and revers- pulmonary collaterals. The arterial oxygen saturation ible, these changes eventually become fixed and perma- is frequently improved under anesthesia in patients nent. Pathological examination reveals media with cyanotic lesions as a consequence of diminished hypertrophy in small muscular arteries and arterioles,

O2 consumption and the resultant elevation of the cellular proliferation in the intima, migration of mixed venous oxygen saturation. Patients with cya- smooth muscle cells into the subendothelium, and ulti- notic lesions have a blunted hypoxic ventilatory mately progressive fibrosis leading to obliteration of response while the response to hypercapnia is pre- these vessels. The reactive nature of PHT is character- served (32). ized by exaggerated responses to sympathetic stimuli As pulmonary blood flow is poor, the rate of rise or (i.e., pain or surgical stress response), acidosis, hyper- fall of the partial pressure of inhalational anesthetic carbia, hypoxia, and hypothermia. Although it may be agents in the blood and brain is slow. Thus, inhala- easier and preferable to prevent exacerbations of PHT tional induction and recovery are slower. Importantly, than it is to treat them, the strategies used are similar. this also leads to difficulty in reducing inhaled anes- It is advisable to use a high FiO2 accompanied by thetic agent concentration in the event of an adverse hyperventilation obtained with the minimal possible event such as hypotension. Intravenous medications intrathoracic pressure (goal PaCO2 of 25–30 mmHg have a theoretically, but perhaps not clinically signifi- and pH >7.45). One must aim for an anesthetic depth cant, more rapid onset as a consequence of R fi L that will blunt the sympathetic response as best as pos- shunting. Conversely, the onset of intravenous medica- sible. The choice of anesthetic medication used to tions may be slowed in the presence of reduced cardiac achieve this goal will depend upon the particular output. Anesthetic-induced hypotension may be more circumstances, but opioids (sufentanil, fentanyl, and profound in patients on diuretics and angiotensin-con- remifentanil) are often helpful. Inotropic support with verting enzyme inhibitors. b-adrenergic agents (dobutamine or isoproterenol) or Patients with cyanotic lesions are at additional risk phosphodiesterase-3 inhibitors (milrinone) may permit of ventricular dysfunction and myocardial ischemia, as the patient to tolerate this depth of anesthesia and help consequences of impaired oxygen transport, decreased the RV to cope. However, one must beware of the sys- diastolic perfusion pressure caused by shunts or collat- temic hypotension that may accompany their use and erals, and from impaired, or occluded myocardial aggravate R fi L shunting. Systemic vasoconstriction

618 Pediatric Anesthesia 21 (2011) 615–622 ª 2011 Blackwell Publishing Ltd R. Seal Adult congenital heart disease with norepinephrine or phenylephrine may be essential Additional considerations in tetralogy of fallot to increase coronary perfusion pressure, relieve suben- survivors docardial ischemia and improve ventricular function. Selective pulmonary vasodilation with inhaled nitric Infant repair of tetralogy of Fallot (TOF) illustrates oxide is a mainstay of current therapy, although aero- the evolution in the management of a more common solized prostanoids have also been used with success. lesion. In comparison with initial Blalock Taussig Oral sildenafil, cGMP-specific phosphodiesterase type shunt palliation followed by late repair, infant TOF 5 inhibitor, has been used for both prophylactic and repair resolves cyanosis, circumvents dysfunctional RV maintenance therapy. The use of other vasodilating remodeling, facilitates growth of the pulmonary vascu- medications has been limited by their systemic vasodi- lature, and may decrease the incidence of aortic root lating properties. dilatation (37,38). Unfortunately, the need for transan- nular patching to relieve RVOTO produces chronic PR, which in turn leads to progressive RV enlargement Additional considerations for patients with and dysfunction, potential secondary tricuspid regurgi- Eisenmenger’s Syndrome tation, worsening exercise intolerance, heart failure, Patients with long-standing untreated L fi R intracar- supraventricular and ventricular tachyarrhythmias, and diac or extracardiac shunts may progress to Eisenm- sudden death (39–41). A mean QRS duration in excess enger’s syndrome with severe, fixed PHT with of 180 ms is associated with an increased risk of sud- equalization of the RV and left ventricular (LV) pres- den cardiac death (42). Patients may remain unaware sures, bidirectional or even reversed (R fi L) shunting of their symptoms until the onset of advanced ventric- and eventual dilation, and failure of the RV accompa- ular dysfunction at which time concomitant LV dys- nied by tricuspid regurgitation. The physiological tres- function becomes common (29). The timing of pass of anesthesia and surgery can easily upset the pulmonary was often delayed precarious balance of systemic, pulmonary, and myo- because of the commitment for subsequent replace- cardial perfusions and lead to cardiovascular collapse ment of the bioprosthetic valve every 10 years. Unfor- in patients with Eisenmenger’s syndrome. Avoidable tunately, delayed surgical correction of PR is noncardiac surgery should not be undertaken and associated with a lack of improvement in RV function referral to the expertise, and resources of a specialist (43). It is hoped that repairs that preserve pulmonary ACHD center should be strongly considered. In addi- valve function when possible, as well as undertaking tion to sildenafil, these patients are often managed early pulmonary valve replacement, may further with Bosentan, a dual endothelin receptor A and B improve TOF outcome (43,44). Anti-arrhythmia proce- antagonist. The continuous infusion of prostacyclin dures at the time of pulmonary valve replacement analogs is also utilized, but carries thrombotic, show promise (45). embolic, and infectious risks as well as the potential for rebound PHT. As Eisenmenger’s syndrome carries Additional considerations for Fontan patients a high mortality (40% prior to age 25 with a median survival in the mid-4th decade) and a markedly Survival following Fontan has improved to over 85% increased to prohibitive anesthetic risk, it is hoped at 10 years and 80% at 15–20 years (46) (15-O) that this condition will become increasingly rare with Arrhythmias, ventricular failure, and thromboembolic time (35). events are the most common etiologies of late death in Fontan patients. Tachycardia and loss of sinus rhythm are poorly tolerated in single-ventricle physiology and Considerations for patients with impaired ven- should be rectified promptly with due concern for the tricular function presence of thrombus. Sinus node dysfunction may Impaired ventricular function is common in ACHD lead to permanent pacing in up to 40% of Fontan (36). In the absence of surgical correction or as patients. a consequence of residual lesions, ventricular remod- Pulmonary thromboembolism may result from eling such as concentric hypertrophy and spherical sluggish flow in the Fontan pathway, the presence of dilation worsens over time. In general, pressure over- intravascular prosthetic material, coagulation abnor- load is less well tolerated than volume overload. malities, and arrhythmias. Systemic thromboembolism This is common in lesions in which a morphologi- may occur through a Fontan fenestration or residual cally right ventricle must function as the systemic ASD, or may originate from a ligated pulmonary ventricle. artery stump, or systemic heart chamber. Oxygen

Pediatric Anesthesia 21 (2011) 615–622 ª 2011 Blackwell Publishing Ltd 619 Adult congenital heart disease R. Seal desaturations below 94% may be indicative of R fi L thesia may be beneficial. In contrast, in pregnant shunting through collaterals or arteriovenous malfor- ACHD patients with palliative L fi R shunts general mations, residual ASDs, or Fontan fenestrations. Pro- anesthesia is often recommended as it is essential to gressive ventricular dysfunction with or without AV avoid dropping the SVR. Risks of pregnancy in Fon- valve regurgitation is common. tan patients include systemic venous congestion and Fontan patients who develop protein-losing enterop- ascites, worsening ventricular function, aggravation of athy (PLE) have a 50% 5-year survival. Features of AV valve regurgitation, atrial arrhythmias, thrombo- PLE include peripheral edema, ascites, pleural and emboli, paradoxical emboli, spontaneous abortion, pericardial effusions, chronic diarrhea, and an alpha-1 intrauterine growth retardation, and preterm birth. antitrypsin level. PLE responds poorly to medial ther- Neonatal complications occur in 20% of pregnancies apy with diuretics and aldosterone antagonists, but with prematurity and intrauterine growth retardation may occasionally respond to the lowering of systemic being most common. The incidence of CHD among venous pressure and improved cardiac output permit- the newborns was 7%. Predictors of neonatal compli- ted by conversion to a fenestrated Fontan or to revi- cations included the following: cyanosis, NYHA class sion surgery in the presence of Fontan obstruction. >II, maternal left heart obstruction, smoking, multiple Chronic systemic venous hypertension may lead to gestation, and anticoagulants use during pregnancy. hepatic dysfunction and eventual cirrhosis. A model of care for complex ACHD Considerations in pregnancy Following the Bethesda Conference, there have been Contraceptive counseling is essential in ACHD. Thir- additional calls for consolidation of the care of com- teen percent of 599 pregnancies in a series of 562 plex ACHD into regional centers that serve a catch- patients with ACHD resulted in pulmonary edema, ment population of approximately 5–10 million arrhythmia, stroke, or cardiac death (47). Management persons (6,7,20,23). This concentration of services can of pregnancy in patients with moderate to severe create the critical mass required to develop a team that ACHD requires a multidisciplinary team. The normal also includes cardiologists, surgeons, anesthesiologists, physiological changes of pregnancy [50% increase in intensive care physicians, radiologists, nurses, perfu- blood volume; 50%, 30–40% increase in cardiac out- sionists, social workers, and other health care workers put; 30% increase in heart rate and decreased systemic with expertise in the care and management of ACHD. vascular resistatnce (SVR)] may lead to pulmonary This expertise ensures the provision of (i) comprehen- overload in patients with L fi R shunts and to sive diagnostic services (including echocardiography, increased shunting and cyanosis in R fi L shunts. cardiac catheterization, electrophysiological testing, Maternal adaptation to the postpartum physiological cardiac MRI and CT, cardiac and pulmonary exercise changes also carries a particularly high mortality risk. testing), (ii) electrophysiological interventions (ablative, Risk factors for elevated maternal morbidity and mor- pacing, and implanted defibrillator therapies), (iii) inter- tality include the following: cyanosis, PHT, poor func- ventional cardiac catheterization procedures, (iv) criti- tional status and/or heart failure (baseline NYHA cal care (including extracorporeal life support, class >II or EF < 40%), poor arrhythmia control, ventricular assist device and artificial heart programs), LV obstructive lesions (mitral valve area <2 cm2, aor- (v) a transplantation program, and (vi) clinical and tic valve area <1.5 cm2, or peak left ventricular out- basic science research. Patients with bicuspid aortic flow tract gradient >30 mmHg by echocardiography), valves, small aterioventricular septal defects, repaired and prior cerebral ischemic events. Maternal mortality ASDs, and mild pulmonic stenosis may be readily in the presence of PHT (i.e., Eisenmenger’s syndrome) cared for in peripheral hospitals. The complex ACHD is estimated at between 30% and 70%. The normal center needs readily available good communications to physiological adaptation to pregnancy is altered with and from these hospitals and their physicians in its coarctation of the aorta with an increased heart rate catchment region. This can be enhanced by multidisci- being necessary to support most of the increase in car- plinary videoconferences that facilitate regularly sched- diac output. Complications (ventricular failure, dys- uled case referral and quality improvement discussions. rhythmias, endocarditis, aortic dissection, and Ideally, such a system requires the support of those thromboses) may arise in almost one-third of pregnant funding the health care system, as well as national spe- ACHD patients with coarctation of the aorta. cialty societies and patient interest groups. In pregnant ACHD patients with L fi R shunts, the As the evolution in ACHD care continues, will ideal lower SVR with spinal or epidural analgesia or anes- anesthetic care of these adults be in the domain and

620 Pediatric Anesthesia 21 (2011) 615–622 ª 2011 Blackwell Publishing Ltd R. Seal Adult congenital heart disease comfort zone of pediatric cardiac anesthesiologists or ACHD center (4,48). To meet the future demands of in the practice realm of adult cardiac anesthesiologists? patients with ACHD, it is possible that there will be a It has been recommended that patients with moderate need to follow the examples of other specialties and or complex ACHD (especially those with significant institute formalized training of those who provide risk factors such as poor functional class, PHT, cyano- anesthetic care for both their cardiac and noncardiac sis or heart failure) undergoing planned complex non- procedures. cardiac surgery should be managed at the dedicated

References 1 Warnes CA, Liberthson R, Danielson GK 14 Saidi A, Kovacs AH. Developing a transi- genital heart disease. J Am Coll Cardiol et al. Task force 1: the changing profile of tion program from pediatric to adult- 1996; 28: 768–772. congenital heart disease in adult life. JAm focused cardiology care: practical consider- 26 Perloff JK, Marelli AJ, Miner PD. Risk of Coll Cardiol 2001; 37: 1170–1175. ations. Congenit Heart Dis 2009; 4: 204–215. stroke in adults with cyanotic congenital 2 Foster E, Graham TP Jr, Driscoll DJ et al. 15 Silversides CK, Kiess M, Beauchesne L et heart disease. Circulation 1993; 87: 1954– Task force 2: special health care needs of al. CCS 2009 Consensus Conference on the 1959. adults with congenital heart disease. JAm Management of Adults with Congenital 27 Flanagan MF, Hourihan M, Keane JF. Coll Cardiol 2001; 37: 1176–1183. Heart Disease: outflow tract obstruction, Incidence of renal dysfunction in adults with 3 Child JS, Collins-Nakai RL, Alpert JS et al. coarctation of the aorta, tetralogy of Fallot, cyanotic congenital heart disease. Am J Task force 3: workforce description and Ebstein anomaly and Marfan’s syndrome. Cardiol 1991; 68: 403–406. educational requirements for the care of Can J Cardiol 2010; 26: e80–e97. 28 Ross EA, Perloff JK, Danovitch GM et al. adults with congenital heart disease. JAm 16 Cannesson M, Earing MG, Collange V Renal function and urate metabolim in late Coll Cardiol 2001; 37: 1183–1187. et al. Anesthesia for noncardiac surgery in survivors with cyanotic congenital heart dis- 4 Landzberg MJ, Murphy DJ Jr, Davidson adults with congenital heart disease. Anes- ease. Circulation 1986; 73: 396–400. WR et al. Task force 4: organization of thesiology 2009; 111: 432–440. 29 Geva T, Sandweiss BM, Gauvreau K et al. delivery systems for adults with congenital 17 Chassot PG, Bettox DA. Anesthesia and Factors associated with impaired clinical heart disease. J Am Coll Cardiol 2001; 37: adult congenital heart disease. J Cardiotho- status in long-term survivors of tetralogy of 1187–1193. rac Vasc Anesth 2006; 20: 414–437. Fallot repair evaluated by magnetic reso- 5 Skorton DJ, Garson A Jr, Allen HD et al. 18 Lovell AT. Anaesthetic implications of nance imaging. J Am Coll Cardiol 2004; 43: Task force 5: adults with congenital heart grown-up congenital heart disease. Br J 1068–1074. disease: access to care. J Am Coll Cardiol Anaesth 2004; 93: 129–139. 30 Triedman JK. Arrhythmias in adults with 2001; 37: 1193–1198. 19 Warnes CA, Williams RG, Bashore TM congenital heart disease. Heart 2002; 87: 6 Webb G. The long road to better ACHD et al. ACC/AHA 2008 Guidelines for the 383–389. care. Congenit Heart Dis 2010; 5: 198–205. Management of Adults With Congenital 31 Dowling M, Hatem AJ, Hardman JG et al. 7 Moodie D. The long road—a destination Heart Disease. J Am Coll Cardiol 2008; 52: Real-time three-dimensional ultrasound- not yet reached. Congenit Heart Di. 2010; 5: 1890–1947. guided central venous catheter placement. 206–207. 20 Report of the British Cardiac Society Work- Anesth Analg 2011; 112: 378–381. 8 Moodie D. Adult congenital patients in the ing Party. Grown-up congenital heart 32 Burrows FA. Physiologic deadspace, venous hospital—where the rubber meets the road. (GUCH) disease: current needs and provi- admixture, and the arterial end-tidal carbon Congenit Heart Dis 2009; 4: 423. sion of service for adolescents and adults dioxide difference in infants and children 9 Marelli AJ, Mackie AS, Ionescu-Ittu R et with congenital heart disease in the UK. undergoing cardiac surgery. Anesthesiology al. Congenital heart disease in the general Heart 2002; 88(Suppl 1): i1–i14. 1989; 70: 219–225. population, changing prevalence and age 21 Silversides CK, Dore A, Poirier N et al. 33 Wilson W, Taubert KA, Gewitz M et al. distribution. Circulation 2007; 115: CCS 2009 Consensus Conference on the Prevention of infective endocarditis. Guide- 163–172. management of adults with congenital heart lines from the American Heart Association. 10 Moller JH, Taubert KA, Allen HD et al. A disease: Shunt lesions. Can J Cardiol 2010; A Guideline From the American Heart Special Writing Group from the Task Force 26: e70–e79. Association Rheumatic Fever, Endocarditis, on Children and Youth, American Heart 22 Silversides CK, Salehian O, Oechslin E et al. and Kawasaki Disease Committee, Council Association. Cardiovascular health and dis- CCS 2009 Consensus Conference on the on Cardiovascular Disease in the Young, ease in children: current status. Circulation Management of Adults with Congenital and the Council on Clinical Cardiology, 1994; 89: 923–930. Heart Disease: complex congenital cardiac Council on Cardiovascular Surgery and 11 Hoffman JI, Christianson R. Congenital lesions. Can J Cardiol 2010; 26: e98–e117. Anesthesia, and the Quality of Care and heart disease in a cohort of 19,502 births 23 Deanfield J, Thaulow E, Warnes C et al. Outcomes Research Interdisciplinary Work- with long-term follow-up. Am J Cardiol Management of grown up congenital heart ing Group. Circulation 2007; 116: 1736– 1978; 42: 641–647. disease. Eur Heart J 2003; 24: 1035–1084. 1754. 12 Hoffman JIE, Kaplan S, Liberthson RR. 24 Kantoch MJ, Collins-Nakai RL, Medwid S 34 Tempe DK, Virmani S. Coagulation abnor- Prevalence of congenital heart disease. Am et al. Adult patients’ knowledge about their malities in patients with cyanotic congenital Heart J 2004; 147: 425–439. congenital heart disease. Can J Cardiol heart disease. J Cardiothorac Vasc Anesth 13 Warnes CA. The adult with congenital heart 1997; 13: 641–645. 2002; 16: 752–765. disease: born to be bad? J Am Coll Cardiol 25 Ammash N, Warnes CA. Cerebrovascular 35 Salehian O, Schwerzmann M, Rambihar S 2005; 46: 1–8. events in adult patients with cyanotic con- et al. Left ventricular dysfunction and

Pediatric Anesthesia 21 (2011) 615–622 ª 2011 Blackwell Publishing Ltd 621 Adult congenital heart disease R. Seal

mortality in adult patients with Eisenmenger before and after palliative and reparative 44 Gengsakul A, Harris L, Bradley TJ et al. syndrome. Congenit Heart Dis 2007; 2: 156– operation in tetralogy of Fallot. Circulation The impact of pulmonary valve replacement 164. 1976; 54: 417–423. after Tetralogy of Fallot repair: a matched 36 Graham TP. Ventricular performance in 41 Ilbawi MN, Idriss FS, DeLeon SY et al. comparison. Eur J Cardiothorac Surg 2007; congenital heart disease. Circulation 1991; Factors that exaggerate the deleterious 32: 462–468. 84: 2259–2274. effects of pulmonary insufficiency on the 45 Therrien JP, Sui SC, Harris L et al. Impact 37 Didonato RM, Jonas RA, Lang P et al. right ventricle after tetralogy repair. Surgical of pulmonary valve replacement on arrhyth- Neonatal repair of tetralogy of Fallot with implications. J Thorac Cardiovasc Surg mia propensity late after repair of tetralogy and without pulmonary atresia. J Thorac 1987; 93: 36–44. of Fallot. Circulation 2001; 103: 2489–2494. Cardiovasc Surg 1991; 101: 126–137. 42 Gatzoulis MA, Till JA, Somerville J et al. 46 McHugh KE, Hillman DG, Gurka MJ et al. 38 Pigula FA, Khalil PN, Mayer JE et al. Mechano-electrical interaction in tetralogy Three-stage palliation of hypoplastic left Repair of tetralogy of Fallot in neonates of Fallot: QRS prolongation relates to right heart syndrome in the University Health and young infants. Circulation 1999; ventricular size and predicts malignant ven- System Consortium. Congenit Heart Dis 100(Suppl 2): 157–161. tricular arrhythmias and sudden death. Cir- 2010; 5: 8–15. 39 Abarbanell GL, Goldberg CS, Devaney EJ culation 1995; 92: 231–237. 47 Siu SC, Sermer M, Colman JM et al. Pro- et al. Early surgical morbidity and mortality 43 Therrien J, Siu SC, McLaughlin PR et al. spective multicenter study of pregnancy out- in adults with congenital heart disease: the Pulmonary valve replacement in adults late comes in women with heart disease. University of Michigan experience. Congenit after repair of tetralogy of Fallot: are we Circulation 2001; 104: 515–521. Heart Dis 2008; 3: 82–89. operating too late? J Am Coll Cardiol 2000; 48 Somerville J. Management of adults with 40 Graham TP Jr, Cordell D, Atwood GF et 36: 1670–1675. congenital heart disease: an increasing prob- al. Right ventricular volume characteristics lem. Annu Rev Med 1997; 48: 283–293.

622 Pediatric Anesthesia 21 (2011) 615–622 ª 2011 Blackwell Publishing Ltd Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Approach to managing children with heart disease for noncardiac surgery Michelle C. White

Department of Paediatric Anaesthesia, Bristol Royal Hospital for Children, Bristol, UK

Keywords Summary congenital heart disease; cardiac; noncardiac surgery; children; anesthesia; complications Congenital heart disease is the commonest birth defect, and advances in modern medicine mean 90% of these children now survive to adulthood. Correspondence Therefore, many children present to their local hospital requiring general Dr Michelle C. White, anesthesia for common childhood conditions. They pose a challenge for Department of Paediatric Anaesthesia, anesthesia because perioperative morbidity and mortality is greater com- Bristol Royal Hospital for Children pared with other children. It is impossible to prescribe a formula for Marlborough Street, Bristol BS2 8BJ, UK Email: [email protected] anesthetizing children with heart disease because of the complexity of heart defects and the variety of noncardiac surgery. There is also a lack of high- Section Editor: Greg Hammer quality data of efficacy of one anesthetic technique over another. Much data come from case series or isolated case reports. In a rapidly advancing Accepted 9 August 2010 field such as cardiac surgery, studies of long-term complications may be out of date by the time they are published, limiting applicability of the doi:10.1111/j.1460-9592.2010.03416.x results. Because of these factors, claims of efficacy and safety of various approaches to managing children with heart disease for noncardiac surgery must be interpreted cautiously. This narrative review aims to present the evidence concerning a range of anesthetic techniques, the long-term compli- cations of congenital heart disease and suggest a physiological and evi- dence-based approach to managing these children.

the majority of cardiac arrests occurred during noncar- Introduction diac surgery; 75% were under 2 years old, and 75% of Congenital heart disease is the commonest birth defect, deaths were accounted for by three distinct defects: occurring in approximately one in 125 live births. aortic stenosis, cardiomyopathy, and single ventricle Extracardiac anomalies requiring surgery within the lesions. Torres et al. (9) report a 19% mortality in chil- first year of life are present in 30% (1,2). Advances in dren with hypoplastic left heart syndrome (HLHS) medical care have resulted in 90% of children with <2 years old presenting for noncardiac surgery. Peri- heart disease surviving to adulthood (3,4). These chil- operative intensive care support is needed by 67% of dren present needing both elective and emergency children with cardiomyopathy (11). However, despite anesthesia because they are subject to the same child- these challenges, neonates with complex disease requir- hood ailments as other children. Some have mild heart ing major surgery can tolerate general anesthesia with disease with no complications, whereas others have few complications (13). complex disease with multiorgan dysfunction (5,6). Children with heart disease undergoing noncardiac Aims and limitations surgery are at increased risk of perioperative morbidity and mortality compared with other children (7–12). This narrative review aims to discuss the literature These children pose a serious challenge for anesthesia. concerning different anesthetic techniques, the major Ramamoorthy et al. (10) reviewed anesthesia-related long-term complications of heart disease and suggest pediatric cardiac arrests. In children with heart disease, an evidence-based approach to managing these chil-

522 Pediatric Anesthesia 21 (2011) 522–529 ª 2010 Blackwell Publishing Ltd M.C. White Managing children for noncardiac surgery dren for noncardiac surgery. However, there are three adults (27,28) and in children with sepsis (29). The use main limitations. of numerous muscle relaxants (suxamethonium, vecur- First, few studies exist evaluating anesthetic techni- onium, rocuronium, and pancuronium) and volatile ques in children with heart disease undergoing noncar- agents are also described (18–25,30). diac surgery. Most data come from cardiac Isoflurane, sevoflurane, and fentanyl/midazolam catheterization. Surgery involving major fluid shifts, infusions for maintenance of anesthesia have been changes in position (prone or Trendelenburg) or evaluated during cardiac catheterization (30), or prior changes in intracavity pressure (laparoscopic techni- to cardiac surgery (31,32). In children with shunt, ques) may produce important hemodynamic and pul- neither isoflurane, sevoflurane nor fentanyl/midazolam monary effects that alter risk. Studies in these have any effect on shunt fraction (31). In children with situations are nonexistent. Therefore, claims of efficacy single ventricle lesions, neither sevoflurane nor fenta- and safety must be interpreted cautiously. Small case nyl/midazolam have any effect on myocardial contrac- series and isolated case reports do exist and are dis- tility (30). In contrast, Rivenes et al. (32) report cussed. Secondly, in a rapidly advancing field such as fentanyl/midazolam reduces cardiac index and contrac- cardiac surgery, studies of long-term complications tility. Rivenes’s study used a heterogeneous group of may reflect outdated treatment strategies. This makes children and so should be interpreted cautiously as dif- application to current practice difficult. Thirdly, ferent lesions may produce different pharmacody- because of the varied and complex nature of pediatric namics. Hannon et al. (33) reported sevoflurane and heart disease, a ‘one-size-fits-all’ approach to manage- isoflurane have a greater depressive effect on myocar- ment is impossible. Therefore, the approach presented dial contractility in ferrets with a here is general rather than specific and based on evi- band than in unbanded animals. In several large cases dence as well as physiology. series of high-risk children for noncardiac surgery, maintenance with isoflurane and sevoflurane is described (13,18,20,34). The evidence cannot recom- Anesthetic techniques mend one agent over another but suggests isoflurane The two most studied induction agents in children with and sevoflurane are routinely used, whereas desflurane heart disease for noncardiac surgery are ketamine and and propofol infusions are not. propofol. In children undergoing cardiac catheteriza- Postoperative analgesia depends on the type of sur- tion, propofol decreased systemic vascular resistance gery. For major surgery, intravenous opioid infusions (SVR) and mean arterial pressure (MAP) (14). Heart are commonly used (13). Regional techniques have rate, pulmonary arterial pressure (PAP), and pulmon- been used but reports are rare. Sacrista et al. (35) ary vascular resistance (PVR) were unchanged. In chil- describe spinal anesthesia alone in an infant with dren with shunt, left-to-right flow decreases and right- HLHS undergoing colostomy formation for anorectal to-left flow increases causing a clinically relevant atresia. The spinal block was measured to the 4th thor- reduction in oxygen saturation. Oklu et al. (15) com- acic vertebra level with no adverse hemodynamic pared propofol with ketamine. Propofol decreased effects. High spinal blocks are well tolerated in healthy both MAP and SVR, whereas ketamine increased children and are not associated with the reductions in MAP with no effect on SVR. Neither drug had any MAP and heart rate seen in adults (36,37). Epidural effect on PAP or PVR. Ketamine is also well tolerated with general anesthesia is reported in neonates with in children with pulmonary hypertension (PHT) complex cardiac defects undergoing general surgery (16,17). In a review of 18 neonates with complex car- (13). Epidural infusions of ropivicaine were used and diac defects undergoing major general surgery, keta- managed in a non-ICU environment with no complica- mine was the commonest induction agent in those not tions reported. A recent review cites ropivacaine as intubated at the time of surgery (13). Thiopentone, having the greatest margin of safety of all long-acting etomidate, and sevoflurane use is also reported for the local anesthetics (38). induction of anesthesia (18–25). These reports cover a In summary, many different anesthetic techniques wide range of cardiac conditions including single ven- have been described. Propofol dramatically reduces tricle lesions and encompass a variety of procedures SVR, whereas ketamine has no effect, making keta- from trachea-oesophageal fistula repair to laparoscopic mine the agent of choice when a reduction in SVR is surgery. Etomidate causes minimal hemodynamic dis- undesirable or in children with PHT (14–17). Isoflur- turbances in children with shunt lesions undergoing ane and sevoflurane have minimal effect on myo- cardiac catheterization (26). However, the adrenal sup- cardial contractility or shunt fraction and their use is pressive effects of etomidate are well described in widely reported for the maintenance of anesthesia

Pediatric Anesthesia 21 (2011) 522–529 ª 2010 Blackwell Publishing Ltd 523 Managing children for noncardiac surgery M.C. White

(13,20,30,31,34). The effects of desflurane and propofol It has been suggested that factors known to lower the infusions are unknown in this patient group. Opioid threshold for VE should be avoided, for example infusions, spinal and epidural anesthesia have all been hypercarbia, acidosis, hypoxia, and large doses of local used effectively (13,35). anesthetic with epinephrine (39).

Four major complications of heart disease in chil- Cardiac failure dren Cardiac failure is the end result of a continually Arrhythmias volume- or pressure-overloaded heart. Limited cardiac Some cardiac operations are associated with an reserve describes the situation where the heart is func- increased risk of late-onset arrhythmias. Surgery invol- tioning at near maximal capacity even while the child ving extensive atrial sutures or ventriculotomy poses is resting (39). Children with limited cardiac reserve the highest risk. Arrhythmias can occur years later and must be identified because they are at high risk of car- be asymptomatic or associated with hemodynamic col- diac failure during anesthesia. Plasma noradrenaline lapse and sudden death. Necrosis and progressive levels are increased in children with cardiac failure fibrosis extending into the conduction system are (58), and anesthetics that reduce sympathetic tone may thought to be the causes (39,40). precipitate severe cardiac failure. In this situation, pro- Atrial arrhythmias are commonest after repair of pofol can have a profoundly deleterious effect on car- sinus venous defects, atrial septal defects, atrial switch diac output, whereas ketamine has minimal effect procedures (Mustard or Senning), and total cavopul- (14,15). Both inhalational and intravenous inductions monary anastomosis (). Damage to are prolonged, so patience is required to prevent exces- the atrioventricular (AV) node and Bundle of His is sive drug administration. The use of vasoactive agents commonest after ventriculotomy or right ventricle to and invasive monitoring may be needed even for minor pulmonary artery conduit (39,40). Table 1 summarizes surgery. Depending on the cause of cardiac failure and the evidence. However, these data need to be inter- the risk of poor coronary perfusion, epinephrine and a preted cautiously, because modern surgical techniques vasoconstrictor such as phenylephrine should be imme- and trends toward earlier repair in infancy may reduce diately available. Major surgery involving significant but not eliminate these risks (41–45). blood loss and/or fluid shifts poses high risk, and full Some children with heart disease have a primary pediatric intensive care support should be available malignant arrhythmia. Prolongation of the QT interval (11). Further management of children with cardiac is associated with torsades de pointes especially in chil- failure is discussed elsewhere in this supplement. dren with long QT syndromes. Optimal anesthesia for these children is not well described. The evidence sug- Pulmonary hypertension (PHT) gests propofol and sevoflurane (55,56) have minimal effect on the QTc interval (QT interval corrected for PHT is defined as having a mean PAP above heart rate), whereas desflurane is known to prolong 25 mmHg at rest or 30 mmHg on exercise (59,60). QTc in normal children (57).All children with heart Children at high risk are those with (i) excessive pul- disease should have a preoperative electrocardiogram monary blood flow (PBF) (left-to-right shunt) or (ii) (ECG). Ventricular ectopics (VE) are an ominous sign prolonged pulmonary venous obstruction or high left as 30% of these patients eventually die suddenly (42). atrial pressure.

Table 1 Arrhythmia risk

Procedure Risk References

Atrial Switch (Mustard or Senning) Approximately 50% have sinus node dysfunction; 10% incidence of (46–48) ventricular dysfunction and arrhythmia Ebsteins Atrial tachycardia. Arrhythmias and sudden death during placement of (49,50) intracardiac lines Single ventricle circulation Up to 30% risk of arrhythmias leading to sudden death (51,52) Tetralogy of Fallot Up to 5% risk of sudden death or sustained ventricular tachycardia. Right (41–44,53,54) ventricular outflow tract obstruction increases risk of ventricular arrhythmias Poor exercise tolerance is suggestive of RV failure is a risk factor for VT and sudden death.

524 Pediatric Anesthesia 21 (2011) 522–529 ª 2010 Blackwell Publishing Ltd M.C. White Managing children for noncardiac surgery

Documented PHT is a clear predictor of periopera- reduced, especially if iron deficiency is present because tive morbidity (12,34,61). Children with suprasystemic of red cell rigidity (70). Hyperviscosity is associated PAP are eight times more likely to experience a major with cerebral vein and sinus thrombosis resulting in complication than those with subsystemic PAP (61). stroke. Children under 5 years are at highest risk espe- Treatment with 100% oxygen, inhaled nitric oxide, cially in the presence of dehydration, fever, or iron intravenous prostacylin, inotropic support of the right deficiency (71,72). All children with cyanosis should ventricle, and other measures to maintain cardiac have hemoglobin and hematocrit measured and preo- output and PBF may all be required. Therefore, full perative fasting kept to a minimum. Hemoglobin pediatric intensive care facilities should be available. greater than 18 gÆdl)1 (or hematocrit >60%) is consid- Children with PHT also have reduced pulmonary ered by some as an indication for preoperative intrave- compliance and increased airway resistance causing nous fluid therapy in order to minimize risk. increased work of breathing (62,63). Therefore, Abnormal laboratory tests of hemostasis are present respiratory tract infections may be poorly tolerated in 20% of children with cyanosis (73). They correlate and have a greater impact on PVR than ordinarily with the degree of cyanosis but the etiology is expected. unknown. Prolongation of prothrombin time or partial PHT develops rapidly in association with specific thromboplastin time is the commonest abnormality. lesions (AVSD, TA, D-TGA with VSD) and in certain Others include thrombocytopenia, platelet dysfunction, patient populations (Down’s Syndrome) (64). It causes hypofibrinogenemia, and accelerated fibrinolysis (74). fixed structural changes in the vascular bed but with Even when coagulation tests are normal and children reactive vascular smooth muscle affected by several are asymptomatic, the risk of excessive postoperative factors. Acidosis, hypercarbia, hypoxia, hypothermia, bleeding is still real. All cyanotic children should be increased sympathetic stimulation, and increased air- assumed at risk. If a child is receiving aspirin therapy way pressure all increase PVR (65). Hydrogen ion con- to maintain shunt patency, the risk of shunt thrombo- centration actually exerts a greater effect than carbon sis is usually greater than that of bleeding, so aspirin dioxide concentration (66). Appropriate management therapy should continue. Also, as the effects of aspirin of these variables reduces PVR, improves RV function, last many days, discontinuing the drug immediately and minimizes right-to-left shunting. before surgery has little effect. The anesthetic management of children (67) and Animal models show reduced ability to increase car- adults (68) with PHT has been recently reviewed. The diac output in response to increased demands (75,76). importance of understanding the specific intracardiac Therefore, premedication is commonly used to mini- anatomy and the physiological consequences of mize distress on induction. In children with right-to- changes in PVR, SVR, and shunt flow cannot be over left shunts, end tidal carbon dioxide (CO2) monitoring emphasized. Different situations require different stra- underestimates arterial CO2 (77). Profound hypoxia tegies. If in doubt, advice from a pediatric cardiac can also occur, especially in conjunction with respira- anesthetist and/or pediatric cardiologist should be tory depressants such as opioids, because the hypoxic sought before proceeding. ventilatory response is blunted (78). So, careful post- operative monitoring including oxygen saturation is recommended (39,79). Cyanosis Cyanosis is a common feature of unrepaired or par- Approach to preoperative assessment and tially palliated congenital heart disease. It is usually anesthetic management the result of decreased PBF and right-to-left shunting but children with increased PBF can also suffer from Children with heart disease presenting for noncardiac cyanosis. Children with cyanosis often have concurrent surgery are at increased risk of perioperative morbidity cardiac failure, PHT, and arrhythmias making them a and mortality. Table 2 summarizes the children at high-risk group. Chronic cyanosis affects most major highest risk (7–12). No data exist on patient outcome organ systems in the body but this discussion focuses or satisfaction regarding anesthetizing these children in primarily on the hematological consequences. their local hospital or moving them to an expert cen- Hypoxia causes an increase in erythropoietin thereby ter. In Slater et al.’s series (18) of laparoscopic surgery increasing hematocrit, hemoglobin, and viscosity (69). in neonates with HLHS, all were anesthetized by a This allows increased oxygen delivery without a sus- pediatric cardiac anesthetist. Conversely, in Walker tained increase in cardiac output. However, if hemato- et al.’s series (13) of major general surgery in neonates crit levels exceed 65%, oxygen delivery can actually be with complex cardiac defects, not all infants were

Pediatric Anesthesia 21 (2011) 522–529 ª 2010 Blackwell Publishing Ltd 525 Managing children for noncardiac surgery M.C. White

Table 2 Factors associated with high risk of perioperative morbidity symptoms and/or is under active cardiology follow- (see text for references) up); perform preoperative ECG in those at risk. Exam- Less than 2 years old ination of peripheral and central veins gives forewarn- Complex lesions (including single ventricle physiology, ing of cannulation difficulties and may influence cardiomyopathy, aortic stenosis) anesthetic technique. Major surgery Drug history: Many children receive numerous medi- Emergency surgery cations such as aspirin diuretics, angiotensin convert- Presence of long-term sequelae (arrhythmia, cardiac failure, ing enzyme (ACE) inhibitors, and anti-arrythmics. pulmonary hypertension, cyanosis) Cardiac medications are not associated with clinically important electrolyte disturbances, so routine preo- anesthetized by specialist cardiac anesthetists, but full perative electrolyte testing is unnecessary (80). Gener- cardiology, cardiac anesthesia, and intensive care sup- ally, all cardiac medications should be given on the port were available. Professional responsibility dictates morning of surgery (40). However, some anesthetists that the anesthetist must understand the cardiac ana- prefer to omit ACE inhibitors based on studies of tomy and physiology. Depending on the lesion, this adults with ischemic heart disease and hypertension in includes ‘balanced’ (parallel) and single ventricle phy- which hypotension on induction of anesthesia is siology; factors affecting SVR, PVR, and shunt flow reported (81–83), although other adult work dismisses including anesthetic drugs and ventilator settings. these claims (81,84,85). Data in children are lacking. These are outlined in more detail elsewhere (39,40,79). Aspirin therapy to prevent shunt thrombosis should Given that high-risk children are likely to require spe- usually be continued. Children on warfarin therapy cialist support and have increased perioperative mor- need to be admitted to hospital for anticoagulant mon- bidity (7–12), an evidence-based approach suggests all itoring and establishment on intravenous heparin prior high-risk children are anesthetized where specialist to surgery. facilities are available. However, this may not be feasi- Premedication: Commonly used to avoid distress, ble in every situation especially those children present- minimize oxygen consumption, and may reduce the ing for emergency surgery. In such situations, good amount of induction agent required so minimizing early communication and co-operation is essential. The reductions in SVR. Cyanotic children and those with local hospital should telephone the specialist center for heart failure will require monitoring of vital signs after advice, and where possible think about early transfer administration and may require oxygen therapy to for any high-risk child who may require surgery. Spe- maintain oxygen saturations. cialist centers should realize their responsibilities and Endocarditis prophylaxis: Appropriate guidelines facilitate the transfer of such children. must be followed (86,87). Meticulous preoperative preparation involves (i) rou- Communication: Good communication between all tine anesthetic history and examination; (ii) evaluation health care professionals, the child, and their family is of long-term complications (anesthetic implications are important including discussion of associated risks. discussed earlier), and other features that put children in a high-risk category (see Table 1); (iii) finally, speci- Conclusions fic attention to the following areas that are easily over- looked: Children with heart disease are at increased risk of Respiratory: Respiratory tract infections have a perioperative complications. The cardiovascular anat- greater impact on PVR than expected in children with omy and physiology is complex and each case requires PHT or cavopulmonary anastomosis (62,63). Oxygen individual evaluation of risk factors. With the excep- saturations compared with predicted values for the tion of ketamine for children with PHT or where redu- specific defect give an indication of the amount of cing SVR is deleterious, there is no evidence to PBF, which in turn influences the risks of PHT, right recommend one anesthetic technique in favor of ventricular failure, and cyanosis. another. There is no evidence of patient outcome or Cardiovascular: Features of arrhythmias, cardiac satisfaction regarding treatment in the local hospital or failure, and PHT should be sought. Review previous expert center. However, high-risk children are likely to echocardiography and cardiac catheterization reports require specialist support, so these facilities should be (these may need repeating if the child has ongoing available where possible.

526 Pediatric Anesthesia 21 (2011) 522–529 ª 2010 Blackwell Publishing Ltd M.C. White Managing children for noncardiac surgery

References 1 Greenwood RD, Rosenthal A, Parisi L children with congenital heart disease. shunt lesions. J Cardiothorac Vasc Anesth et al. Extracardiac abnormalities in infants Anesth Analg 1999; 89: 1411–1416. 2010. Apr 21 [Epub ahead of print]. with congenital heart disease. Pediatrics 15 Oklu E, Bulutcu FS, Yalcin Y et al. Which 27 Daniell H. Opioid and benzodiazepine con- 1975; 55: 485–492. anesthetic agent alters the hemodynamic sta- tributions to etomidate-associated adrenal 2 Hoffman JI, Christianson R. Congenital tus during pediatric catheterization? Com- insufficiency. Intensive Care Med 2008; 34: heart disease in a cohort of 19,502 births parison of propofol versus ketamine. J 2117–2118. with long-term follow-up. Am J Cardiol Cardiothorac Vasc Anesth 2003; 17: 686– 28 Hildreth AN, Mejia VA, Maxwell RA et al. 1978; 42: 641–647. 690. Adrenal suppression following a single dose 3 Deanfield J, Thaulow E, Warnes C et al. 16 Williams GD, Maan H, Ramamoorthy C of etomidate for rapid sequence induction: a Management of grown up congenital et al. Perioperative complications in children prospective randomized study. J Trauma heart disease. Eur Heart J 2003; 24: 1035– with pulmonary hypertension undergoing 2008; 65: 573–579. 1084. general anesthesia with ketamine. Pediatr 29 den BM, Hokken-Koelega AC, Hazelzet JA 4 Warnes CA, Liberthson R, Danielson GK Anesth 2010; 20: 28–37. et al. One single dose of etomidate nega- et al. Task force 1: the changing profile of 17 Williams GD, Philip BM, Chu LF et al. tively influences adrenocortical performance congenital heart disease in adult life. JAm Ketamine does not increase pulmonary vas- for at least 24h in children with meningo- Coll Cardiol 2001; 37: 1170–1175. cular resistance in children with pulmonary coccal sepsis. Intensive Care Med 2008; 34: 5 Dimopoulos K, Diller GP, Koltsida E et al. hypertension undergoing sevoflurane 163–168. Prevalence, predictors, and prognostic value anesthesia and spontaneous ventilation. 30 Ikemba CM, Su JT, Stayer SA et al. Myo- of renal dysfunction in adults with congeni- Anesth Analg 2007; 105: 1578–1584. table. cardial performance index with sevoflurane- tal heart disease. Circulation 2008; 117: 18 Slater B, Rangel S, Ramamoorthy C et al. pancuronium versus fentanyl-midazolam- 2320–2328. Outcomes after laparoscopic surgery in neo- pancuronium in infants with a functional 6 Khairy P, Landzberg MJ. Adult congenital nates with hypoplastic heart left heart syn- single ventricle. Anesthesiology 2004; 101: heart disease: toward prospective risk assess- drome. J Pediatr Surg 2007; 42: 1118–1121. 1298–1305. ment of a multisystemic condition. Circula- 19 Mariano ER, Chu LF, Albanese CT et al. 31 Laird TH, Stayer SA, Rivenes SM et al. tion 2008; 117: 2311–2312. Successful thoracoscopic repair of esopha- Pulmonary-to-systemic blood flow ratio 7 Hennein HA, Mendeloff EN, Cilley RE geal atresia with tracheoesophageal fistula in effects of sevoflurane, isoflurane, halothane, et al. Predictors of postoperative outcome a newborn with single ventricle physiology. and fentanyl/midazolam with 100% oxygen after general surgical procedures in patients Anesth Analg 2005; 101: 1000–1002. table. in children with congenital heart disease. with congenital heart disease. J Pediatr Surg 20 Mariano ER, Boltz MG, Albanese CT et al. Anesth Analg 2002; 95: 1200–1206. table. 1994; 29: 866–870. Anesthetic management of infants with pal- 32 Rivenes SM, Lewin MB, Stayer SA et al. 8 Baum VC, Barton DM, Gutgesell HP. Influ- liated hypoplastic left heart syndrome Cardiovascular effects of sevoflurane, iso- ence of congenital heart disease on mortality undergoing laparoscopic nissen fundoplica- flurane, halothane, and fentanyl-midazolam after noncardiac surgery in hospitalized chil- tion. Anesth Analg 2005; 100: 1631–1633. in children with congenital heart disease: an dren. Pediatrics 2000; 105: 332–335. 21 Rice-Townsend S, Ramamoorthy C, Dutta echocardiographic study of myocardial con- 9 Torres A Jr, DiLiberti J, Pearl RH et al. S. Thoracoscopic repair of a type D esopha- tractility and hemodynamics. Anesthesiology Noncardiac surgery in children with hypo- geal atresia in a newborn with complex con- 2001; 94: 223–229. plastic left heart syndrome. J Pediatr Surg genital heart disease. J Pediatr Surg 2007; 33 Hannon JD, Cody MJ, Sun DX et al. 2002; 37: 1399–1403. 42: 1616–1619. Effects of isoflurane and sevoflurane on 10 Ramamoorthy C, Haberkern CM, Bhanan- 22 Caliskan E, Bozdogan N, Kocum A et al. intracellular calcium and contractility in ker SM et al. Anesthesia-related cardiac Anesthetic management of adenotonsillect- pressure-overload hypertrophy. Anesthesiol- arrest in children with heart disease: data omy in a child with bidirectional superior ogy 2004; 101: 675–686. from the Pediatric Perioperative Cardiac cavapulmonary shunt. Pediatr Anesth 2008; 34 Taylor CJ, Derrick G, McEwan A et al. Arrest (POCA) registry. Anesth Analg 2010; 18: 996–997. Risk of cardiac catheterization under 110: 1376–1382. 23 Hanamoto H, Tachibana K, Kinouchi K anaesthesia in children with pulmonary 11 Kipps AK, Ramamoorthy C, Rosenthal et al. Postoperative management of cleft lip hypertension. Br J Anaesth 2007; 98: 657– DN et al. Children with cardiomyopathy: repair using dexmedetomidine in a child 661. complications after noncardiac procedures with bidirectional superior cavopulmonary 35 Sacrista S, Kern D, Fourcade O et al. with general anesthesia. Pediatr Anesth shunt. Pediatr Anesth 2008; 18: 350–352. Spinal anaesthesia in a child with hypoplas- 2007; 17: 775–781. 24 Hardy C. Pyloromyotomy in an infant with tic left heart syndrome. Paediatr Anaesth 12 Warner MA, Lunn RJ, O’Leary PW et al. hypoplastic left heart syndrome status-post 2003; 13: 253–256. Outcomes of noncardiac surgical procedures hybrid procedure: not just another case? 36 Imbelloni LE, Vieira EM, Sperni F et al. in children and adults with congenital heart Pediatr Anesth 2008; 18: 993–994. Spinal anesthesia in children with isobaric disease. Mayo Perioperative Outcomes 25 Sinha R, Roy K, Rajeshwari S et al. Anes- local anesthetics: report on 307 patients Group. Mayo Clin Proc 1998; 73: 728–734. thetic management of a child with congeni- under 13 years of age. Pediatr Anesth 2006; 13 Walker A, Stokes M, Moriarty A. Anesthe- tal heart disease and pulmonary artery 16: 43–48. sia for major general surgery in neonates banding for ophthalmic surgery. Pediatr 37 Wu CL, Fleisher LA. Outcomes research in with complex cardiac defects. Pediatr Anesth Anesth 2008; 18: 994–996. regional anesthesia and analgesia. Anesth 2009; 19: 119–125. 26 Dhawan N, Chauhan S, Kothari SS et al. Analg 2000; 91: 1232–1242. 14 Williams GD, Jones TK, Hanson KA et al. Hemodynamic responses to etomidate in 38 Zink W, Graf BM. The toxicity of local The hemodynamic effects of propofol in pediatric patients with congenital cardiac anesthetics: the place of ropivacaine and

Pediatric Anesthesia 21 (2011) 522–529 ª 2010 Blackwell Publishing Ltd 527 Managing children for noncardiac surgery M.C. White

levobupivacaine. Curr Opin Anaesthesiol of Fallot: a clinicopathologic study. Circula- 67 Friesen RH, Williams GD. Anesthetic man- 2008; 21: 645–650. tion 1983; 67: 626–631. agement of children with pulmonary arterial 39 Frankville D. Anesthesia for children and 54 Chandar JS, Wolff GS, Garson A Jr et al. hypertension. Pediatr Anesth 2008; 18: 208– adults with congenital heart disease. In: Ventricular arrhythmias in postoperative tet- 216. Lake CL, Booker PD, eds. Pediatric Car- ralogy of Fallot. Am J Cardiol 1990; 65: 68 Fischer LG, Van Aken H, Burkle H. Man- diac Anesthesia. Philadelphia: Lippincott 655–661. agement of pulmonary hypertension: physio- Williams & Wilkins, 2005: pp. 601–632. 55 Curry TB, Gaver R, White RD. Acquired logical and pharmacological considerations 40 Diaz LK, Hall S. Anesthesia for non-cardiac long QT syndrome and elective anesthesia for anesthesiologists. Anesth Analg 2003; 96: surgery and magnetic resonance imaging. in children. Pediatr Anesth 2006; 16: 471– 1603–1616. In: Andropoulos DB, Stayer SA, Russell 478. 69 Gidding SS, Stockman JA III. Erythropoie- IA, eds. Anesthesia for Congenital Heart 56 Whyte SD, Booker PD, Buckley DG. The tin in cyanotic heart disease. Am Heart J Disease. Massachusetts: Blackwell Publish- effects of propofol and sevoflurane on the 1988; 116: 128–132. ing, 2005: pp. 427–452. QT interval and transmural dispersion of 70 Linderkamp O, Klose HJ, Betke K et al. 41 Walsh EP, Rockenmacher S, Keane JF et repolarization in children. Anesth Analg Increased blood viscosity in patients with al. Late results in patients with tetralogy of 2005; 100: 71–77. cyanotic congenital heart disease and iron Fallot repaired during infancy. Circulation 57 Aypar E, Karagoz AH, Ozer S et al. The deficiency. J Pediatr 1979; 95: 567–569. 1988; 77: 1062–1067. effects of sevoflurane and desflurane 71 Wedemeyer AL, Edson JR, Krivit W. Coa- 42 Garson A Jr, Randall DC, Gillette PC et al. anesthesia on QTc interval and cardiac gulation in cyanotic congenital heart dis- Prevention of sudden death after repair of rhythm in children. Pediatr Anesth 2007; 17: ease. Am J Dis Child 1972; 124: 656–660. tetralogy of Fallot: treatment of ventricular 563–567. 72 Phornphutkul C, Rosenthal A, Nadas AS arrhythmias. J Am Coll Cardiol 1985; 6: 58 Ross RD, Daniels SR, Schwartz DC et al. et al. Cerebrovascular accidents in infants 221–227. Plasma norepinephrine levels in infants and and children with cyanotic congenital heart 43 Hokanson JS, Moller JH. Adults with tet- children with congestive heart failure. Am J disease. Am J Cardiol 1973; 32: 329–334. ralogy of Fallot: long-term follow-up. Car- Cardiol 1987; 59: 911–914. 73 Colon-Otero G, Gilchrist GS, Holcomb GR diol Rev 1999; 7: 149–155. 59 Rich S. Primary pulmonary hypertension: et al. Preoperative evaluation of hemostasis 44 Lucron H, Marcon F, Bosser G et al. executive summary from world symposium in patients with congenital heart disease. Induction of sustained ventricular tachycar- – primary pulmonary hypertension. World Mayo Clin Proc 1987; 62: 379–385. dia after surgical repair of tetralogy of Fal- Health Organisation. 1998. Available at: 74 Henriksson P, Varendh G, Lundstrom NR. lot. Am J Cardiol 1999; 83: 1369–1373. http://www.who.int/cardiovascular_diseases/ Haemostatic defects in cyanotic congenital 45 Joffe H, Georgakopoulos D, Celermajer DS resources/publications/en/. Accessed 6 Sep- heart disease. Br Heart J 1979; 41: 23–27. et al. Late ventricular arrhythmia is rare tember, 2010. 75 Barragry TP, Blatchford JW, Tuna IC et al. after early repair of tetralogy of Fallot. J 60 Simonneau G, Galie N, Rubin LJ et al. Left ventricular dysfunction in a canine Am Coll Cardiol 1994; 23: 1146–1150. Clinical classification of pulmonary hyper- model of chronic cyanosis. Surgery 1987; 46 Flinn CJ, Wolff GS, Dick M et al. Cardiac tension. J Am Coll Cardiol 2004; 43: 5S– 102: 362–370. rhythm after the Mustard operation for 12S. 76 Graham TP Jr, Erath HG Jr, Buckspan GS complete transposition of the great arteries. 61 Carmosino MJ, Friesen RH, Doran A et al. et al. Myocardial anaerobic metabolism dur- N Engl J Med 1984; 310: 1635–1638. Perioperative complications in children with ing isoprenaline infusion in a cyanotic ani- 47 Myridakis DJ, Ehlers KH, Engle MA. Late pulmonary hypertension undergoing noncar- mal model: possible cause of myocardial follow-up after venous switch operation diac surgery or cardiac catheterization. dysfunction in cyanotic congential heart dis- () for simple and com- Anesth Analg 2007; 104: 521–527. ease. Cardiovasc Res 1979; 13: 401–406. plex transposition of the great arteries. Am 62 Bancalari E, Jesse MJ, Gelband H et al. 77 Fletcher R. The relationship between the J Cardiol 1994; 74: 1030–1036. Lung mechanics in congenital heart arterial to end-tidal PCO2 difference and 48 Merlo M, de Tommasi SM, Brunelli F et al. disease with increased and decreased pul- hemoglobin saturation in patients with con- Long-term results after atrial correction of monary blood flow. J Pediatr 1977; 90: 192– genital heart disease. Anesthesiology 1991; complete transposition of the great arteries. 195. 75: 210–216. Ann Thorac Surg 1991; 51: 227–231. 63 Howlett G. Lung mechanics in normal 78 Blesa MI, Lahiri S, Rashkind WJ et al. 49 Kumar AE, Fyler DC, Miettinen OS et al. infants and infants with congenital heart Normalization of the blunted ventilatory Ebstein’s anomaly. Clinical profile and nat- disease. Arch Dis Child 1972; 47: 707–715. response to acute hypoxia in congenital cya- ural history. Am J Cardiol 1971; 28: 84–95. 64 Clapp S, Perry BL, Farooki ZQ et al. notic heart disease. N Engl J Med 1977; 50 Paranon S, Acar P. Ebstein’s anomaly of Down’s syndrome, complete atrioventricular 296: 237–241. the tricuspid valve: from fetus to adult: con- canal, and pulmonary vascular obstructive 79 White MC. Anaesthetic implications of con- genital heart disease. Heart 2008; 94: 237– disease. J Thorac Cardiovasc Surg 1990; genital heart disease for children undergoing 243. 100: 115–121. non-cardiac surgery. Anaesth Intensive Care 51 Moodie DS, Ritter DG, Tajik AH et al. 65 Rudolph AM, Yuan S. Response of the pul- Med 2009; 10: 504–509. Long-term follow-up after palliative opera- monary vasculature to hypoxia and H+ ion 80 Hastings LA, Wood JC, Harris B et al. Car- tion for univentricular heart. Am J Cardiol concentration changes. J Clin Invest 1966; diac medications are not associated with 1984; 53: 1648–1651. 45: 399–411. clinically important preoperative electrolyte 52 Moodie DS, Ritter DG, Tajik AJ et al. 66 Morray JP, Lynn AM, Mansfield PB. Effect disturbances in children presenting for car- Long-term follow-up in the unoperated uni- of pH and PCO2 on pulmonary and sys- diac surgery. Anesth Analg 2008; 107: 1840– ventricular heart. Am J Cardiol 1984; 53: temic hemodynamics after surgery in chil- 1847. 1124–1128. dren with congenital heart disease and 81 Colson P, Ryckwaert F, Coriat P. Renin 53 Deanfield JE, Ho SY, Anderson RH et al. pulmonary hypertension. J Pediatr 1988; angiotensin system antagonists and anesthe- Late sudden death after repair of tetralogy 113: 474–479. sia. Anesth Analg 1999; 89: 1143–1155.

528 Pediatric Anesthesia 21 (2011) 522–529 ª 2010 Blackwell Publishing Ltd M.C. White Managing children for noncardiac surgery

82 Ryckwaert F, Colson P. Hemodynamic eral anesthesia. Anesth Analg 2005; 101: lines from the American Heart Association: effects of anesthesia in patients 622–628. table. a guideline from the American Heart Asso- with ischemic heart failure chronically 85 Pigott DW, Nagle C, Allman K et al. Effect ciation Rheumatic Fever, Endocarditis and treated with angiotensin-converting of omitting regular ACE inhibitor medication Kawasaki Disease Committee, Council on enzyme inhibitors. Anesth Analg 1997; 84: before cardiac surgery on haemodynamic Cardiovascular Disease in the Young, and 945–949. variables and vasoactive drug requirements. the Council on Clinical Cardiology, Council 83 Comfere T, Sprung J, Kumar MM et al. Br J Anaesth 1999; 83: 715–720. on Cardiovascular Surgery and Anesthesia, Angiotensin system inhibitors in a general 86 NICE Guideline CG64, 2008. Prophylaxis and the Quality of Care and Outcomes surgical population. Anesth Analg 2005; 100: against infective endocarditis. http://www.ni- Research Interdisciplinary Working Group. 636–644. table. ce.org.uk/Guidance/CG64. J Am Dent Assoc 2007; 138: 739–760. 84 Reich DL, Hossain S, Krol M et al. Predic- 87 Wilson W, Taubert KA, Gewitz M et al. tors of hypotension after induction of gen- Prevention of infective endocarditis: guide-

Pediatric Anesthesia 21 (2011) 522–529 ª 2010 Blackwell Publishing Ltd 529 Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Treatment and monitoring of coagulation abnormalities in children undergoing heart surgery Philip Arnold

Jackson Rees Department of Paediatric Anaesthesia, Alder Hey Children’s Hospital, Liverpool, UK

Keywords Summary blood transfusion; hematology; coagulation; massive transfusions; congenital heart Bleeding is a considerable clinical problem during and after pediatric heart disease; antifibrinolytic drugs; cardiac surgery. While the primary cause of bleeding is surgical trauma, its treat- surgery; thromboelastogram ment is often complicated by the presence of coagulopathy. The principle causes of coagulopathy are discussed to provide a context for treatment. Correspondence The role of laboratory and point of care tests, which aim to identify the Dr. Philip Arnold, cause of bleeding in the individual patient, is also discussed. An attempt is Consultant in Paediatric Cardiac Anaesthesia, Jackson Rees Department of made to examine the current evidence for available therapies, including use Paediatric Anaesthesia, Alder Hey Children’s of blood products and, more recently proposed, approaches based on Hospital, Eaton Road, Liverpool L12 2AP, human or recombinant factor concentrates. UK Email: [email protected]

Section Editor: Chandra Ramamoorthy

Accepted 23 October 2010 doi:10.1111/j.1460-9592.2010.03461.x

Introduction Models of coagulation Bleeding remains a major clinical problem during and Traditional models of coagulation have presented coa- after pediatric heart surgery. Young age, low weight, gulation as a linear cascade of proteolytic enzymes. use of deep hypothermia, high preoperative hemato- The implication has been that each step is necessary crit, and complex surgery are risk factors for severe and of equal importance. More recent descriptive mod- bleeding (1–3). Bleeding occurs because of surgical els of coagulation have stressed the importance of pla- trauma leading to opening of blood vessels; all bleed- telet and cellular elements (such as white cells and ing is ‘surgical bleeding’. However, achieving accepta- endothelium) and have recognized that coagulation ble hemostasis is often complicated by the presence of occurs not as a linear process but as a complex web of coagulopathy of variable severity. All patients will interactions between procoagulant factors, inhibitors, have coagulation abnormalities if sensitive tests are and the fibrinolytic system (Figure 1) (4). This distinc- applied and most patients will have coagulation tion has clinical relevance in explaining the observa- abnormalities demonstrated by routine laboratory and tions that functionally normal coagulation can exist point of care tests. Optimal care demands cooperation despite low levels of clotting proteins (in neonates and between the anesthetist and surgeon to apply therapy moderate liver dysfunction) (5), that deficits in one which will give an acceptable level of hemostasis with- part of coagulation can be compensated for by other out additional risk of injury, either from bleeding or parts of coagulation (use of fibrinogen to correct for from adverse effects of therapy aimed at reducing thrombocytopenia (6) or use of activated factor VIIa bleeding. Therapy should be aimed neither at ‘perfect’ in hemophilia), and questions some approaches to correction of coagulation abnormalities or unrealistic acquired bleeding (use of laboratory coagulation tests surgical hemostasis. (5) and some use of fresh frozen plasma). A difficulty

494 Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd P. Arnold Treatment and monitoring of coagulation abnormalities

Activated Blood vessel platelets Collagen VIIa Platelets vWF a MA -angle A 30 min 30 Clot + Fibrin Amplification R time X Factors: V, X, Lys 30 = MA/A30 (%) Thrombin VIIa TF Xa VIII, IX, XI Fibroblast Ca Figure 2 A normal thromboelastogram (TEG) trace is shown. The output of TEG is graphical; however, a number of numerical values PC aPC Phospholipid can be derived to describe the curve. R-time is the time from TM beginning the sample until the amplitude of the trace reaches Endothelium 1 mm; it describes the initiation of clotting. Maximum amplitude Inhibition (MA) is the largest amplitude reached; it describes the strength of the clot formed. The a angle describes the steepness of the initial Figure 1 This figure demonstrates the major pathways involved in part of the curve; this is often referred to as the ‘clot kinetics’. Lys normal coagulation. Coagulation is initiated by leakage of factor VIIa 30 is the ratio of the amplitude of the trace 30 min after the maxi- and platelets from traumatized blood vessels. When platelets come mum amplitude to the maximum amplitude; this describes the sta- into contact with von Willibrands factor (VWF) and collagen, they bility of the clot. undergo a process of activation, which includes the release of procoagulants, a dramatic increase in surface area and expression of adhesion molecules (for other platelets and fibrin). Factor VIIa, Coagulopathy and cardiac bypass usually present in low concentrations within blood vessels, makes contact with tissue factor (TF) bound to cells (principally fibroblasts In the absence of anticoagulants, blood will clot in but subsequently white cells). This leads to the production of a extracorporeal circuits. This has been ascribed to the small quantity of thrombin, which initiates a positive feedback loop activation of the ‘intrinsic’ pathway involving prekal- (involving a number of factors, further conversion of factor VII to likrein and factor XII upon contact with the artificial VIIa and activation of platelets) producing large quantities of throm- surface. However, patients deficient in factor XII or bin. This leads to the production of fibrin (from fibrinogen), which prekallikrein clot at the same rate as other patients, together with activated platelets forms clot. The major inhibitor is and the ‘intrinsic’ pathway appears to play little role the Protein C (PC) system, which inactivates factor VIII and V. In the absence of endothelium bound thrombomodulin (TM) (for either in normal coagulation or in clotting inside extra- example in invitro blood tests), little active protein C is produced. corporeal circuits (7,8). The lining of artificial circuits Subsequent stabilization of the clot by factor XIII, inhibitors other differs from endothelium in two respects: proteins will than protein C, and the fibrinolytic system are not shown for bind freely to their surface and they lack any inhibi- simplicity. tory effect on coagulation. Binding of fibrinogen (together with fibronectin and von Willibrand factor) with this descriptive model of coagulation is its com- leads to adherence of white cells, red cells, and plate- plexity. Complex systems may perform in unexpected lets (7). The adherent platelets undergo activation ways, study of parts of such systems in isolation may encouraging further adhesion, releasing procoagulants, be misleading and such a complex model is of little and presenting a large phospholipid surface upon practical use when managing a bleeding patient. A which further activation occurs. Adherent monocytes simpler, more practical approach to coagulation is illu- express tissue factor allowing coagulation to proceed strated by the thromboelastogram (TEG) (Figure 2). as described previously. While heparin dramatically Coagulation is broken down into three stages repre- reduces thrombin formation and the formation of senting initiation, clot strength, and clot stability. fibrin clots, it is an imperfect anticoagulant and does Therapy can then be aimed specifically at each of these not prevent initial protein binding or activation of coa- stages when abnormalities exist. Ideally, the ability of gulation or platelets. This leads to defects in platelet these interventions to stop bleeding can be studied in numbers, in function of platelets, in fibrinogen concen- properly conducted prospective trials in specific patient tration, and in the activation of fibrinolytic systems. populations. In reality, few such studies have been per- Activation of fibrinolysis will lead to dissolution of formed and much of our therapy is based on much clots, further reduction in platelet function, and weaker evidence from laboratory data, observational increased concentrations of fibrin degradation pro- studies, and expert opinion. ducts, which inhibit coagulation further (9).

Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd 495 Treatment and monitoring of coagulation abnormalities P. Arnold

As well as the effects of interaction between blood with moderate hepatic disease (5). In these patients, and extracorporeal surfaces, other factors will add to the long PT reflects low concentrations of coagulation coagulation anomalies. Blood spilt into the pericardial factors which, in vivo, are balanced by low concentra- space and returned via ‘pump’ is a further tions of inhibitors. A long PT is also common in the (and possibly more important) cause of activation (10). absence of abnormal bleeding after pediatric heart sur- Clotting factors and platelets will undergo dilution on gery and this may reflect a similar situation. Secondly, initiating bypass (11,12), red cells and platelets may be deficiency of coagulation factors may not be the pri- damaged from suction, bleeding itself may add to coa- mary cause of bleeding. Congenital hemophilias are gulation abnormalities, extensive surgical dissection represented by profound deficiency of a single clotting will add to bleeding, and many patients will have pre- factor. Acquired coagulopathy after surgery represents existing coagulation abnormalities because of cyanotic multiple coagulation defects with multiple origins. Not heart disease (13–15), critical illness, or medications. all these defects will be of equal importance, and cor- rection will not have an equal impact on bleeding. After cardiac surgery, deficiencies of fibrinogen and Monitoring of coagulation functional platelets are likely to make a greater contri- Tests of coagulation are performed with the aim of bution to bleeding than deficiency of other factors identifying the coagulation abnormalities most likely (3,16,17). Basing treatment on PT may not lead to to be contributing to bleeding. If results from these optimal correction of bleeding. tests can be reported to the clinician in a short time, Thrombin time reflects fibrinogen activity but it is therapy can be directed more effectively at the specific also prolonged by inhibition because of fibrinogen cause of bleeding, leading to both more rapid correc- degradation products (FDP) or heparin. The most tion of coagulopathy and avoiding unnecessary ther- common method for the determination of fibrinogen apy. Tests can be divided between those conducted concentration, the Clauss method, is derived from the primarily in hematology laboratories and those avail- thrombin time. It is not interpretable in the presence able at the patient’s bedside (point of care devices). of heparin and will be inaccurate in the presence of The objective in using point of care tests is to make FDPs (common after heart surgery) (18). the results available to clinicians more rapidly; though, as the tests available are also different, they will have Point of care monitoring their own advantages and limitations. Point of care monitoring of coagulation may be aimed at the control of heparin administration or at the iden- Laboratory-based coagulation tests tification of coagulopathy. Devices available for moni- The most commonly performed laboratory tests rele- toring of heparin have been reviewed previously vant to coagulation are the full blood count (to give (19,20). For the purpose of this article, only the latter platelet concentration) and clotting studies [prothrom- will be discussed. Over the last few years, improvement bin time (PT), activated partial thromboplastin time in design has allowed a number of devices, previously (APTT), thrombin time, and fibrinogen concentration]. suitable only for use in laboratories, to be developed Clotting studies are performed using patient’s plasma for point of care use. Devices with potential for use and so are insensitive to effects owing to abnormalities during cardiac surgery include point of care equiva- of platelets or cellular elements. Citrated plasma is lents of laboratory tests (such as APTT and PT), mixed with an excess of calcium and a reagent. Use of thromboelastography (TEG and ROTEM), and speci- different reagents aims to test different parts of the fic tests of platelet function (Sonoclot analyzer, PFA- coagulation pathway. The two most important tests 100 and MultiPlate platelet aggregometer) (21). Cur- used in acquired coagulopathy are the PT and fibrino- rently, only variations of thromboelastography have gen concentration. PT is widely used for monitoring been widely employed and will be the focus of this dis- acquired coagulopathy in a number of situations, cussion (22,23). The differences between the TEG and including postcardiac surgery. The reagent used, rotational thromboelastography (ROTEM) have been thromboplastin, contains tissue factor and phospholi- recently reviewed (24). While important differences pids; activation therefore occurs by the normal ‘tissue exist and results from these devices are not inter- factor’ pathway. Problems arise in the interpretation of changeable (25), the principle of their clinical applica- this test for two reasons. First, PT is prolonged in a tion is similar. The discussion will focus on TEG while number of situations despite functionally normal coa- important differences in clinical application will be gulation, including healthy neonates and in patients highlighted.

496 Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd P. Arnold Treatment and monitoring of coagulation abnormalities

In conventional TEG, a pin is lowered into a cup any effect of heparin, and comparison to an unmodi- containing whole blood. The cup is rotated in each fied TEG will identify patients in which reversal of direction approximately 4. As the blood begins to heparin has been inadequate. An estimate of fibrinogen clot, the pin will begin to move with the cup, the activity can be made by the addition of blockers of force of the movement being a marker of the strength platelet function; the resulting clot will, therefore, of the clot at that moment in time. This force is used reflect the activity of fibrinogen alone. to construct a graph of clot strength over time. A typical (Kaolin activated) TEG is shown in Figure 1. Integration of monitoring into clinical care Initially, a straight line indicates no clot being present (the pin does not move), subsequently, a clot begins A rational approach to bleeding after heart surgery to form and moves the pin to reach a maximum should include a clinical assessment of bleeding at that amplitude taken to indicate maximum clot strength. point in time (30,31), an appreciation of risk factors, The stability of the clot is then indicated by the and use of specific tests to identify the presence and degree to which the clot retains this strength. An acti- mechanisms of coagulopathy. In the absence of clinical vator is usually added to the cup to speedup analysis, bleeding, therapy is not required even if specific tests and other substances may also be added to modify demonstrate an abnormality. In the presence of bleed- the test. Kaolin is the most common activator used ing, tests can help to differentiate the specific cause. with conventional TEG, while the ROTEM is com- The most common causes of bleeding immediately monly performed as a battery of tests including kaolin after bypass are surgical causes, platelet abnormalities, (inTEM) and tissue factor (exTEM). A recent modifi- fibrinogen insufficiency, inadequate reversal of heparin, cation of TEG is ‘rapid’ TEG using a mixed activator and combinations of these factors (16,32). In the aimed at giving information on clot strength more absence of clotting abnormalities, it is more likely that rapidly (26). further surgical efforts will reduce bleeding than the The output of the TEG is primarily graphical. To administration of blood products and early recognition allow comparison of results and to develop protocols of this can lead to more appropriate treatment (33). based on TEG, it is necessary to reduce this graphical However, in small children, this is an uncommon situa- output to numerical values (Figure 2); computerization tion and abnormalities are likely to be present even if has allowed for rapid calculation of these values. Simi- the primary cause of bleeding is surgical. If bleeding is lar values can be calculated for ROTEM although the persistent despite the administration of appropriate values are not interchangeable and different nomencla- clotting factors and normalization of tests, then further ture is used. Population ‘normals’ have been described surgical efforts are indicated. for these values in adults, children (27), and children Barriers exist to applying such an approach and it with congenital heart disease (28). Some care is may not always be appropriate. Time is required to required in the use of such reference values, as these perform tests and to initiate treatment (in particular, will be different if different activators and additives are use of blood products) and delaying treatment, while used. In the interpretation of unmodified TEG values, awaiting diagnostic information may be counterpro- a prolonged R-time is an indication of residual antic- ductive. This can be minimized by the use of tests that oagulant or deficiency of clotting factors, a low maxi- will give rapid results (26) or by performing tests dur- mum amplitude indicates poor clot strength, and ing bypass, which will predict coagulopathy after increased degree of lysis (high Lys 30 or Lys 60) indi- bypass (17). Inadequate reversal of heparin is an easily cates the presence of fibrinolysis. Low amplitude (MA) reversible cause of bleeding, which can be rapidly con- is the TEG value most closely correlated with bleeding firmed by point of care tests such as heparinase-modi- after heart surgery (16,23). A low MA may be the fied TEG. Some patients, such as neonates undergoing result of low platelet number, poor platelet function, complex procedures, are at very high risk of bleeding. low fibrinogen, poor function of fibrinogen, or a com- If excessive bleeding is present, after reversal of bination of these factors. Moganasundram et al. (16) heparin, initial empirical transfusion of platelets or demonstrated a close correlation between TEG MA cryoprecipitate is indicated (3,30) and waiting for coa- and the product of platelet count and fibrinogen con- gulation tests may only lead to delay in treatment. If centration in children. Similar findings of an interac- the patient fails to respond to initial treatment, the tion between platelets and fibrinogen have been cause for ongoing bleeding is likely to be less clear and confirmed in adults (29). coagulation tests can help to guide further therapy. Adaptation of the TEG can add further informa- Figure 3 demonstrates the ability of TEG to follow tion. Addition of heparinase to the cup will neutralize changes in coagulation during a case.

Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd 497 Treatment and monitoring of coagulation abnormalities P. Arnold

(a) (b)

(c) (d)

Figure 3 The ability of thromboelastogram (TEG) to follow sive bleeding. Bleeding continued and a TEG was performed to changes in coagulation through a case is demonstrated. Dotted identify the cause of bleeding. This demonstrated a low amplitude lines show a ‘normal’ trace. The patient was a newborn who had and excess fibrinolysis (b). Tranexamic acid was given with some undergone an aortic arch repair. The baseline trace (a) demon- effect (c) followed by the infusion of cryoprecipitate (d). Following strates a mildly elongated R-time but is otherwise normal. After this, bleeding reduced dramatically and the TEG returned to the bypass platelets were administered empirically because of exces- baseline.

In other situations and when bleeding is of lesser appropriate for an elderly adult with liver disease or a severity, greater consideration can be given to the need neonate undergoing heart surgery. Algorithms based for treatment. Approaches to bleeding based on clini- on point of care monitoring have also been introduced cal assessment of bleeding and results of coagulation (Table 2). These have been successful in limiting unne- tests can avoid unnecessary transfusion. Transfusion cessary transfusion in adults (37–39) and children (32). algorithms have been introduced to this end. Tradi- The algorithms introduced have largely been extrapo- tionally, these algorithms have been based on labora- lated from laboratory-based guidelines and may not tory tests (Table 1) and have been developed primarily fully exploit the advantages of tests such as thromboe- through clinical consensus (34–36). Often the clinical lastograms. Table 2 gives advice on TEG-guided trans- evidence (for the chosen indication, the specific pro- fusion. Evidence that the introduction of transfusion duct, or the dose) is weak. In addition, the guidelines algorithms based on point of care monitoring can are frequently applied to highly heterogeneous groups improve outcome other than reducing unnecessary of patients; it is unlikely that the same management is transfusion is limited (40).

Table 1 Guidelines for the transfusion of blood products. Following Specific agents bypass, it is assumed that significant platelet dysfunction will be Blood products present and the higher target for platelet transfusion (100 · 109)is Platelet dysfunction and thrombocytopenia are com- used mon after cardiac bypass in infants. Hence, in the pre- Indication Product and dose sence of bleeding, transfusion of platelets is logical.

) Coagulation variables that relate to platelet number PT >1.5 times control FFP 10–15 mlÆkg 1 APTT >2.0 times control FFP 10–15 mlÆkg)1 and function (such as TEG MA) are corrected by the Platelet count Platelet concentrate 10 mlsÆkg)1* infusion of platelets, and clinical experience indicates <50 · 109–100 · 109 that platelet transfusion is highly efficacious at redu- Fibrinogen <0.8–1.0 gÆl)1 Cryoprecipitate 1 unit per 10 kg cing bleeding. This has been confirmed in a small study of children postcardiac surgery (3). A platelet count of PT, prothrombin time; FFP, fresh frozen plasma; APTT, activated 9 )1 partial thromboplastin time. <108 · 10 l have been identified as a predictor of *Platelet products will vary between institutions and local guide- bleeding, while clot strength measured by TEG reduces ) lines on dosage should be developed. steeply at platelet counts below 120 · 109 l 1. The cur-

498 Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd P. Arnold Treatment and monitoring of coagulation abnormalities

Table 2 Commonly measured TEG parameters, with guidance on their significance and suggested action. Thresholds for transfusion or other treatments should be seen only as a guide, and treatment should be guided by the clinical situation and known risk factors for coagul- opathy. Treatment is only indicated in the presence of active bleeding

TEG parameter Additive ‘Normal’ value Interpretation Suggested action

MA (kaolin) Kaolin 54–72 mm Reflects platelet number, >48 mm platelets or cryoprecipitates platelet function, fibrinogen are not indicated concentration and <43 and if patient is bleeding, there is dysfunction likely benefit from platelet or cryoprecipitate infusion R-time (kaolin) Kaolin 4–8 min Similar to ACT is a complex If long, consider residual heparin parameter affected by (check heparinase TEG) or factor anticoagulation, factor deficiency deficiency, and presence of FFP may be indicated if >15 min but inhibitors. Also operator other abnormalities should be corrected variable first If severe bleeding and long R-time, consider use of Prothrombin Complex Concentrates Lys 30/Lys 60 Kaolin 0–8% High-level diagnostic of Treat with antifibrinolytic drugs 0–15% abnormal fibrinolysis Heparinase-modified Kaolin As R-time Useful in heparinized patient. If unmodified R-time is >1.5 times R-time Heparinase Compare to plain TEG MA on bypass shows heparinase-modified R-time, further kaolin R-time correlation with bleeding protamine should be given post bypass If ran together with unmodified TEG, a shorter R-time indicates residual heparin or contamination Functional Tissue factor Expressed as a Identifies portion of clot ‘Traditional’ management is to Fibrinogen Platelet blocker fibrinogen strength attributable to supplement fibrinogen until >1 gÆl)1 (ReoPro) concentration fibrinogen to give estimate in the presence of bleeding. If active, of fibrinogen activity bleeding likely to be further benefit in targeting concentration 2 gÆl)1 or higher. Can be achieved with cryoprecipitate or fibrinogen concentrate

TEG, thromboelastogram; FFP, fresh frozen plasma. rent recommendation of targeting a platelet count of to demonstrate benefit, although the trials examined ) approximately 100 · 109 l 1 appears reasonable. Use were small and conducted in heterogeneous populations of platelets for initial treatment of presumed coagulo- (41). In a small observational study, a number of pathic bleeding (in the absence of specific clotting patients had coagulopathic bleeding after transfusion of tests) also seems reasonable. platelets; if these patients were then given FFP, bleed- The use of fresh frozen plasma (FFP) is based on the ing increased, while if cryoprecipitate was given, bleed- observation that concentration of clotting factors is ing improved (3). The lack of efficacy may be related to often low immediately post bypass. The case for use of dose. FFP is frequently given in doses below the recom- FFP either empirically or guided by coagulation tests is mended dose of 10–15 mlÆkg)1 (42); though, this dose considerably weaker than for platelets. PT, APTT, and may itself be inadequate to restore factor concentra- TEG-R-time are clotting tests that correlate best with tion. Modeling of FFP administration in major trauma factor deficiency. A PT >1.5 times control is taken as suggests that much larger doses may be required to an indication that FFP should be transfused (36). How- boast or retain factor levels (43). It is possible that a ever, these tests are often significantly prolonged in the dose of 30–40 mlÆkg)1 may be more appropriate; absence of bleeding and, when taken post bypass, cor- though, administration of such a large volume is unli- relate poorly with excessive bleeding. A meta-analysis kely to be desirable except in the context of massive of the use of FFP to treat acquired coagulopathy failed ongoing bleeding. Dilution of platelets and red cells,

Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd 499 Treatment and monitoring of coagulation abnormalities P. Arnold due to administration of large volumes of FFP, may in safe alternative to cryoprecipitate (51–53). Use in car- fact exacerbate bleeding. Addition of FFP to the bypass diac surgery in children and adults has been increasing circuit may allow the administration of much larger (54,55). It is expected that this agent will be available doses, however benefit appears to be confined to smal- in the United States and United Kingdom in 2010. ler patients (44,45). Unlike cryoprecipitate, it contains only fibrinogen, and Not all coagulation factors are of equal importance a direct comparison of these two agents in acquired during bleeding. Fibrinogen is normally present in coagulopathy has not been made. Potentially, it allows much higher concentrations than other clotting factors for more rapid correction of fibrinogen concentration and while other factors are mainly involved in initiat- as it will not require defrosting and can be adminis- ing or amplifying thrombin formation, fibrinogen is a tered in a smaller volume. Combining the availability substrate for the production of fibrin. Deficiencies of of this agent with rapid measurement of fibrinogen fibrinogen are reflected in reduced strength of clot and concentration (using modified TEG) is an attractive are associated with increased bleeding (16). Cryopreci- approach (55–58). A small randomized study compar- pitate is the most concentrated fibrinogen replacement ing prophylactic administration of fibrinogen concen- widely available with a fibrinogen concentration 4–8 trate to placebo in adult cardiac patients demonstrated times that of FFP; though, a unit of cryoprecipitate a reduction in bleeding (59). Recombinant fibrinogen contains less fibrinogen than a unit of FFP, and multi- is currently under development but it will be several ple units will be required in larger patients. One unit years before it is available. for each 10 kg should raise fibrinogen concentration The ease of administration and low volume of fibrino- by 0.5–1.0 gÆl)1. As well as fibrinogen, cryoprecipitate gen concentrates compared to FFP or cryoprecipitate is a source of Von Willibrands factor, Factors VIII can allow targeting of higher fibrinogen concentrations and XIII, although the clinical significance of this is than have been used previously (57,60). It is possible unclear. Cryoprecipitate has been effective in treating that such use could reduce both bleeding and the need bleeding resistant to platelet concentrates alone during for platelet transfusion. In TEG studies, high concentra- pediatric heart surgery (3). tions of fibrinogen can restore clot strength despite low platelet numbers. In animal models of bleeding and thrombocytopenia, fibrinogen concentrates were super- Human-fractionated factor concentrates ior to stored platelet concentrates in resolving bleeding Human-fractionated factor concentrates are produced (6). Such an approach requires further evaluation from pooled human plasma. They undergo pasteuriza- including the assessment of effectiveness, safety, and tion, which will neutralize blood-borne viruses but identification of appropriate targets. (potentially) not prions. They are presented as a dried powder requiring reconstitution. They do not require Recombinant factor concentrates cross-matching and have a long shelf life avoiding some of the logistical problems with the supply of Use of factor VIIa in pediatric heart patients has been other blood products. previously reviewed (61). Numerous case reports have Prothrombin complex concentrates (PCCs) are mix- shown effectiveness in treating severe bleeding in a tures of inactive factor II (prothrombin), VII, IX, and number of situations including pediatric heart surgery. X plus the coagulation inhibitors Protein S and C. A Significant questions concerning the indication for this number of different brands of PCC are available, agent, dose, and safety remain. Prophylactic use in a which have different relative concentrations of these relatively low dose in children was ineffective in redu- factors. In some older products, low concentration of cing bleeding (62). In adults undergoing heart surgery, Protein S and C has been associated with increased rVIIa was effective in reducing bleeding, although the thrombogenicity, while low concentration of factor VII rate of serious adverse effects was higher than placebo has been associated with a lack of effect (46). PCCs (63). Because of concern over thrombotic complica- are currently considered safe and reliable for urgent tions and ineffectiveness in preventing bleeding, it reversal of oral anticoagulants (47,48). The use of these would be reasonable to confine use to treatment of ser- agents in other causes of acquired coagulopathy is less ious life threatening bleeding. certain although case series in adult patients have been Human and recombinant factor XIII has been encouraging (49,50). Widespread use in children after described for use during heart surgery (64–66) and is heart surgery requires more evaluation. supported by animal work (67). Factor XIII functions Fibrinogen concentrate has been available in conti- by stabilizing and strengthening bonds in the fibrin nental Europe for over 20 years and appears to be a polymer and would, therefore, be expected to contri-

500 Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd P. Arnold Treatment and monitoring of coagulation abnormalities bute to clot strength and stability. Association between Conclusion factor XIII concentration and bleeding have been Coagulopathy is a common cause of excess bleeding inconsistent (68,69). Factor XIII also has a more gen- after heart surgery. Treatment should aim to reduce eral role in healing and may reduce myocardial edema bleeding and should only be initiated in the presence and pleural effusions in children after heart surgery of bleeding. ‘Traditional’ treatment using blood pro- (70,71). The role of factor XIII replacement is uncer- ducts is largely successful, and empirical use of these tain; though, the development of a recombinant con- products is often appropriate, particularly in the initial centrate of this factor may lead to increased interest. treatment of severe bleeding and in high-risk patients. A strategy based on coagulation testing and applica- Other drugs tion of specific therapies is likely to be more successful in patients who fail to respond to initial empirical Treatment with antifibrinolytics such as aprotinin, tra- treatment. Alternatively, in lower-risk patients, the nexamic acid, or epsilon aminocaproic acid is common empirical use of blood products is likely to lead to during pediatric heart surgery. Use of aprotinin in chil- excessive transfusion, and an approach based on a dren has reduced dramatically because of safety con- clinical assessment of bleeding, an appreciation of risk cerns in adults (72–75), despite retrospective studies that factors, and use of tests is likely to reduce transfusion. show a lack of serious toxicity in children (76). Aproti- Points of care tests of coagulation, in particular the nin is no longer available in many countries. Tranexa- TEG, have theoretical and practical advantages over mic acid and epsilon aminocaproic acid have evidence laboratory-based tests. of efficacy in children although probably less than apro- Management of bleeding after pediatric heart sur- tinin (77–79). Only a single study has addressed dosing gery using strategies based on infusion of more specific of tranexamic acid in children (80) and dosing remains factor concentrates and fibrinogen is promising. As extremely variable (abstract APA 2010). yet, these approaches are largely untested either for Desmopressin increases plasma levels of von Willi- efficacy or safety. Trials of therapy aimed at treatment brand factor with potential for reducing bleeding. In of bleeding are challenging; though, these problems are adults undergoing cardiac surgery, any benefit is con- not insurmountable and the benefits from such trials fined to patients with high anticipated blood loss (81). will be considerable. Trials of desmopression in children and adults under- going surgery for congenital heart disease have failed to show a beneficial effect (82–84).

References 1 Szekely A, Cserep Z, Sapi E et al. Risks and thrombocytopenia. J Thromb Haemost 2007; 12 Chan AK, Leaker M, Burrows FA et al. predictors of blood transfusion in pediatric 5(5): 1019–1025. Coagulation and fibrinolytic profile of pae- patients undergoing open heart operations. 7 Paparella D, Brister SJ, Buchanan MR. diatric patients undergoing cardiopulmonary Ann Thorac Surg 2009; 87(1): 187–197. Coagulation disorders of cardiopulmonary bypass. ThrombHaemost 1997; 77(2): 270–277. 2 Williams GD, Bratton SL, Ramamoorthy bypass: a review. Intensive Care Med 2004; 13 Osthaus WA, Boethig D, Johanning K et al. C. Factors associated with blood loss and 30(10): 1873–1881. Whole blood coagulation measured by mod- blood product transfusions: a multivariate 8 Burman JF, Chung HI, Lane DA et al. ified thrombelastography (ROTEM) is analysis in children after open-heart surgery. Role of factor XII in thrombin generation impaired in infants with congenital heart Anesth Analg 1999; 89(1): 57–64. and fibrinolysis during cardiopulmonary diseases. Blood Coagul Fibrinolysis 2008; 3 Miller BE, Mochizuki T, Levy JH et al. Pre- bypass. Lancet 1994; 8931: 1192–1193. 19(3): 220–225. dicting and treating coagulopathies after 9 Williams GD, Bratton SL, Nielsen NJ et al. 14 Tempe DK, Virmani S. Coagulation cardiopulmonary bypass in children. Anesth Fibrinolysis in pediatric patients undergoing abnormalities in patients with cyanotic con- Analg 1997; 85(6): 1196–1202. cardiopulmonary bypass. J Cardiothorac genital heart disease. J Cardiothorac Vasc 4 Hoffman M. A cell-based model of coagula- Vasc Anesth 1998; 12(6): 633–638. Anesth 2002; 16(6): 752–765. tion and the role of factor VIIa. Blood Rev 10 Chung JH, Gikakis N, Rao AK et al. Peri- 15 Guay J, Rivard GE. Mediastinal bleeding 2003; 17(Suppl 1): S1–S5. cardial blood activates the extrinsic coagula- after cardiopulmonary bypass in pediatric 5 Tripodi A, Chantarangkul V, Mannucci tion pathway during clinical patients. Ann Thorac Surg 1996; 62(6): PM. Acquired coagulation disorders: revis- cardiopulmonary bypass. Circulation 1996; 1955–1960. ited using global coagulation/anticoagula- 93(11): 2014–2018. 16 Moganasundram S, Hunt BJ, Sykes K et al. tion testing. Br J Haematol 2009; 147(1): 11 Kern FH, Morana NJ, Sears JJ et al. Coa- The relationship among thromboelastogra- 77–82. gulation defects in neonates during cardio- phy, hemostatic variables, and bleeding after 6 Velik-Salchner C, Haas T, Innerhofer P et pulmonary bypass. Ann Thorac Surg 1992; cardiopulmonary bypass surgery in children. al. The effect of fibrinogen concentrate on 54(3): 541–546. Anesth Analg 2010; 110(4): 995–1002.

Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd 501 Treatment and monitoring of coagulation abnormalities P. Arnold

17 Williams GD, Bratton SL, Riley EC et al. 30 McEwan A. Aspects of bleeding after car- 43 Alam HB, Bice LM, Butt MU et al. Testing Coagulation tests during cardiopulmonary diac surgery in children. Pediatr Anesth of blood products in a polytrauma model: bypass correlate with blood loss in children 2007; 17(12): 1126–1133. results of a multi-institutional randomized undergoing cardiac surgery. J Cardiothorac 31 Nuttall GA, Oliver WC, Ereth MH et al. preclinical trial. J Trauma 2009; 67(4): 856– Vasc Anesth 1999; 13(4): 398–404. Coagulation tests predict bleeding after car- 864. 18 Nieuwenhuizen W. Biochemistry and mea- diopulmonary bypass. J Cardiothorac Vasc 44 Oliver WC Jr, Beynen FM, Nuttall GA surement of fibrinogen. Eur Heart J 1995; Anesth 1997; 11(7): 815–823. et al. Blood loss in infants and children for 16(Suppl A): 6–10; discussion. 32 Romlin SHWea. Routine intraoperative open heart operations: albumin 5% versus 19 Gruenwald CE, Manlhiot C, Crawford- thromboelastography reduces blood transfu- fresh-frozen plasma in the prime. Ann Lean L et al. Management and monitoring sions in paediatric cardiac surgery. J Cardi- Thorac Surg 2003; 75(5): 1506–1512. of anticoagulation for children undergoing othorac Vasc Anesth 2009; 23(3): S1. 45 McCall MM, Blackwell MM, Smyre JT cardiopulmonary bypass in cardiac surgery. 33 Davidson SJ, McGrowder D, Roughton M et al. Fresh frozen plasma in the pediatric J Extra Corpor Technol 2010; 42(1): 9–19. et al. Can ROTEM thromboelastometry pump prime: a prospective, randomized 20 Prisco D, Paniccia R. Point-of-care testing predict postoperative bleeding after cardiac trial. Ann Thorac Surg 2004; 77(3): 983–987; of hemostasis in cardiac surgery. Thromb J surgery? J Cardiothorac Vasc Anesth 2008; discussion 7. 2003; 1(1): 1. 22(5): 655–661. 46 Hellstern P. Production and composition of 21 Enriquez LJ, Shore-Lesserson L. Point-of- 34 O’Shaughnessy DF, Atterbury C, Bolton prothrombin complex concentrates: correla- care coagulation testing and transfusion Maggs P et al. Guidelines for the use of tion between composition and therapeutic algorithms. Br J Anaesth 2009; 103(Suppl fresh-frozen plasma, cryoprecipitate and efficiency. Thromb Res 1999; 95(4 Suppl 1): 1): i14–i22. cryosupernatant. Br J Haematol 2004; S7–S12. 22 Miller BE, Guzzetta NA, Tosone SR et al. 126(1): 11–28. 47 Hanley JP. Warfarin reversal. J Clin Pathol Rapid evaluation of coagulopathies after 35 Murphy MF, Brozovic B, Murphy W et al. 2004; 57(11): 1132–1139. cardiopulmonary bypass in children using Guidelines for platelet transfusions. British 48 Bershad EM, Suarez JI. Prothrombin com- modified thromboelastography. Anesth Committee for Standards in Haematology, plex concentrates for oral anticoagulant Analg 2000; 90(6): 1324–1330. Working Party of the Blood Transfusion therapy-related intracranial hemorrhage: a 23 Martin P, Horkay F, Rajah SM et al. Mon- Task Force. Transfus Med 1992; 2(4): 311– review of the literature. Neurocrit Care itoring of coagulation status using thrombe- 318. 2010; 12(3): 403–413. lastography during paediatric open heart 36 Contreras M, Ala FA, Greaves M et al. 49 Fraser TA, Corke CF, Mohajeri M et al. A surgery. Int J Clin Monit Comput 1991; 8(3): Guidelines for the use of fresh frozen retrospective audit of the use of Prothrom- 183–187. plasma. British Committee for Standards in binex-HT for refractory bleeding following 24 Jackson GN, Ashpole KJ, Yentis SM. The Haematology, Working Party of the Blood adult cardiac surgery. Crit Care Resusc TEG vs the ROTEM thromboelastography/ Transfusion Task Force. Transfus Med 2006; 8(2): 141–145. thromboelastometry systems. Anaesthesia 1992; 2(1): 57–63. 50 Bruce D, Nokes TJ. Prothrombin complex 2009; 64(2): 212–215. 37 Anderson L, Quasim I, Soutar R et al. An concentrate (Beriplex P/N) in severe bleed- 25 Venema LF, Post WJ, Hendriks HG et al. audit of red cell and blood product use after ing: experience in a large tertiary hospital. An assessment of clinical interchangeability the institution of thromboelastometry in a Crit Care 2008; 12(4): R105. of TEG (R) and RoTEM (R) thromboelas- cardiac intensive care unit. Transfus Med 51 Bevan DH. Cryoprecipitate: no longer the tographic variables in cardiac surgical 2006; 16(1): 31–39. best therapeutic choice in congenital fibrino- patients. Anesth Analg 2010; 111(2): 339– 38 Ak K, Isbir CS, Tetik S et al. Thromboelas- gen disorders? Thromb Res 2009; 124(Suppl 344. tography-based transfusion algorithm 2): S12–S16. 26 Chavez JJ, Foley DE, Snider CC et al. reduces blood product use after elective 52 Dickneite G, Pragst I, Joch C et al. Animal A novel thrombelastograph tissue factor/ CABG: a prospective randomized study. J model and clinical evidence indicating low kaolin assay of activated clotting times Card Surg 2009; 24(4): 404–410. thrombogenic potential of fibrinogen con- for monitoring heparin anticoagulation 39 Avidan MS, Alcock EL, Da Fonseca J et al. centrate (Haemocomplettan P). Blood Coa- during cardiopulmonary bypass. Anesth Comparison of structured use of routine gul Fibrinolysis 2009; 20(7): 535–540. Analg 2004; 99(5): 1290–1294; table of laboratory tests or near-patient assessment 53 Tziomalos K, Vakalopoulou S, Perifanis V contents. with clinical judgement in the management et al. Treatment of congenital fibrinogen 27 Chan KL, Summerhayes RG, Ignjatovic V of bleeding after cardiac surgery. Br J deficiency: overview and recent findings. et al. Reference values for kaolin-activated Anaesth 2004; 92(2): 178–186. Vasc Health Risk Manag 2009; 5: 843–848. thromboelastography in healthy children. 40 Schochl H, Nienaber U, Hofer G et al. Goal- 54 Weinkove R, Rangarajan S. Fibrinogen con- Anesth Analg 2007; 105(6): 1610–1613; table directed coagulation management of major centrate for acquired hypofibrinogenaemic of contents. trauma patients using thromboelastometry states. Transfus Med 2008; 18(3): 151–157. 28 Haizinger B, Gombotz H, Rehak P et al. (ROTEM)-guided administration of fibrino- 55 Solomon C, Pichlmaier U, Schoechl H et al. Activated thrombelastogram in neonates gen concentrate and prothrombin complex Recovery of fibrinogen after administration and infants with complex congenital heart concentrate. Crit Care 2010; 14(2): R55. of fibrinogen concentrate to patients with disease in comparison with healthy children. 41 Stanworth SJ, Brunskill SJ, Hyde CJ et al. severe bleeding after cardiopulmonary Br J Anaesth 2006; 97(4): 545–552. Is fresh frozen plasma clinically effective? A bypass surgery. Br J Anaesth 2010; 104(5): 29 Lang T, Johanning K, Metzler H et al. The systematic review of randomized controlled 555–562. effects of fibrinogen levels on thromboelas- trials. Br J Haematol 2004; 126(1): 139–152. 56 Brenni M, Worn M, Bruesch M et al. Suc- tometric variables in the presence of throm- 42 Kakkar N, Kaur R, Dhanoa J. Improve- cessful rotational thromboelastometry- bocytopenia. Anesth Analg 2009; 108(3): ment in fresh frozen plasma transfusion guided treatment of traumatic haemorrhage, 751–758. practice: results of an outcome audit. Trans- hyperfibrinolysis and coagulopathy. Acta fus Med 2004; 14(3): 231–235. Anaesthesiol Scand 2010; 54(1): 111–117.

502 Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd P. Arnold Treatment and monitoring of coagulation abnormalities

57 Rahe-Meyer N, Pichlmaier M, Haverich A tive bleeding after coronary surgery with 76 Guzzetta NA, Evans FM, Rosenberg ES et al. Bleeding management with fibrinogen extracorporeal circulation. Thorac Cardio- et al. The impact of aprotinin on postopera- concentrate targeting a high-normal plasma vasc Surg 2006; 54(1): 26–33. tive renal dysfunction in neonates under- fibrinogen level: a pilot study. Br J Anaesth 66 Levy JH, Gill R, Nussmeier NA et al. going cardiopulmonary bypass: a 2009; 102(6): 785–792. Repletion of factor XIII following cardio- retrospective analysis. Anesth Analg 2009; 58 Rahe-Meyer N, Solomon C, Winterhalter M pulmonary bypass using a recombinant A- 108(2): 448–455. et al. Thromboelastometry-guided adminis- subunit homodimer. A preliminary report. 77 Schouten ES, van de Pol AC, Schouten AN tration of fibrinogen concentrate for the Thromb Haemost 2009; 102(4): 765–771. et al. The effect of aprotinin, tranexamic treatment of excessive intraoperative bleed- 67 Ponce R, Armstrong K, Andrews K et al. acid, and aminocaproic acid on blood loss ing in thoracoabdominal aortic aneurysm Safety of recombinant human factor XIII in and use of blood products in major pedia- surgery. J Thorac Cardiovasc Surg 2009; a cynomolgus monkey model of extracor- tric surgery: a meta-analysis. Pediatr Crit 138(3): 694–702. poreal blood circulation. Toxicol Pathol Care Med 2009; 10(2): 182–190. 59 Karlsson M, Ternstrom L, Hyllner M et al. 2005; 33(6): 702–710. 78 Tama´s Breuer KM, Wilhelm M, Wiesner G Prophylactic fibrinogen infusion reduces 68 Blome M, Isgro F, Kiessling AH et al. Rela- et al. Aprotinin is superior to tranexamic bleeding after coronary artery bypass sur- tionship between factor XIII activity, fibri- acid in paediatric cardiac surgery. J Cardi- gery. A prospective randomised pilot study. nogen, haemostasis screening tests and othorac Vasc Anesth 2008; 22(3 Suppl): Thromb Haemost 2009; 102(1): 137–144. postoperative bleeding in cardiopulmonary O–34. 60 Fenger-Eriksen C, Ingerslev J, Sorensen B. bypass surgery. Thromb Haemost 2005; 79 Arnold DM, Fergusson DA, Chan AK Fibrinogen concentrate – a potential univer- 93(6): 1101–1107. et al. Avoiding transfusions in children sal hemostatic agent. Expert Opin Biol Ther 69 Brody JI, Pickering NJ, Fink GB. Concen- undergoing cardiac surgery: a meta-analysis 2009; 9(10): 1325–1333. trations of factor VIII-related antigen and of randomized trials of aprotinin. Anesth 61 Warren OJ, Rogers PL, Watret AL et al. factor XIII during open heart surgery. Analg 2006; 102(3): 731–737. Defining the role of recombinant activated Transfusion 1986; 26(5): 478–480. 80 Chauhan S, Bisoi A, Kumar N et al. Dose factor VII in pediatric cardiac surgery: 70 Wozniak G, Noll T, Akinturk H et al. Fac- comparison of tranexamic acid in pediatric where should we go from here? Pediatr Crit tor XIII prevents development of myocar- cardiac surgery. Asian Cardiovasc Thorac Care Med 2009; 10(5): 572–582. dial edema in children undergoing surgery Ann 2004; 12(2): 121–124. 62 Ekert H, Brizard C, Eyers R et al. Elective for congenital heart disease. Ann N Y Acad 81 Cattaneo M, Harris AS, Stromberg U et al. administration in infants of low-dose recom- Sci 2001; 936: 617–620. The effect of desmopressin on reducing binant activated factor VII (rFVIIa) in car- 71 Schroth M, Meissner U, Cesnjevar R et al. blood loss in cardiac surgery – a meta-ana- diopulmonary bypass surgery for congenital Plasmatic [corrected] factor XIII reduces lysis of double-blind, placebo-controlled heart disease does not shorten time to chest severe pleural effusion in children after trials. Thromb Haemost 1995; 74(4): 1064– closure or reduce blood loss and need for open-heart surgery. Pediatr Cardiol 2006; 1070. transfusions: a randomized, double-blind, 27(1): 56–60. 82 Seear MD, Wadsworth LD, Rogers PC parallel group, placebo-controlled study of 72 Fergusson DA, Hebert PC, Mazer CD et al. et al. The effect of desmopressin acetate rFVIIa and standard haemostatic replace- A comparison of aprotinin and lysine analo- (DDAVP) on postoperative blood loss after ment therapy versus standard haemostatic gues in high-risk cardiac surgery. N Engl J cardiac operations in children. J Thorac replacement therapy. Blood Coagul Fibrino- Med 2008; 358(22): 2319–2331. Cardiovasc Surg 1989; 98(2): 217–219. lysis 2006; 17(5): 389–395. 73 Mangano DT, Miao Y, Vuylsteke A et al. 83 Oliver WC Jr, Santrach PJ, Danielson GK 63 Gill R, Herbertson M, Vuylsteke A et al. Mortality associated with aprotinin during et al. Desmopressin does not reduce bleed- Safety and efficacy of recombinant activated 5 years following coronary artery bypass ing and transfusion requirements in congeni- factor VII: a randomized placebo-controlled graft surgery. JAMA 2007; 297(5): 471–479. tal heart operations. Ann Thorac Surg 2000; trial in the setting of bleeding after cardiac 74 Mangano DT, Tudor IC, Dietzel C. The 70(6): 1923–1930. surgery. Circulation 2009; 120(1): 21–27. risk associated with aprotinin in cardiac sur- 84 Reynolds LM, Nicolson SC, Jobes DR et al. 64 Godje O, Haushofer M, Lamm P et al. The gery. N Engl J Med 2006; 354(4): 353–365. Desmopressin does not decrease bleeding effect of factor XIII on bleeding in coronary 75 Henry D, Carless P, Fergusson D et al. The after cardiac operation in young children. surgery. Thorac Cardiovasc Surg 1998; safety of aprotinin and lysine-derived antifi- J Thorac Cardiovasc Surg 1993; 106(6): 46(5): 263–267. brinolytic drugs in cardiac surgery: a meta- 954–958. 65 Godje O, Gallmeier U, Schelian M et al. analysis. CMAJ 2009; 180(2): 183–193. Coagulation factor XIII reduces postopera-

Pediatric Anesthesia 21 (2011) 494–503 ª 2010 Blackwell Publishing Ltd 503 Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Analgesia and sedation after pediatric cardiac surgery Andrew R. Wolf & Lara Jackman

Paediatric Intensive Care Unit, Bristol Children’s Hospital, Bristol, UK

Keywords Summary sedation; analgesia; pediatric; intensive care; cardiac; withdrawal In recent years, the importance of appropriate intra-operative anesthesia and analgesia during cardiac surgery has become recognized as a factor in Correspondence postoperative recovery. This includes the early perioperative management Andrew Wolf, of the neonate undergoing radical surgery and more recently the care sur- Paediatric Intensive Care Unit, Bristol Royal rounding fast-track and ultra fast-track surgery. However, outside these Children’s Hospital, Upper Maudlin Street, areas, relatively little attention has focused on postoperative sedation and Bristol BS2 8BJ, UK Email: [email protected] analgesia within the pediatric intensive care unit (PICU). This reflects per- ceived priorities of the primary disease process over the supporting struc- Section Editor: Chandra Ramamoorthy ture of PICU, with a generic approach to sedation and analgesia that can result in additional morbidities and delayed recovery. Management of the Accepted 23 October 2010 marginal patient requires optimisation of not only cardiac and other atten- dant pathophysiology, but also every aspect of supportive care. Individua- doi:10.1111/j.1460-9592.2010.03460.x lized sedation and analgesia strategies, starting in the operating theater and continuing through to hospital discharge, need to be regarded as an impor- tant aspect of perioperative care, to speed the process of recovery.

child who may be unable to move or communicate dis- General principles tress as a result of the use of muscle relaxants, while Infants and children admitted to the intensive care unit the unparalyzed child may ‘fight’ the ventilator leading (ICU) require treatment for their primary disease and to ineffective ventilation, accidental extubation, or the maintenance of bodily functions (fluid balance, energy loss of invasive access or monitors. In intensive care, intake, temperature control) to optimize recovery. agitation and inadequate sedation has been correlated Additional treatments provide analgesia, reduction in with adverse short- and longer-term outcome (1,2). conscious level and when indicated, muscle relaxation. Recent data has highlighted that intensive care and While this triad forms the basis of classical anesthesia, surgery is associated with long-term dysregulation of in pediatric intensive care they are used for longer peri- nociceptive mechanisms, which may change behavior ods, often below the levels required for surgery and and responses to future sensory and pain stimuli (3). with different goals. This brings to the fore specific By contrast, oversedation delays recovery, promotes problems related to the drugs. These include differing tolerance to the drugs and leads to distressing symp- pharmacokinetics and drug responses owing to age toms on their withdrawal (agitation, seizures, halluci- and individual pathophysiology, tolerance and with- nations, psychosis, fever, and tachycardia) (4). drawal, toxicity associated with long-term use, and the Maintaining ideal analgesia while at the same time need to effectively monitor drug effect against delivery. promoting earlier extubation and PICU discharge can Regrettably, administration of sedative drugs in the be difficult to achieve. Infants have an almost binary pediatric intensive care unit (PICU) is often state of consciousness (5), and a normally active 3- approached in a generic way and as an afterthought, year-old cannot easily be persuaded to remain quies- with more attention paid to the primary disease, which cent for long periods in the ICU environment using can lead to avoidable morbidity. nonpharmacological comforting measures alone. Undersedation and oversedation are both harmful. Furthermore, in the current healthcare environment, Inadequate sedation is unacceptable in a vulnerable there are considerable consumer/parental pressures to

Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd 567 Analgesia and sedation after pediatric cardiac surgery A.R. Wolf and L. Jackman ensure that all discomforts whether real or perceived good outcome in the high-risk surgical infant. Less is are avoided. Recent advances in drug choices, patient known, however, about the benefits of high-dose monitoring, and other therapeutic options, however, opioids in the context of general pediatric intensive offer new prospects for the future. care. Certainly, high-dose opioid administration has Simplistically, analgesic drugs should be given for been shown to be beneficial in the prevention and pain relief, sedative drugs for reduction in conscious treatment of specific situations such as a pulmonary level, and muscle relaxants used only for specific situa- hypertensive crisis in the at-risk infant (11) and may tions when paralysis is essential (e.g. low cardiac out- benefit patients with a low cardiac output state, or put states, when it may be necessary to anesthetize, those with a critically balanced pulmonary to systemic paralyze, and cool a child). However, there is some shunt. Data is limited, however, outside these specific cross-over in these roles: morphine has sedative prop- areas. erties, ketamine provides analgesia and anesthesia, and Control of stress responses can be achieved at much even muscle relaxants may have an additive effect on lower doses than those used in the earlier studies, and reduction of conscious state through deafferentation exposure to high-dose opioids such as fentanyl (6). Individual drugs have secondary properties that (50 lgÆkg)1 or higher) is associated with hypotension may be exploited therapeutically in specific situations, (12). The issue of benefits vs side effects of opioids in for example the use of high-dose opioids in the man- the critical care setting has been highlighted by the agement of pulmonary hypertension. Sedation and results of the ‘NEOPAIN’ study (13) in the preterm analgesia strategies need to be viewed as a collabora- neonate. The findings have been subject to consider- tive effort between anesthetist and intensivist to opti- able debate, but suggest that in preterm neonates given mize the ‘postoperative recovery gradient’ for the morphine in a more liberal fashion than controls, individual. For example, the perioperative strategy for observed comfort was improved at the expense of sys- an infant VSD may include on table or early extuba- temic hypotension. This was associated with increased tion within 4 h, with continued sedation and analgesia risk of significant short- and long-term neurological to ensure that the infant can cope with the constraints injury and adverse outcomes. While management of of intensive care management. By contrast, the com- the newborn is a special case, the inference in PICU is plex, postoperative neonate with a low cardiac output clear: provision of opioids mandates careful manage- state may have an ongoing requirement for high-dose ment to avoid hemodynamic instability whatever dose opioids, sedation, and paralysis in PICU. is provided. Children with severe cardiovascular com- promise need careful hemodynamic monitoring and the availability of supportive therapy when first Analgesia exposed to intense analgesia. Analgesic drugs comprise opioids, local anesthetic Neonates have both pharmacodynamic and pharma- agents, prostaglandin synthetase inhibitors (NSAID’s), cokinetic susceptibility to sedatives and opioids, which acetaminophen, ketamine, and alpha-2 agonists (cloni- may deter the clinician from providing effective dosing. dine and dexmedetomidine) (Table 1 and 2). Opioids However, unless effective plasma concentrations of the remain the primary analgesic agents after cardiac sur- drug have already been achieved during surgery, initial gery because of their high efficacy. Co-analgesia with loading doses may be required in the ICU. By con- NSAID’s and acetaminophen plays a key role in redu- trast, in older infants and young children, it can be dif- cing opioid requirements and side effects (7), which is ficult to maintain comfort and a relatively quiescent particularly useful for fast-track surgery. state without large drug doses, often in combination Opioids have properties other than analgesia that with multiple sedative agents. This reflects increased can be exploited in the care for the critically ill child. volume of distribution, with high clearance, that may Potent opioids such as fentanyl, sufentanil, and remi- equal or exceed that of adults, and a relative pharma- fentanil delivered at doses higher than that required codynamic drug resistance. It is not until adolescence simply for analgesia can obtund hemodynamic and that responses and handling of the drugs approaches adverse hormonal and metabolic responses in the criti- that of the adult. cally ill. Studies by Anand and others demonstrated Pharmacokinetics of the opioids are not only age that high-dose opioids used during surgery were asso- dependent but also affected by cardiac failure and car- ciated with reduced stress response, nitrogen loss, post- diopulmonary bypass (CPB) surgery. Studies with operative complications, and mortality (8–10) This led morphine and remifentanil have shown that after car- to a view that high-dose opioid analgesia, which aims diac surgery (14,15) there is a combination of an to obtund measured responses to surgery are key to increase in volume of distribution and fall in drug

568 Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd A.R. Wolf and L. Jackman Analgesia and sedation after pediatric cardiac surgery

Table 1 Opioid analgesia

Drug Indications Dose Elimination Comments

Fentanyl Analgesia 1–5 lgÆkgÆh)1 Hepatic metabolism Large bolus doses can cause Infusion hypotension Intense analgesia/ 10–15 lgÆkgÆh)1 Ventilatory depression anesthesia in ventilated patients Bolus Fentanyl Pulmonary hypertension 10–50 lgÆkg)1 Neonates may have long elimination half-lives, with delayed recovery Morphine Analgesia with sedation Loading dose Hepatic followed by Delayed recovery in neonates 50–200 lgÆkg)1 renal excretion of Controlled analgesia in the Infusion active metabolite Nausea and vomiting can be extubated patient 5–80 lgÆkgÆh)1 (morphine- problematic in older children Neonates: lower 6-glucuronide) Reduced doses may be needed with infusion rates renal impairment because of 5–20 lgÆkgÆh)1 accumulation of morphine- 6-glucuronide Oral morphine Longer term analgesia 200–500 lgÆkg)1 Oral doses need to be larger than IV once absorption has 4-hourly doses because of reduced recovered bioavailability and first pass metabolism Alfentanil Rapid offset, useful for Loading dose: Hepatic metabolism Small volume of distribution and Infusion fast-track surgery 50–100 lgÆkg)1 short elimination half-life make its Infusion: Highly protein bound offset very rapid 0.5–2 lgÆkgÆmin)1 with a small volume of distribution. Remifentanil Infusion in ventilated Analgesia: Metabolized rapidly by Alternative analgesia is required patients 0.1–0.3 lgÆkgÆmin)1 plasma and tissue before the infusion is stopped Intense, rapid onset/offset Anesthesia: cholinesterases. Unsuitable for the extubated, (largely independent of 0.5–1.5 lgÆkgÆmin)1 spontaneously breathing patient age or infusion duration) Codeine Oral medication 4-6 hourly Neonate: Hepatic and renal Also used for treatment of diarrhoea phosphate 0.5–1 mgÆkg)1 clearance because of constipating effects Child up to 12 years: 0.5–1 mgÆkg)1 Adult max dose: 240 mg >12 years: 30–60 mg Tramadol Oral medication 8 hourly >12 years: Hepatic and renal Only patients >12 years 50–100 mg 4–6 hourly clearance Adult max dose: 400 mg Nausea, vomiting, constipation Respiratory depression in large doses clearance. Care therefore must be taken to provide opioids, such as fentanyl and remifentanil (16). Neo- adequate drug loading early on, while individualizing nates undergoing ECMO require five times the initial continuous infusions over time to reflect reduced drug fentanyl infusion rate by day 6 to achieve an equiva- clearance, particularly in the setting of low cardiac lent level of sedation, because of a combination of output, and hepatic or renal impairment. Accumula- enhanced elimination (17) and pharmacodynamic toler- tion of the active metabolite of morphine, morphine-6- ance (18). glucuronide, in renal failure, can lead to excessive and Morphine is more water soluble than fentanyl, with long-lasting sedation that delays extubation. slow onset and offset, and pronounced sedative effects, All opioids are associated with tolerance, resulting which facilitates effective analgesia following anesthe- in increasing requirements to maintain adequate sia. Loading doses of 100–200 lgÆkg)1, followed by analgesia/sedation. This is accelerated by high-dose infusions of 10–60 lgÆkgÆh)1, provide reliable analgesia. opioid techniques during anesthesia, which can cause Excessive sedation, respiratory depression, nausea and ‘acute tolerance’, an effect which has been shown to be vomiting can be problematic. Patient-controlled more pronounced in shorter acting, higher potency analgesia, using a low background can be highly effec-

Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd 569 Analgesia and sedation after pediatric cardiac surgery A.R. Wolf and L. Jackman

Table 2 Nonopioid analgesia

Drug Indications Dose Comments

Paracetamol Hyperthermia 0–3 months loading dose: oral Significant but low analgesic potency 20 mgÆkg)1, rectal 30 mgÆkg)1 Co-analgesia with opioids Max daily 90 mgÆkg)1 (max adult Reduced doses may be needed in the 4 g daily) for 48 h reduce critically ill fluid restricted child to avoid thereafter to 60 mgÆkg)1 hepatic dysfunction and toxicity Diclofenac* Opioid sparing analgesia Oral or rectal: 1 mg kg)1 Has effects on gastric mucosa and platelet Max daily: 3 mgÆkg)1 function Not used under 1 year or with Can be nephrotoxic significant asthma Ibuprofen* As above Oral or rectal: 10 mgÆkg)1 As above Max daily: 40 mgÆkg)1 Ketamine Alternative intravenous analgesia IV infusion: Can be used in spontaneously breathing children to opioids (NMDA receptor 10–45 lgÆkgÆmin)1 Associated with dysphoria when used as a antagonist) sole agent May provide useful bronchodilation Clonidine Less analgesic potency than Intravenous infusion: Can cause hypotension and bradycardia morphine but can be used as 0.5–3.0 lgÆkgÆh)1 co-analgesia (orally), for longer term sedation, or for withdrawal Oral: Rebound hypertension has been described from opioids (a-2-agonist) 2–5 lgÆkg)1 4-hourly in adults

*These drugs are used extensively outside their product licence but within recommended guidelines from the Royal College of Paediatrics and Child Health. Prevention and Control of Pain in Children 1997: BMJ Publishing Group, London. [Correction added after online publication 14 December 2010: unit of dosage for clonidine changed from lg.kg.min)1 to lg.kg.h)1]

tive in older children, with nurse-controlled analgesia of 3–10 min, after single injection, is relatively age an option for the younger child. independent and not affected by renal and hepatic Fentanyl is a potent opioid, with rapid onset and function, as it is metabolized by plasma and tissue offset after short-term infusion (hours) due to its lipid cholinesterases (20). Peripheral accumulation does not solubility and extensive peripheral uptake. After occur; hence, the half-life remains independent of infu- lengthy infusions (days), recovery can be prolonged sion duration, allowing rapid offset, which remains due to its relatively long elimination half-life. While predictable, despite reduction in clearance associated analgesia can be achieved with doses of 1–5 lgÆkgÆh)1, with CPB (21). Effective plasma concentrations can be infusion rates of 10–20 lgÆkgÆh)1, may be beneficial in rapidly achieved by continuous infusion at doses of infants with low cardiac output states, critically 0.05–0.4 lgÆkgÆmin)1, with higher doses of 1 lgÆkgÆ balanced pulmonary and systemic circulations, or pul- min)1, required to obtund stress responses (22). It monary hypertension, to reduce metabolic demands has limited sedative effects, necessitating use with a and obtund hemodynamic responses to stimuli. Five to hypnotic. ten micrograms per kilograms boluses may be required Despite its desirable properties, its role in postopera- for procedures such as endotracheal suctioning and tive pain management remains to be established, with physiotherapy. High doses may provide complete cost and drug tolerance (23), restricting its utility to anesthesia and sedation for neonates, while older relatively short-term analgesia. Furthermore, its infants and children also require a hypnotic agent. potency and effects on ventilation limit its use in spon- Alfentanil is highly protein bound, resulting in a taneously breathing children and sick neonates, who small volume of distribution. It undergoes rapid hepa- can develop bradycardia and hypotension on initial tic metabolism to inactive compounds (19). Although exposure (22). In a study of ‘fast-track’ pediatric car- less potent than fentanyl, its rapid onset and offset diac surgery, it lacked superiority over fentanyl in lend it to short-term anesthesia and analgesia. terms of recovery and was associated with a significant Remifentanil is a synthetic opiate, offering potent reduction in heart rate (24). It has potential utility for analgesia, with rapid, predictable offset, making it procedural analgesia, in ventilated patients in PICU. potentially attractive for perioperative analgesia in the Given as a single intranasal dose, it provides an excel- pediatric cardiac patient. Its short elimination half-life, lent pharmacokinetic profile providing analgesia for

570 Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd A.R. Wolf and L. Jackman Analgesia and sedation after pediatric cardiac surgery

15 min. This approach avoids fluctuant levels and the Regional anesthesia for cardiac surgery remains in attendant side effects of single intravenous dose or the domain of a few enthusiasts rather than as a main- infusion of remifentanil. Verghese et al. (25) have stream technique. A variety of retrospective publica- reported this technique for intubation during inhala- tions using caudal, epidural, or spinal techniques have tional anesthesia with 5% sevoflurane after a single suggested value in the management of children after dose of 4 lgÆkg)1 of remifentanil. This exiting new cardiac surgery, but the data remains limited and lar- approach would allow modulation of analgesia for gely uncontrolled (32–34). A prospective randomized procedures such as physiotherapy and endotracheal trial of spinal anesthesia with an indwelling catheter tube change, without having to modify existing infu- demonstrated reduced stress responses and lower blood sions or break into the intravenous circuit to give addi- lactate concentrations compared to a high-dose con- tional bolus doses. ventional opioid technique in children under 3 years Once the alimentary route is effective, there is some undergoing major cardiac surgery (35). A similar tech- evidence that oral opioid administration results in nique to that described is currently being used by our reduced longer term tolerance (26–28). Certainly, the group to provide stress reduction in the operating use of enteral drug administration reduces the infection theater, facilitate early extubation and minimize the risk associated with long-term intravenous access and requirement for systemic analgesia. line manipulations (29) and should be used wherever possible. Morphine undergoes considerable first pass Sedation metabolism by the liver, if absorbed through the ali- mentary system, resulting in considerably greater dose Sedation is a broad term when used in the context of requirements than by intravenous routes, with total PICU. It may facilitate several goals including daily doses needing to be increased by a factor of five (Table 3): to ten to achieve the same level of analgesia. 1. Unconsciousness (virtual anesthesia) or reduction in conscious level 2. Reduced awareness Local analgesia techniques 3. Loss of explicit and implicit memory Afferent pain conduction can be eliminated with effec- 4. Compliance with the need to lie in a confined space, tive local anesthesia, with minimal side effects. Simple attached to monitors and invasive lines local anesthetic techniques can provide an important 5. Prevention of distress during procedures such as adjunct to any analgesia or sedation regimen, by redu- physiotherapy, radiological scanning, or minor sur- cing the background drug requirement and that for gical intervention, which may require enhanced procedural analgesia. Local anesthesia can be given by levels a wide variety of routes that include topical, infiltra- Different drugs fulfill these roles to varying extents. tion, peripheral nerve block or central continuous Benzodiazepines, for example, provide anterograde block (epidural or spinal). Topical anesthesia is used amnesia, with reduced or complete unconsciousness at regularly on the general pediatric wards and in the different doses, while phenothiazines and butyrophe- emergency department (ED). The use of these agents nones (chlorpromazine and haloperidol), used as major in PICU is often ignored except in nonintubated tranquillizing drugs in schizophrenia, have psychotro- patients. The pain associated with venous cannulation pic properties that render the patient disinterested in is sufficiently distressing, even in a moderately sedated activity. Some analgesic drugs reduce both pain and child, to necessitate additional sedation. If agents such consciousness: ketamine provides analgesia and a dis- as Ametop and EMLA are routinely applied in sociative anesthesia/sedation, clonidine produces advance of an anticipated procedure, the use of intra- analgesia and a calmed relaxed state, and morphine venous analgesia and sedation can be minimized. has additional sedative properties. Therefore, choice of Newer, faster acting agents are now available, includ- a sedative regimen needs to be tailored to the indivi- ing lidocaine-adrenaline-tetracaine gel (LAT gel), dual rather than generic. which is effective in as little as 10 min (30,31). This Neonates are a special group, in that morphine agent is finding increasing favor in the ED for wound alone can often provide sufficient analgesia and seda- suturing, with good effect. It has been suggested that tion. Outside this period, however, an analgesic and a its use should be extended to other areas, such as sedative drug are almost always necessary. Where neu- PICU, for its rapid onset of action and excellent romuscular blockade is required, a sedative drug given analgesic effects. Care does need to be given, however, at an adequate dose becomes mandatory to prevent to site of use, owing to the adrenaline component. awareness.

Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd 571 Analgesia and sedation after pediatric cardiac surgery A.R. Wolf and L. Jackman

Table 3 Sedative drugs

Drug Benefits Dose Comments

Benzodiazepines Amnesic benefit in Intravenous infusion: High risk of withdrawal with prolonged use or Midazolam addition to sedation max 200 lgÆkgÆh)1 high doses (>100 lgÆkgÆh)1). Intranasal: Helps control seizures 0.1 mgÆkg)1 each nostril Lorazepam Oral drug useful in Risk of respiratory depression – particularly midazolam withdrawal with repeated doses Diazepam Status epilepticus (rectal) Rectal: Neonate: max 2.5 mg 1 month–3 years: 5 mg 3–12 years: 5–10 mg 12–18 years: 10 mg Propofol Short-term action – rapid Intravenous infusion: Hypotension wake up max 4 mgÆkgÆh)1 Titrate to effect Check for rising lactate or acidosis, Risk of propofol infusion syndrome. Lipid load limit infusion duration with infusion. Contraindicated by FDA Phenothiazines Dissociated sedation Oral or IV: Risk of extrapyramidal side effects (chlorpromazine 500 lgÆkg)1 4–6 hourly hydrochloride) Butyrophenones Dissociated sedation Intravenous infusion: Risk of extrapyramidal side effects (haloperidol) 1 month–12 years: 25–85 lgÆkg)1 over 24 h 12–18 years: 1.5–5 mg over 24 h Clonidine Withdrawal Intravenous infusion: Can cause hypotension and bradycardia 0.5–3.0 lgÆkg)1Æh)1 Analgesic properties with Oral: Rebound hypertension described in adults opiate-sparing effect 2–5 lgÆkg)1 4-hourly Volatile agents Rapid clearance for quick Dose titrated to effect Can be useful for bronchospasm and seizures wake up Depends on delivery system and Requires vaporizer and scavenger systems gas flows Chloral hydrate Well tolerated Oral or rectal: Delayed clearance in neonates – be wary with 25–50 mgÆkg)1 regular dosing

[Correction added after online publication 14 December 2010: unit of dosage for clonidine changed from lg.kg.min)1 to lg.kg.h)1]

In the past, patients in adult ICU have been given Early extubation an opioid in combination with a low dose of anesthetic agent to ensure pain relief, hemodynamic stability, and Early extubation, with good analgesia that allows tolerance to the constraints of ITU. The potentially rapid mobilisation and recovery, is likely to be benefi- lethal side effects of anesthetic drugs used over days cial in terms of reducing postoperative complications, have only emerged after reviews of death rates and although the results of more general outcome studies analysis of recurring adverse events. These have remain inconclusive (2,40). It should be added that a included immunomodulation by barbiturates (36,37), precondition of early extubation is that the postopera- adrenocortical suppression by etomidate (38) and more tive repair and cardiovascular function are adequate to recently, mitochondrial dysfunction with propofol, in support an awake, mobilizing infant. Accelerated con- both adults and children (39). valescence confers benefits in both medical and eco- As with opioids, delivery via the enteral route may nomic terms, in that earlier discharge from the ICU avoid complications associated with intravenous deliv- improves throughput. The last 10 years has seen a ery, provided enteric function is intact. Many of the widespread acceptance of fast-track pediatric surgery, sedative agents are well absorbed and tolerated enter- with rapid extubation not only for simple cases but ally, including benzodiazepines, clonidine, and chloral also for more complex procedures (41,42). There may hydrate. An additional advantage to oral sedation is be specific advantages in immediate extubation for that it can be weaned gradually and continued follow- some cardiac conditions. For example, those with sin- ing discharge from ICU, if necessary. gle ventricle physiology (following Glenn shunt/Fontan

572 Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd A.R. Wolf and L. Jackman Analgesia and sedation after pediatric cardiac surgery repair), and Fallot’s Tetralogy, when reduced resolve. Examples include those with cardiac failure, intrathoracic pressure, associated with spontaneous pulmonary hypertension, lung or airway disease, or breathing, improve pulmonary blood flow and oxyge- ongoing sepsis. These patients are the most at risk nation (43). from drug tolerance and subsequent withdrawal. Some A variety of techniques have been described, includ- agents are considered more problematic than others: ing the use of regional analgesia; however, the difficul- the incidence of withdrawal with midazolam has been ties lie not with on table extubation but with reported as frequently as 35%. Limiting midazolam managing pain and sedation in ICU. Children need to infusion to 100 lgÆkgÆh)1 reduces the likelihood of remain sufficiently sedated after surgery to prevent withdrawal, but stopping the drug slowly does not restlessness with attempts to remove vital monitoring reduce the risk (45). or invasive catheters. In the authors’ experience, this Techniques that can be used in PICU to limit the usually necessitates a combination of both analgesia problems associated with long-term sedation include: (opioids, acetaminophen, and NSAIDs) and sedation 1. Drug cycling: changing pharmacological drug (a benzodiazepine or alpha-2 agonist). Excessive use of groups routinely to reduce the emergence of toler- potent opioids during surgery promotes acute toler- ance ance, leading to a requirement for increased analgesia 2. Sedation holidays: temporary cessation of sedative after surgery (16). The debate on whether ‘pharmaco- drugs to evaluate emergence. logical restraint’ is always preferable to limited physi- 3. Nonpharmacological techniques: oral sucrose, redu- cal restraint, such as bandages around the hands, arm cing environmental stress (noise, light, interruption splints or even true restraint, remains unresolved and of day–night cycle), swaddling, etc. provokes strong views. 4. Transfer to oral sedation where possible. Continuing anesthesia for a short time after surgery In adult intensive care, percutaneous tracheostomy in PICU allows time to assess the surgical repair and is often performed at an early stage. This allows the ensure that the patient has been optimized prior to patient to be ventilated while awake with minimal extubation. Several techniques have been described, sedation, allowing the patient to breath with support including infusions of propofol (5) and remifentanil ventilator modes, cough and clear their own secretions with midazolam (15) that allow a controlled emer- more effectively, reducing the risks of nosocomial chest gence. The use of propofol in PICU remains conten- infections. Unfortunately, this technique cannot be tious because of the issues surrounding ‘Propofol applied in children because of the technical issues asso- Infusion Syndrome’ (39,44). Despite these concerns, ciated with smaller airways and the continuing need and its contraindications for use by regulatory authori- for sedation. Surgical tracheostomy and insertion of a ties in the USA and UK, it continues to be used short- long-term indwelling central venous catheter can, how- term, at low dose (<4 mgÆkgÆh)1) for specific cases. ever, allow the patient who requires longer term venti- Patients on a propofol infusion should be closely mon- lator support to be managed with little or no sedation. itored for rising lactate, acidosis, reduced urine output Neuromuscular blockade can be useful in patients or dysrythmias. that are difficult to ventilate, such as for inverse venti- Remifentanil combined with low-dose midazolam latory ratios, high-frequency oscillation, or to deliber- provides very reliable analgesia and sedation for early ately hyper or hypoventilate them depending on their extubation. Recent studies have shown that there is a underlying pathology (46). It continues to be used in good correlation with drug infusion rate, plasma level, children that require hypothermia owing to dysrhyth- and COMFORT score (15). Before withdrawal of mia or to reduce metabolic rate and oxygen demand. remifentanil, a loading dose of opioid is needed to Neuromuscular blockade is used more commonly than maintain analgesia. in adult ICU, but as a specialty we are now using it less frequently (47), with recent reports estimating the use of long-term neuromuscular blockade in 14–16% Sedation and analgesia in the long-stay patient of ventilated days in PICU (48). Avoidance of pro- As the management of the cardiac surgical patient con- longed use is preferable because of the risks of critical tinues to improve, the need to ventilate infants over care polyneuropathy (49) and drug accumulation, days and weeks has diminished. However, those with which may result in delayed extubation. residual problems that cannot be improved immedi- Withdrawal phenomena remain a major concern in ately by further intervention or who remain marginal those patients who received sedation over many days. for cardiac or noncardiac reasons may enter a phase Symptoms can include sweating, tachycardia, hyperten- of chronic sedation and ventilation as their issues sion, agitation, posturing, withdrawal, vomiting, and

Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd 573 Analgesia and sedation after pediatric cardiac surgery A.R. Wolf and L. Jackman diarrhea. Occasionally, this prompts concerns of, and with NMJ blockade in awake volunteers, suggesting it investigation for, cardiac, neurological, or gastrointest- may not identify those inadequately sedated under inal disease. The alpha-2 agonist clonidine, chlorpro- muscle relaxant. The BIS monitor is age dependent mazine, and haloperidol can be used effectively to (58), and currently there is insufficient evidence to sup- moderate these effects, and the patients discharged port routine use of the BIS monitor, or any other neu- to wards on a weaning oral regimen over a period of rophysiological sedation scoring technique, in PICU 7–14 days. Concerns remain about abrupt cessation of (59). clonidine and the risk of rebound hypertension. Future prospects Assessment of pain and sedation The concept of using the volatile agents for longer- Sedation is administered to critically ill children term sedation on ICU’s is not a new one. Barriers in according to predicted requirement. This may not, the past have been ways to safely deliver the drug however, reflect actual requirements, with significant (vaporizers) as well as protecting the staff and environ- individual variation. Assessment of depth of sedation, ment (scavenging systems). A few ICU ventilators in with titration of analgesic and sedative drugs, is impor- the past such as the servo 900c were able to provide a tant to ensure comfort and avoid adverse outcomes, plug on vaporizer to deliver isoflurane to the critically associated with under or over-sedation. ill patient, but these are now generally obsolete. Adult Several scales, assessing behavioral and/or physiolo- data has shown that isoflurane can be an effective and gical measures have been developed and validated for practical sedative drug in the ICU when combined this purpose. These include ‘The COMFORT Scale’, with opioids with few side effects (60,61). More an objective measure of distress in ventilated pediatric recently, Meiser et al. (62) have shown the rapidly patients, validated in all age groups (50,51). It com- eliminated inhalational agent desflurane to have char- prises eight variables, each rated 1–5: alertness, calm- acteristics that make it superior to other more conven- ness/agitation, respiratory response, physical tional agents. The data in PICU is very limited. The movement, heart rate, blood pressure, muscle tone, early 10 patient study by Arnold et al. (63) used rela- and facial tension, the scale ranging from 0 to 40, with tively high doses of isoflurane as a single agent in a target range of 17–26. As with other scoring systems, patients who had already become tolerant to opioids. it is limited by inter-observer variability, provides only Their results showed that while the drug was effective, intermittent data and cannot be used in the context of there was a high incidence (50%) of tolerance agitation neuromuscular (NMJ) blockade. Cardiovascular and dystonia on withdrawal of the drug. More recent responses are also difficult to assess in patients who reports, particularly for treatment of status asthmati- are paced. cus, as part of a more balanced sedative regimen, have Neurophysiological methods and auditory evoked been more encouraging (64,65). However, there is potentials have been evaluated in the research domain. major lack of good prospective data to evaluate the Bispectral Index (BIS) monitoring utilizes data from potential of these drugs in PICU. electroencephalogram (EEG) to measure depth of The renewed interest in gaseous sedative agents is anesthesia. This technique assumes changes in fre- likely to continue with developments of new delivery quency are related and looks for phase coupling systems within ICU. Early studies with xenon show among frequency bands (biocoherence) (52,53). In great potential in the clinical areas of the developing awake individuals, there is minimal synchronization neonate and neurological injury (66,67). Unlike many because of multiple signal generators within the brain, PICU sedative agents, including midazolam, ketamine, whereas during sleep, there is less activity and the and other volatile anesthetic agents, xenon has been EEG reflects coupling between signal generators shown in animal models to provide anesthesia and (52,53). A dimensionless value, ‘the BIS number’, is sedation without producing apoptosis (68). Further calculated, which ranges from 0 to 100, 0 indicating an studies have found that xenon can offer protection isoelectric state, with 100 correlating with a fully from hypoxic injury. Hobbs et al. (69) demonstrated awake individual. Values of 40–60 are seen with gen- that rat pups exposed to hypoxic injury had a signifi- eral anesthesia (54). While it may provide an accurate cantly better outcome with the combination of cooling assessment of depth of sedation for single agents such to 32C and xenon. The functional improvement with as midazolam, propofol, and volatile agents, this is not this regimen was almost complete, sustained long-term, the case for opiates or ketamine (55,56). Furthermore, and accompanied by greatly improved histopathology. Messner et al. (57) demonstrated BIS Index to decline, Chakkrapani and colleagues demonstrated similar neu-

574 Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd A.R. Wolf and L. Jackman Analgesia and sedation after pediatric cardiac surgery roprotective benefits of xenon, with synergistic effects viable in PICU. It is rare, expensive and requires a seen with hypothermia in asphyxiated newborn piglets closed system of delivery, which necessitates the devel- (70). Issues still remain whether this drug is to become opment of new ventilatory systems.

References 1 Tonner P, Weiler N. Sedation and anlagesia preterm neonates: primary outcomes from 26 Nestler E. Molecular basis of long-term in the intensive care unit. Curr Opin Anaes- the NEOPAIN randomised trial. Lancet plasticity underlying addiction. Nat Rev thesiol 2003; 16: 113–121. 2004; 363: 1673–1682. Neurosci 2001; 2: 119–128. 2 Kehlet H, Holte K. Effect of postoperative 14 Lynn A, Nespeca NK, Bratton SL. Clear- 27 Eisch A, Barrott M, Schad CA. Opiates analgesia on surgical outcome. Br J Anaesth ance of morphine in postoperative infants inhibit neurogenesis in the adult rodent den- 2001; 87: 62–72. during intravenous infusion: the influence of tate gyrus. Proc Natl Acad Sci USA 2000; 3 Walker S, Franck LS, Fitzgerald M et al. age and surgery. Anesth Analg 1998; 86: 97: 7579–7584. Long-term impact of neonatal intensive care 958–963. 28 Pu L, Bao G-B, Xu N-J et al. Hippocampal and surgery on somatosensory perception in 15 Rigby-Jones A, Priston MJ, Sneyd JR et al. long-term potentiation is reduced by chronic children born extremely preterm. Pain 2009; Remifentanil-midazolam sedation for pedia- opiate treatment and can be restored by re- 2: 78–87. tric patients receiving mechanical ventilation exposure to opiates. J Neurosci 2002; 22: 4 Tobias J. Tolerance, withdrawal, and physi- after cardiac surgery. Br J Anaesth 2007; 99: 1914–1921. cal dependency after long-term sedation and 252–261. 29 Pronovost P, Needham D, Berenholtz S analgesia of children in the pediatric inten- 16 Guignard B, Bossard AE, Coste C et al. et al. An intervention to decrease catheter- sive care unit. Crit Care Med 2000; 28: Acute opioid tolerance: intraoperative remi- related bloodstream infections in the ICU. 2122–2132. fentanil increases postoperative pain and N Engl J Med 2006; 355: 2725–2732. 5 Murray D, Thorne GC, Rigby-Jones AE morphine requirement. Anaesthesiology 30 Chipont Benabent E, Garcia-Hermosa P, et al. Electroencephalograph variables, drug 2000; 93: 409–417. Alio Y et al. [Suture of skin lacerations concentrations and sedation scores in chil- 17 Arnold J, Truog R, Scavone J et al. using LAT gel (lidocaine, adrenaline, tetra- dren emerging from propofol infusion Changes in the pharmacodynamic response caine)]. Arch Soc Esp Oftalmol 2001; 76: anaesthesia. Pediatr Anesth 2004; 14: 143– to fentanyl in neonates during continuous 505–508. 151. infusion. J Pediatr 1991; 119: 639–643. 31 Ernst A, Marvez-Valls E, Nick TG et al. 6 Bonhomme V, Hans P. Muscle relaxation 18 Greeley W, Debruijn N. Changes in sufenta- Topical lidocaine adrenaline tetracaine and depth of anaesthesia: where is the miss- nil pharmacokinetics within the neonatal (LAT gel) versus injectable buffered lido- ing link? Br J Anaesth 2007; 99: 456–460. period. Anesth Analg 1998; 67: 86–90. caine for local anaesthesia in laceration 7 Hiller A, Meretoja OA, Piiparinen S et al. 19 Bower S, Hull CJ. Comparative pharmaco- repair. West J Med 1997; 167: 79–81. The analgesic efficacy of acetaminophen, kinetics of fentanyl and alfentanil. Br J 32 Thammasitboon S, Lutfi R, Ely BA et al. ketoprofen, or their combination for pedia- Anaesth 1982; 54: 871–877. Epidural analgesia promotes immediate tra- tric surgical patients having soft tissue or 20 Ross A, Davis P, Dear G et al. Pharmacoki- cheal extubation without complications in orthopedic procedures. Anesth Analg 2006; netics of remifentanil in anesthetised pedia- infants, children and young adults after car- 102: 1365–1371. tric patients undergoing elective surgery or diac surgery. Pediatr Anesth (in press). 8 Anand K, Hickey PR. Halothane-morphine diagnostic procedures. Anesth Analg 2001; 33 Peterson K, DeCampli WM, Pike NA. A compared with high dose sufentanil for 93: 1393–1401. report of two hundred twenty cases of regio- anaesthesia and postoperative analgesia in 21 Davis P, Wilson A, Siewers R et al. The nal anesthesia in pediatric cardiac surgery. neonatal cardiac surgery. N Engl J Med effects of cardiopulmonary bypass on remi- Anesth Analg 2000; 90: 1014–1019. 1992; 326: 1–9. fentanil kinetics in children undergoing 34 Hammer G, Ngos K, Macario A. A retro- 9 Gruber E, Laussen P, Casta A et al. Stress atrial septal defect repair. Anesth Analg spective examination of regional plus gen- response in infants undergoing cardiac sur- 1999; 89: 904–908. eral anesthesia in children undergoing open gery: a randomised study of fentanyl bolus, 22 Weale N, Rogers C, Cooper R et al. Effect of heart surgery. Anesth Analg 2000; 90: 1020– fentanyl infusion and fentanyl-midazolam remifentanil infusion rate on stress response 1024. infusion. Anesth Analg 2001; 92: 882–890. to the pre-bypass phase of pediatric cardiac 35 Humphreys N, Bays S, Parry AJ et al. 10 Hickey P, Hansen D. High-dose fentanyl surgery. Br J Anaesth 2004; 92: 1–8. Spinal anesthesia with an indwelling cathe- reduces intraoperative ventricular fibrillation 23 Vinik H, Kissin I. Rapid development of ter reduces the stress response in pediatric in neonates with hypoplastic left heart syn- tolerance to analgesia during remifentanil open heart surgery. Anesthesiology 2005; drome. J Clin Anesth 1991; 3: 285–300. infusion in humans. Anesth Analg 1998; 86: 103: 1113–1120. 11 Hickey PR, Hansen DD, Wessell DL et al. 1307–1311. 36 Sanders R, Hussell T, Maze M. Sedation Blunting of stress responses in the pulmon- 24 Friesen R, Veit A, Archibald D et al. A and immunomodulation. Crit Care Clin ary circulation of infants by fentanyl. Anesth comparison of remifentanil and fentanyl for 2009; 25: 551–570. Analg 1985; 64: 1137–1142. fast track paediatric cardiac surgery. Pae- 37 Galley H, DiMatteo MA, Webster NR. 12 Duncan H, Clode A, Weir PM et al. Infant diatr Anaesth 2003; 13: 122–125. Immunomodulation by anaesthetic, sedative stress responses in the pre bypass phase of 25 Verghese ST, Hannallah RS, Brennan M and analgesic agents: does it matter? Inten- open heart surgery: a comparison of differ- et al. The effect of intranasal administration sive Care Med 2000; 26: 267–274. ent fentanyl doses. Br J Anaesth 2000; 84: of remifentanil on intubating conditions and 38 Fraser R, Watt I, Gray CE et al. The effect 556–565. airway responses after sevoflurane induction of etomidate on adrenocortical function in 13 Anand K, Whit Hall R, Desai N et al. of anaesthesia in children. Anesth Analg dogs before and during hemorrhagic shock. Effects of morphine analgesia in ventilated 2008; 107: 1176–1181. Endocrinology 1984; 115: 2266–2270.

Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd 575 Analgesia and sedation after pediatric cardiac surgery A.R. Wolf and L. Jackman

39 Vasile B, Rasulo F, Candiani A et al. The 49 Tabarki B, Coffinieres A, van den Bergh P 60 Millane T, Bennett ED, Grounds RM. Iso- pathophysiology of propofol infusion syn- et al. Critical illness neuromuscular disease: flurane and propofol for long-term sedation drome: a simple name for a complex syn- clinical, electrophsiological, and prognostic in the intensive care unit. A crossover study. drome. Intensive Care Med 2003; 29: 1417– aspects. Arch Dis Child 2002; 86: 103–107. Anaesthesia 1992; 47: 768–774. 1425. 50 Ambuel B, Hamlett KW, Marx CM et al. 61 Spencer E, Willatts SM. Isoflurane for pro- 40 Scott N, Terfrey DJ, Ray DA et al. A pro- Assessing distress in pediatric intensive care longed sedation in the intensive care unit; spective randomised study of the potential enviroments: the COMFORT scale. J efficacy and safety. Intensive Care Med benefits of thoracic epidural anesthesia and Pediatr Psychol 1992; 17: 95–109. 1992; 18: 415–421. analgesia in patients undergoing coronary 51 Van Dijk M, de Boer JB, Koot HM et al. 62 Meiser A, Sirtl C, Bellgardt M et al. Des- artery bypass grafting. Anesth Analg 2001; The reliability and validity of the COM- flurane compared with propofol for post- 93: 523–535. FORT scale as a postoperative pain instru- operative sedation in the intensive care unit. 41 Vricella L, Dearani JA, Gundry SR et al. ment in 0 to 3-year-old infants. Pain 2000; Br J Anaesth 2003; 90: 273–280. Ultra fast track in elective congenital car- 84: 367–377. 63 Arnold J, Truog RD, Rice SA. Prolonged diac surgery. Ann Thorac Surg 2000; 69: 52 Raimer PL, Shanley TP. Bispectral index administration of isoflurane in pediatric 865–871. monitoring in the PICU. In: Shanley TP, patients during mechanical ventilation. Br J 42 Heinle J, Diaz LK, Fox LS et al. Early Zimmerman JJ, eds. Current Concepts in Anaesth 1993; 90: 273–280. extubation after cardiac operations in neo- Pediatric Critical Care. Illinois, IL: SCCM, 64 Wheeler D, Clap CR, Ponaman ML. Iso- nates and young infants. J Thorac Cardio- 2008: pp. 89–93. flurane therapy for status asthmaticus in vasc Surg 1998; 114: 413–418. 53 De Silva A. Brain electrical activity moni- children: a case series and protocol. Pediatr 43 Shekerdemian L, Bush A, Shore DF et al. toring. In: Steotling RK, Miller RD, eds. Crit Care Med 2000; 1: 5–59. Cardiopulmonary interactions after Basics of Anesthesia. Philadelphia, PA: Else- 65 Shanker V, Churchwell KB, Deshpande JK. Fontan operations. Circulation 1997; 96: vevier, 2007: pp. 314–316. Isoflurane therepy for severe refractory sta- 3934–3942. 54 Rampril I. A primer for EEG signal proces- tus asthmaticus in children. Intensive Care 44 Wolf A, Potter F. Propofol infusions in chil- sing in anesthesia. Anesthesiology 1998; 89: Med 2004; 32: 927–933. dren: when does an anaesthetic tool become 980–1002. 66 Bedi A, Murray JM, Dingley J et al. Use of an intensive care liability. Pediatr Anesth 55 Shields CH, Stayadi-Park G, McGown MY xenon as a sedative for patients receiving 2004; 17: 675–683. et al. Clinical Utility of the Bispectral Index critical care. Crit Care Med 2003; 31: 2470– 45 Jenkins I, Playfor S, Bevan C et al. for the score when compared to the University of 2477. Sedation Working Party, Pediatric Intensive Michigan Sedaton Scale in assessing the 67 Sanders R, Ma D, Maze M. Xenon: elemen- Care Society Study Group, UK. Current depth of outpatient pediatric sedation. Clin tal anaesthesia in clinical practice. Br Med United Kingdom sedation practice in pedia- Pediatr 2005; 44: 229–236. Bull 2004; 71: 115–135. tric intensive care. Pediatr Anesth 56 Aneja R, Heard A, Fletcher J et al. Seda- 68 Ma D, Williamson P, Januszewski A et al. 2007;17:675–683. tion monitoring of children by the Bispec- Xenon mitigates isoflurane-induced neuronal 46 Playfor S, Jenkins I, Boyles C et al. Consen- tral Index in the paediatric intensive care apoptosis in the developing rodent brain. sus guidelines for sustained neuromuscular unit. Pediatr Crit Care Med 2003; 4: 60–64. Anesthesiology 2007; 106: 746–753. blockade in critically ill children. Pediatr 57 Messner M, Besse U, Romstock J et al. The 69 Hobbs C, Thoreson M, Tucker A et al. Anesth 2007; 17: 881–887. Bispectral index declines during neuromus- Xenon and hypothermia combine additively, 47 Merriman HM. The techniques used to cular block in fully awake persons. Anesth offering long-term functional and histo- sedate ventilated patients. A survey of meth- Analg 2003; 97: 488–491. pathologic neuroprotection after neonatal ods used in 34 ICUs in Great Britain. Inten- 58 Davidson A, McCann ME, Devavaram P hypoxia/ischaemia. Stroke 2008; 39: 1307– sive Care Med 1981; 7: 217–224. et al. The differences in the bispectral index 1313. 48 Martin LD, Bratton SL, Quint P et al. Pro- between infants and children during emer- 70 Chakkarapani E, Dingley J, Liu X et al. spective documentation of sedative, analge- gence from anesthesia after circumcision Xenon enhances hypothermic neuroprotec- sic, and neurouscular blocking agent use in surgery. Anesth Analg 2001; 93: 326–330. tion in asphyxiated newborn pigs. Ann Neu- infants and children in the intensive care 59 Playfor S. The use of bispectral index moni- rol 2010; 68: 330–341. unit: a multicenter perspective. Pediatr Crit tors in paediatric intensive care. Crit Care Care Med 2001; 2: 205–210. 2005; 9: 25–26.

576 Pediatric Anesthesia 21 (2011) 567–576 ª 2010 Blackwell Publishing Ltd Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Pediatric heart transplantation: demographics, outcomes, and anesthetic implications* Annette Y. Schure & Barry D. Kussman

Department of Anesthesiology, Perioperative and Pain Medicine, Children’s Hospital Boston, Boston, MA, USA

Keywords Summary pediatric cardiac transplantation; pediatric anesthesia; transplant outcomes; The evolving demographics, outcomes, and anesthetic management of cardiomyopathies; congenital heart disease; pediatric heart transplant recipients are reviewed. As survival continues to failed Fontan improve, an increasing number of these patients will present to our operat- ing rooms and sedation suites. It is therefore important that all anesthesiol- Correspondence ogists, not only those specialized in cardiac anesthesia, have a basic Annette Y. Schure, understanding of the physiologic changes in the transplanted heart and the Department of Anesthesiology, Perioperative and Pain Medicine, Children’s anesthetic implications thereof. Hospital Boston, 300 Longwood Avenue, Boston, MA 02115, USA Email: [email protected]

Section Editor: Greg Hammer

*Useful Data Sources ISHLT: International Society for Heart and Lung Transplantation: http://www.ishlt.org. PHTS: Pediatric Heart Transplant Study Group: http://www.uab.edu/ctsresearch/phts/. OPTN: Organ Procurement and Transplanta- tion Network: http://www.optn.org. UNOS: United Network for Organ Sharing: http://www.unos.org. SRTR: Scientific Registry of Transplant Recipi- ents: http://www.ustransplant.org.

Accepted 9 August 2010 doi:10.1111/j.1460-9592.2010.03418.x

then, >8000 children have been registered and fol- Introduction lowed. Children account for 12.5% of cardiac trans- The first ‘successful’ pediatric heart transplant was in plantations, with 450 pediatric transplants reported 1967 in an 18 -day-old baby with tricuspid atresia who yearly from 80 centers (most in Europe and North survived 6 h (1). It was not until the late seventies that America) (Figure 1) (2). This article reviews current survival became realistic with the introduction of trends and outcomes for pediatric cardiac transplanta- cyclosporine, advances in donor management, organ tion and discusses important anesthetic implications. preservation, and recipient selection, development of the transvenous cardiac bioptome for rejection surveil- Indications and demographics lance, and clarification of brain death criteria. In 1982, the International Society of Heart and Lung Trans- The major indications for heart transplantations in plantation (ISHLT) opened a voluntary registry for infants and children are cardiomyopathy, congenital pediatric heart transplantations (<18 years); since heart disease (CHD), and re-transplantation (3,4).

594 Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd A.Y. Schure and B.D. Kussman Pediatric heart transplantation

Figure 1 Age distribution of pediatric heart recipients (by year of transplant) (2, 21).

sion. Survival rates for 1, 5, and 10 years after Cardiomyopathies (CM) diagnosis are 80%, 39%, and 20%, respectively, thus CM fall into three types: dilated, hypertrophic, and transplantation is considered early in the disease pro- restrictive, each with different risk factors, clinical cess (14–16). courses, outcomes, and treatment options.

Dilated cardiomyopathy (DCM). DCM is the most CHD common pediatric cardiomyopathy (>50%), account- Initially, a primary indication for management of ing for 75% of transplantations for CM. It is caused hypoplastic left heart syndrome (17), transplantation by neuromuscular disorders, viral myocarditis, che- was limited by the low number of available organs for motherapeutic agents, metabolic diseases, and genetic infants and high waiting list mortality. Advances in factors (5). Treatment includes b-blockers, vasodila- cardiac surgery and pediatric cardiology for complex tors, diuretics, and biventricular pacing, with most stu- CHD have significantly improved survival, resulting in dies showing a 5-year survival rate of 40–80% (6,7); a shift away from heart transplantation in infants as a the subgroup of viral myocarditis has the best chance primary therapeutic modality (Figure 2). Current indi- of recovery within the first 2 years (freedom from cations are previously repaired or palliated CHD with death or transplantation of 70% at 1 year and 50% at poor ventricular function, with single ventricle lesions 5 years after the diagnosis) (8,9). (especially the failing Fontan) accounting for the Hypertrophic cardiomyopathy (HCM). HCM constitu- majority (36%) of CHD patients transplanted (17–19). tes about 25% of pediatric CM and is characterized by ventricular hypertrophy not caused by an underlying Re-transplantation obstruction or stenosis. Seventy-five percent of cases are idiopathic, with inborn errors of metabolism (Pompe Retransplantation is a small, but steadily growing per- disease), malformation syndromes (Noonan, Beckwith- centage of heart transplantation (6–7%). The major Wiedemann), and neuromuscular disorders (Friedreich’s indications are posttransplant coronary vasculopathy ataxia) also being associated with HCM (10,11). Chil- (51%) and graft failure (16%) (2). Unfortunately, sur- dren with inborn errors of metabolism and those with vival after re-transplantation is inferior to that seen early presentation in infancy, lower shortening fraction, after primary transplantation (20). and higher posterior wall thickness on echocardiogra- The pattern of indications for pediatric heart trans- phy are at high risk of death. Progression to a dilated or plantation has changed over the past 23 years. From restrictive cardiomyopathy is often the indication for 1988 to 1995, 78% of transplants occurred in infants for transplantation (12). CHD (mostly hypoplastic left heart syndrome) and 16% for CM (21). More recently (1996–2008), 63% of trans- Restrictive cardiomyopathy (RCM). Restrictive cardio- plantations in infants occurred for CHD and 31% for myopathy (RCM), a diastolic cardiac dysfunction CM. In older children, cardiomyopathy is the major defined as restrictive filling with normal ventricular size indication (64%), with CHD accounting for 24% and and wall thickness, is rare in children (2.5–3%) re-transplantation 7% (21). North America has the (10,13). It responds poorly to medical or surgical treat- highest proportion of infant heart transplantations ment and can lead to significant pulmonary hyperten- (27%) compared to the rest of the world (11%) (2).

Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd 595 Pediatric heart transplantation A.Y. Schure and B.D. Kussman

Figure 2 Diagnosis in pediatric heart recipi- ents (age <1 year) (2, 21).

waiting, 63% were transplanted, 8% recovered, and Contraindications 12% remained listed (28). Most of the children who In addition to general contraindications (active malig- died on the waiting list weighed <15 kg. Multivariate nancy, uncontrolled infection, multi-organ failure, psy- predictors of waiting list mortality (hazard ratio 1.9– chosocial factors), irreversible pulmonary hypertension 3.1) included extracorporeal membrane oxygenation (>6 Wood UnitsÆm)2) or severe organ dysfunction (ECMO), CHD, ventilator support, dialysis, listing sta- (cirrhosis, renal insufficiency, major neurodevelopmen- tus 1A, and nonwhite race/ethnicity. The level of inva- tal disorder or stroke) can be exclusion criteria (3,4). sive hemodynamic support was a stronger predictor of Current controversies include indications for combined mortality at 30 days than UNOS status. For example, organ transplants (heart–kidney/liver/lung), controlled an infant (<10 kg) on ECMO for CHD had a 12-fold infections (HIV, hepatitis), or malignancies (22–24). increased risk of death compared to a child >10 kg with cardiomyopathy on inotropic support, although both would be listed as UNOS 1A. This study high- Timing of transplantation and waiting list mortality lighted a major problem of the current organ alloca- Generally, transplantation is considered when it offers tion system, namely the increasing conflict between an important survival advantage over alternative man- medical urgency and waiting list seniority. agement options. According to the American Heart Association, heart transplantation is a Class 1 (Level Outcomes B) recommendation for children with stage D heart failure and Class 1 (Level C) for stage C heart failure Survival (3,25) (Table 1). The United Network for Organ Shar- The average survival (time at which 50% of recipients ing (UNOS) will list these children as Status 1A or 1B remain alive) varies with the age of the recipient at in their medical urgency allocation algorithm (26) transplant. The average survival is 18 years for (Table 2). infants, 15 years for children aged 1–10 years, and As of June 11, 2010, there were 3142 patients on the 11 years for teenagers (Figure 3) (2). The highest risk US waiting list for heart transplantation, including 262 of dying is within the first 6 months and is mainly infants and children. The median waiting time for chil- caused by acute rejection or infection. Risk factors dren is about 80 days, compared to 170 days for adults for 1-year mortality include the degree of pretrans- <65 years of age (27). Unfortunately, children have plant support (ECMO, ventilator, dialysis), a diagno- the highest waiting list mortality. An analysis of 3098 sis of CHD, re-transplantation, severe infections, children listed for transplant in the United States panel reactive antibodies ‡10%, increased donor age, between 1999 and 2006 found that 17% died while and earlier era of transplant. Inotropic support, hos-

Table 1 Heart failure staging in children. Stage Interpretation Clinical Examples Modified from (25) A At risk of developing heart failure Congenital heart defects Family history of cardiomyopathy Anthracycline exposure B Abnormal cardiac structure/function Univentricular hearts No symptoms of heart failure Asymptomatic cardiomyopathy C Abnormal cardiac structure/function Repaired or unrepaired congenital Past or present symptoms of heart failure heart disease Cardiomyopathies D Abnormal cardiac structure/function Same as stage C

Continuous i.v. infusion of inotropes or PGE1 Mechanical ventilatory and/or circulatory support

596 Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd A.Y. Schure and B.D. Kussman Pediatric heart transplantation

Table 2 United Network for Organ Sharing allocation algorithm for pediatric medical urgency. Modified from (26)

Status 1A I. At least one of the following devices/therapies Mechanical circulatory support Balloon pump Mechanical ventilation <6 months old, PGE, or pulmonary hypertension at >50% systemic pressures High-dose or multiple i.v. inotropes (Dobutamine or Dopamine >7.5 mcgÆkg)1Æmin)1) Unresponsive, recurrent life-threatening arrhythmias (if all thoracic organ transplant center within the organ procurement organization agree) II. Heart status 1A Justification Form within 24 h of listing Status 1B At least one of the following Low-dose i.v. inotropes (Dobutamine or Dopamine <7 mcgÆkg)1Æmin)1) <6 months old, not meeting criteria as Status 1A Falls of growth curve and exhibits poor systemic ventricular function, or has failed previous surgical intervention Status 2 All other actively listed children pitalization, previous malignancy, ischemia time, cyto- Transplant morbidity megalovirus, or HLA mismatch had no effect on 1- year mortality. Survival rates have improved in the With improving survival, the focus of posttransplant recent era, primarily because of increased survival care is shifting toward transplant morbidity (2). during the first 6 months. Preliminary data suggest a 30% improvement in the risk-adjusted 5-year survival Functional status. There are only limited data available rate. On the other hand, media and long-term out- for children who have survived at least 10 years after comes have not been significantly affected by transplantation; 92% are reported to have no limita- advances in posttransplant care. Centers with higher tions on physical activity, and 1% requires total assis- pediatric transplant volume generally have better tance. Growth rate is normal in the majority of survival rates (2). transplant recipients. Up to a quarter of patients may have difficulties with emotional adjustment. Causes of death Rehospitalization. Within the first year, 50% of chil- Nearly 50% of deaths within 30 days are caused by dren need to be hospitalized: 35% for infection, 25% graft failure (primary or secondary to rejection) and for rejection, and 15% for both. By 10 years, this technical factors. Acute rejection, infection, and multi- number drops to about 25% (infections [36%], rejec- ple organ failure each account for about 10%. During tion [15%], or both [4%]) and is similar to adult data the first 3 years, acute rejection is the leading cause of (2). death, whereas after this period almost 60% of deaths are attributable to coronary allograft vasculopathy Rejection. Rejection is one of the major complications (CAV) or graft failure (2). and remains an ever present threat. Approximately 30–

Figure 3 Kaplan Meier survival by era for transplants from January 1982 through June 2007. (Half life: average survival, the time at which 50% of recipients remain alive.) Mod- ified from (2, 21).

Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd 597 Pediatric heart transplantation A.Y. Schure and B.D. Kussman

35% of children experience at least one episode of duction of IL-2. Cyclosporine therapy is often acute rejection within the first year, despite the increas- complicated by hypertension, nephrotoxicity, neuro- ing use of induction agents. The incidence begins to toxicity, liver dysfunction, hyperlipidemia, hypertricho- decline after 3 years, reaching about 12% between 5 sis, gingival hyperplasia, and posttransplant and 10 years. Of note, adolescents are notorious for lymphoproliferative disease (PTLD). Tacrolimus low compliance (up to 20%) with antirejection regi- (FK506) has a slightly different side-effect profile than mens. Rejection is not only one of the major causes of cyclosporine (higher incidence of diabetes and neuro- death, but it is also associated with the development of toxicity with similar nephrotoxicity) and might be cardiac allograft vasculopathy and therefore graft sur- more effective. Consequently, it is increasingly repla- vival (29,30). cing cyclosporine in many protocols (32–34). Antiproli- Recipient CD4 T cells recognize foreign antigen in ferative agents inhibit B- and T-cell proliferation; the donor heart, triggering T-cell activation, interleu- MMF is replacing azathioprine in many centers kin-2 (IL-2) secretion, and activation of monocytes/ because of less bone marrow suppression and nephro- macrophages, B cells, and cytotoxic CD8 cells. Preven- toxicity. Sirolimus, because of its adverse effects on tion and ongoing suppression of such activation is the wound healing, bone marrow function, and triglyceride fundamental concept behind immunosuppressive ther- levels, has mainly been used for rejection therapy. This apy. The majority of centers use an induction phase might change in the future because adult studies sug- followed by a maintenance regimen. gest that sirolimus may reduce the progression of cor- The induction phase is purported to reduce the inci- onary vasculopathy (35). dence of early rejection and possibly delay or avoid Many pediatric centers use a combination of triple treatment with nephrotoxic agents. This is generally immunosuppression therapy for maintenance for the first achieved by the depletion of the T-cell pool with mono- year: 38% received cyclosporine, 58% tacrolimus, 20% clonal or polyclonal antibodies or prevention of IL-2 azathioprine, 59% MMF, 55% prednisone, and 8% secretion with specific IL-2 receptor antagonists. Com- sirolimus. Between years 1 and 5 after transplant, there pared with classical agents like cyclosporine, renal dys- is a trend toward reducing the number of agents, early function is less likely with these agents. Induction agents steroid withdrawal, and reduced use of cyclosporine include the following: (i) OKT3 is a murine monoclonal and azothioprine in favor of tacrolimus and MMF (2). antibody directed against CD3 molecules on T cells; (ii) Rabbit or equine antithymocyte globulin (ATG) are poly- Rejection surveillance. Despite induction therapy, the clonal IgG antibodies extracted from thymocytes; (iii) first-year rejection rate (30–35%) has not decreased Basiliximab and daclizumab, in contrast to OKT3 and (2). The diagnosis of rejection is made on clinical signs, ATG, are specific IL-2 receptor antibodies that are echocardiographic evidence of ventricular dysfunction, usually well tolerated and increasingly being used; cur- and (36). Rejection episodes rently, 22% of pediatric heart transplant recipients are can be classified into acute vs chronic, cellular vs anti- treated with these specific IL-2 receptor antibodies. The body mediated, and with or without hemodynamic most recent ISHLT registry reports that 60% of pedia- compromise (31). Rejection with hemodynamic com- tric cardiac recipients received an induction therapy in promise seems to occur twice as often in children: 11% 2008, a significant increase from 37% in 2001. Neverthe- compared to 5% in adults (37,38). Treatment strategies less, the incidence of rejection episodes between dis- for rejection episodes are based upon the etiology and charge and 1 year has not decreased, nor has the choice severity, and usually include a 3-day pulse with ster- of induction agents influenced survival (2). oids or ATG, and if antibody-mediated rejection, the Several classes of drugs are currently used for main- addition of i.v. immunoglobulin, plasmapheresis, or tenance immunosuppression, including corticosteroids, both. Bortizemab, a proteasome inhibitor directly tar- calcineurin inhibitors (cyclosporine, tacrolimus/FK- geting the antibody-producing plasma cells, is the new- 506), antiproliferative agents (azathioprine, mycophe- est addition to the antirejection armamentarium and nolate mofetil [MMF]), and target of rapamycin inhi- currently undergoing investigation (39–42). For chil- bitors (sirolimus) (31). Corticosteroids are nonspecific dren, rejection surveillance with serial endomyocardial antiinflammatory agents, but because of their deleter- biopsies is complicated by patient size, difficult vascu- ious side effects, especially on growth and glucose lar access, and the need for anesthesia or deep seda- metabolism, most transplant centers try to restrict or tion. The timing of biopsies is tailored to the avoid their use for routine immunosuppression. probability of rejection; the first is usually obtained 7– Cyclosporine and tacrolimus (FK506) inhibit the tran- 14 days after transplant, and further biopsies are per- scription of the IL-2 gene, thereby reducing the pro- formed at 4-, 6-, 9-, and 12-week posttransplant, then

598 Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd A.Y. Schure and B.D. Kussman Pediatric heart transplantation at 6 and 9 months. After the first year, most centers antibodies. However, the immature immune system of continue surveillance with 6 monthly biopsies and newborns does not produce isohemagglutinins and annual coronary angiography. In the near future, gene levels remain low until 12–14 months of age; in addi- expression profiling of peripheral blood specimens will tion, the complement system is not fully competent in reduce the need for frequent endomyocardial biopsies young infants. As a result, infants can be recipients of (43,44). ABO-mismatched organs. This concept has been suc- cessfully used over the last decade, and the existing Cardiac allograft vasculopathy. At 10 years, 34% of data suggest that outcomes are similar to matched reci- children have CAV (2), with the onset of CAV influ- pients as long as important criteria and strategies are enced by age at transplantation. Freedom from CAV 8- followed: the ABO-mismatched organ recipient must be year posttransplant is higher in infants and younger chil- <15 months of age, have low or no isohemagglutinins dren (71% and 74%, respectively) than in children levels, and fulfill all other transplant criteria (46,47). >11 years (56%). Once CAV occurs, the 3-year graft During the pretransplant phase, they should only survival is 45% for all pediatric age groups. As adults receive ABO-compatible blood products and never have a 50% graft loss at 9 years after CAV diagnosis, it any whole blood. At the onset of cardiopulmonary is uncertain whether CAV causes a more aggressive rate bypass, recipients typically receive a two times blood of graft deterioration in younger patients or whether volume exchange transfusion, which can be repeated CAV is diagnosed earlier in adults as a result of more several times depending on the level of isohemaggluti- aggressive surveillance (45). It is also unclear why short nins checked periodically during surgery and in the ischemic times (<2 h) increase the incidence of CAV in postoperative period. The removed red blood cells can children younger but not older than 10 years. be ‘washed’ in a cell saver and retransfused.

Renal dysfunction. Ten years after transplant, 11% of Donation after cardiac death children and adolescents compared with 60% of adults have severe renal dysfunction (dialysis, kidney trans- The concept of donation after cardiac death or trans- plant, serum creatinine >2.5 mgÆdl)1). plantation after declaration of cardiocirculatory death (DCD) was reintroduced in the late nineties. With the Malignancy. Eight percent of children develop malig- consent of the family, nonbrain-dead children with a nancies at 10 years after transplant in contrast to 32% terminal diagnosis (e.g. severe cerebral hemorrhage) of adults. Almost all pediatric malignancies are lym- are taken to the operating room, prepared for organ phomas in contrast to the skin or nonlymphoma retrieval, and support is withdrawn by the primary tumors seen in adults. care or critical care physician. Death is declared after cardiopulmonary arrest, and there is an additional Hypertension. Eight years after transplantation, 69% waiting period of several minutes before organs are of pediatric survivors had hypertension, compared to harvested. Concerns about myocardial hypoxic- 94% of adults at 5 years. ischemic injury during the prearrest time initially dis- couraged the harvesting of hearts, but evidence sug- gests a ‘safe’ hypoxic-ischemic time of up to 30 min New developments (48). Despite ongoing ethical and medical controver- ABO-incompatible transplantation in infants, donation sies, the first experiences with DCD in the pediatric after cardiac death, and the increasing number of population have been encouraging (49). One report single ventricle patients with a failed Fontan palliation compared outcomes of three DCD infant cardiac represent some of the latest developments in pediatric transplantations with 17 infants with conventional cardiac transplantation. donation after brain death (50). The 6 -month survival rate was 100% in the DCD group and 84% in the con- ventional group; there was no difference in the number ABO-incompatible heart transplantation in infants of rejection episodes or echocardiographic indicators A shortage of donor organs, especially for infants, and of ventricular dysfunction. the high waiting list mortality have triggered the search for new concepts to enlarge the available donor pool Failed Fontan (46). In the adult population, ABO-incompatible solid organ transplantations are absolutely contraindicated Children with single ventricle hearts and failed Fontan because of hyperacute rejection mediated by preformed palliation represent the largest and fastest growing

Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd 599 Pediatric heart transplantation A.Y. Schure and B.D. Kussman group of transplant candidates with CHD (51). Indica- and efferent denervation have multiple effects on circu- tions include severe ventricular dysfunction, atrioven- latory control mechanisms, including altered cardiovas- tricular valve regurgitation, pulmonary hypertension, cular responses to exercise, cardiac , or long-term complications of Fontan physiology such and responses to cardiac pharmacologic agents as protein-losing enteropathy, plastic bronchitis, (Table 3) (54). Exercise testing demonstrates that intractable arrhythmias, thromboembolism, hepatic transplanted children can usually achieve only 60–70% and renal dysfunction, ascites, and pleural effusions. of normal capacity. Exercise generates an increase in These children (and adults) have had numerous hospi- cardiac output by increases in stroke volume, a highly talizations and are often in poor physical and mental preload-dependent process, with tachycardia occurring condition; they are ‘high-risk’ transplant candidates only later in response to circulating catecholamines; and require thorough preoperative evaluation. The risk the peak heart rate achieved is lower than normal. of bleeding is markedly increased because of underly- There are several case reports of profound bradycardia ing coagulopathies, multiple previous surgeries, the and even cardiac arrest after administration of neostig- presence of aortopulmonary collaterals, and the need mine for reversal of neuromuscular blockade (55–59). for pulmonary artery reconstruction. Malnourishment The denervated heart is extremely sensitive to adeno- and electrolyte imbalances can aggravate preexisting sine, with the magnitude and duration of effect on the arrhythmias and predispose to poor wound healing AV node being three to five times greater; the dose and infection. Adequate nutritional support and should be reduced by 50%. The lack of reflex tachy- advanced imaging studies, including cardiac catheteri- cardia can lead to profound hypotension with direct zation with liver biopsy and coil occlusion of collat- vasodilators (nitroglycerine, nitroprusside, hydrala- erals, are important aspects of pretransplant zine). The incidence, timing, and extent of sympathetic management (52). A retrospective multi-institutional reinnervation are still being investigated, but positive review of pediatric Fontan patients listed for cardiac effects on cardiac performance during exercise have transplantation between 1993 and 2001 reported the been demonstrated (60). survival rate on the waiting list as 78% at 6 months and 74% at 12 months, which is comparable to other Table 3 Physiology of the transplanted heart. Modified from (54) transplant candidates (19). Eighty percent of deaths on the waiting list occurred within the first 6 months Elevated filling pressures of listing, with risk factors being ventilatory support, Low normal left ventricular ejection fraction UNOS status 1, age <4 years, and shorter interval Restrictive physiology (stiff heart) Increased afterload (hypertension) since the Fontan procedure (<6 months). The actual Afferent denervation survival rates after transplantation were 77% at Silent ischemia 1 year, slightly less but not significantly different from Altered cardiac baro-and mechanoreceptors data for other recipients [Non-CHD (91%) or other Less stress-induced increase in systemic vascular resistance CHD indications (85%)]. Interestingly, protein-losing Increased blood volume (decreased natriuresis and diuresis) enteropathy resolved in all children who survived Efferent denervation 30 days. There are also reports from individual insti- Resting tachycardia (loss of vagal tone) Impaired chronotropic response to stress (dependent on tutions (53). circulating catecholamines) Electrophysiology Physiology of the transplanted heart Sinus node dysfunction in immediate postoperative period Normal AV node conduction After transplantation, the function of the surgically Shift from b1tob2 receptors denervated heart is dependent on an intact Frank Altered response to medications Starling mechanism (ability of the myocardium to No heart rate response to atropine or glycopyrrolate Decreased response to digitalis; possible severe bradycardia increase contractility and hence stroke volume in /cardiac arrest with neostigmine response to stretch (preload)) and stimulation from Exacerbated response to Ca++channel blockers, b-blockers, endogenous circulating catecholamines. The trans- adenosine planted heart is characterized by elevated filling pres- Exacerbated response to direct acting sympathomimetic agents sures, increased end-diastolic and end-systolic volumes, Decreased response to indirect acting agents (dopamine, low normal left ventricular ejection fraction, and a ephedrine) restrictive physiology (diastolic dysfunction). The left Possible sympathetic reinnervation Enhanced contractile response and exercise tolerance ventricular end-diastolic pressure is typically around Higher peak heart rates during exercise 12 mmHg 4–8 weeks after transplantation. Afferent

600 Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd A.Y. Schure and B.D. Kussman Pediatric heart transplantation

pressive drug levels. Close monitoring and frequent Anesthetic considerations for the transplanted dose adjustments are often required and should be car- heart ried out in consultation with the transplant cardiology Anesthesia for patients with a transplanted heart is team. reviewed in detail elsewhere (61–63). The preoperative evaluation should specifically focus on cardiac func- Anesthesia for pediatric cardiac transplantation tion, current rejection status, presence of infections, evidence of cardiac allograft vasculopathy, status of Anesthesia can be quite challenging, depending on the vascular access, organ dysfunction, side effects of ster- indication for transplant and the age and clinical status oids and other immunosuppressive agents (frequently of the child. Children with CM can present in critical hypertension, renal dysfunction, bone marrow suppres- condition on inotropic support and pulmonary vasodi- sion), and psychological status. A thorough review of lators (milrinone, dobutamine) or with a ventricular recent surveillance testing (echocardiography, ECG, assist device being used as a bridge to transplant. Chil- endomyocardial biopsies, cardiac catheterizations, and dren with previously repaired or palliated CHD are coronary angiography) is essential. Any decrease in frequently underweight, malnourished, and anemic exercise tolerance or new onset of dysrhythmias should from longstanding low cardiac output, feeding pro- raise a high index of suspicion for rejection and/or blems, or protein-losing enteropathy (failing Fontan) CAV. Consultation with the patient’s cardiologist as (51–53). They are often very scared and in our experi- to their current status can be extremely beneficial and ence very tolerant to sedatives and analgesics. Vascular is therefore strongly recommended. access (arterial and/or venous) can be limited as the The anesthetic plan depends on the patient’s physi- radial arteries may be occluded from previous use and cal status and the nature of the procedure. The impor- the femoral arteries and veins from previous cardiac tance of adequate preload, slower adaptive responses catheterizations and surgical access. A thorough preo- (denervation and dependency on circulating catecho- perative assessment and discussion with the surgical lamines), reduced inotropic and chronotropic reserve, team to specify potential alternative sites for central and altered responses to cardiac pharmacologic agents venous and arterial monitoring as well as peripheral is discussed earlier. Use of hemodynamic responses to bypass cannulation are essential to avoid surprises and noxious stimuli to assess adequacy of depth of anesthe- confusion in the operating room. The potential for sia is limited. massive bleeding is significant as a result of scarring Many anesthesia techniques have been successfully and adhesions from multiple previous surgeries, collat- used for children with transplanted hearts. Adequate eral vessels, and preexisting coagulopathy; this can hydration and appropriate drug selection are impor- prolong the dissection phase. Close coordination with tant to avoid hypotension. Hypotension is best mana- the donor retrieval team is important to minimize the ged with a fluid bolus and a direct-acting ischemic time for the donor organ. In addition, many sympathomimetic, if necessary, while maintaining an children are sensitized from previous blood trans- adequate depth of anesthesia. The increased risk of fusions and may require exchange transfusions or plas- infection requires strict attention to aseptic techniques mapharesis just prior to surgery. Standard anesthetic and appropriate antibiotic coverage. For the same rea- techniques for cardiac surgery are employed. As failure sons, the oral route is preferred to the nasal for endo- of the donor right ventricle postcardiopulmonary tracheal intubation. A careful risk/benefit analysis bypass is not unusual, attention needs to be focused must precede the use of invasive monitoring. When on lowering the pulmonary vascular resistance and planning central venous access and pressure monitor- inotropic support (dopamine, epinephrine, milrinone) ing, it should be remembered that endomyocardial of ventricular function (64). Nitric oxide is a useful biopsies are typically performed via the femoral vessels adjunct, and occasionally placing the patient on in infants and young children, and via the right inter- ECMO to allow for myocardial recovery and normali- nal jugular vein in older children and adults. Consid- zation of pulmonary artery pressures and resistance is eration should be given to preserving these sites for necessary in the setting of actual or impending right later use, if possible. ventricular failure. Maintenance of adequate immunosuppression in the Children presenting for re-transplantation are a perioperative period is essential. Postoperative gastric combination of all these challenges (poor cardiac func- dysfunction can delay or impair the absorption of tion from acute or chronic rejection, history of multi- cyclosporine and tacrolimus, and many medications ple previous surgeries, sensitization, etc.) plus the used in the perioperative period can affect immunosup- altered physiology of the transplanted heart and the

Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd 601 Pediatric heart transplantation A.Y. Schure and B.D. Kussman side effects of the immunosuppressive therapy on var- away from ‘irreparable’ congenital heart defects in ious organ systems. infants toward CM and increasingly to ‘failed’ repairs or palliations in older children or adolescents. Unfor- tunately, the mortality on the waiting list is still very Summary high: 20%, emphasizing the need for further changes Recent advances in donor allocation, critical care man- in the organ allocation and new ‘bridging’ solutions agement, and immunosuppressive therapy have led to like ventricular assist devices. Finally, improving trans- increased survival after pediatric cardiac transplanta- plant morbidity by finding the perfect balance between tions: The 50% survival rate for infants is currently effective immunosuppression and acceptable long-term 18 years and will most likely continue to improve. In side effects will be the major challenge for the coming the pediatric population, the indications for heart years. transplantations have slightly shifted over the years:

References 1 Kantrowitz A, Haller JD, Joos H et al. 10 Nugent AW, Daubeney PE, Chondros P et institutional study. J Heart Lung Transplant Transplantation of the heart in an infant al. The epidemiology of childhood cardio- 2006; 25: 1420–1424. and an adult. Am J Cardiol 1968; 22: 782– myopathy in Australia. N Eng J Med 2003; 21 International Society for Heart and Lung 790. 348: 1639–1646. Transplantation. Pediatric Heart Transplant 2 Kirk R, Edwards LB, Aurora P et al. Regis- 11 Towbin JA, Lowe AM, Colan SD et al. Statistics 2009. Available at: www.ishlt.org. try of the International Society for Heart Incidence, causes, and outcomes of dilated Accessed 23 April, 2010. and Lung Transplantation: Twelfth Official cardiomyopathy in children. JAMA 2006; 22 Porrett PM, Desai SS, Timmins KJ et al. Pediatric Heart Transplantation Report- 296: 1867–1876. Combined orthotopic heart and liver trans- 2009. J Heart Lung Transplant 2009; 28: 12 Colan SD, Lipshultz SE, Lowe AM et al. plantation: the need for exception status list- 993–1006. Epidemiology and cause-specific outcome ing. Liver Transpl 2004; 10: 1539–1544. 3 Canter CE, Shaddy RE, Bernstein D et al. of hypertrophic cardiomyopathy in children: 23 Bisleri G, Morgan JA, Deng MC et al. Indications for heart transplantation in findings from the Pediatric Cardiomyopathy Should HIV-positive recipients undergo pediatric heart disease: a scientific statement Registry. Circulation 2007; 115: 773–781. heart transplantation? J Thorac Cardiovasc from the American Heart Association 13 Lipshultz SE, Sleeper LA, Towbin JA et al. Surg 2003; 126: 1639–1640. Council on Cardiovascular Disease in the The incidence of pediatric cardiomyopathy 24 Ward KM, Binns H, Chin C et al. Pediatric Young; the Councils on Clinical Cardiology, in two regions of the United States. N Eng heart transplantation for anthracycline car- Cardiovascular Nursing, and Cardiovascular J Med 2003; 348: 1647–1655. diomyopathy: cancer recurrence is rare. J Surgery and Anesthesia; and the Quality of 14 Kimberling MT, Balzer DT, Hirsch R et al. Heart Lung Transplant 2004; 23: 1040–1045. Care and Outcomes Research Interdisciplin- Cardiac transplantation for pediatric restric- 25 Rosenthal D, Chrisant MR, Edens E et al. ary Working Group. Circulation 2007; 115: tive cardiomyopathy: presentation, evalua- International Society for Heart and Lung 658–676. tion, and short-term outcome. J Heart Lung Transplantation: practice guidelines for 4 Huddleston CB. Indications for heart trans- Transplant 2002; 21: 455–459. management of heart failure in children. J plantation in children. Prog Pediatr Cardiol 15 Weller RJ, Weintraub R, Addonizio LJ et Heart Lung Transplant 2004; 23: 1313–1333. 2009; 26: 3–9. al. Outcome of idiopathic restrictive cardio- 26 Renlund DG, Taylor DO, Kfoury AG et al. 5 Canter C, Naftel DC. Recipient characteris- myopathy in children. Am J Cardiol 2002; New UNOS rules: historical background tics. In: Fine R, Webber S, , Harmon W, 90: 501–506. and implications for transplantation man- Olthoff K, Kelly D, eds. Pediatric Solid 16 Russo LM, Webber SA. Idiopathic restric- agement. United Network for Organ Shar- Organ Transplantation, 2nd edn. Malden, tive cardiomyopathy in children. Heart ing. J Heart Lung Transplant 1999; 18: Mass: Blackwell, 2007: 259–264. [Chapter 30] 2005; 91: 1199–1202. 1065–1070. 6 Lipshultz SE. Ventricular dysfunction clini- 17 Bailey LL, Gundry SR, Razzouk AJ et al. 27 OPTN. Organ Procurement and Transplan- cal research in infants, children and adoles- Bless the babies: one hundred fifteen late tation Network. Current U.S. Waiting List. cents. Prog Pediatr Cardiol 2000; 12: 1–28. survivors of heart transplantation during the Available at: http://optn.transplant.hrsa.- 7 Tsirka AE, Trinkaus K, Chen SC et al. first year of life. The Loma Linda Univer- gov/latestData/rptData.asp. Accessed 11 Improved outcomes of pediatric dilated car- sity Pediatric Heart Transplant Group. J June, 2010. diomyopathy with utilization of heart trans- Thorac Cardiovasc Surg 1993; 105: 805–814. 28 Almond CS, Thiagarajan RR, Piercey GE et plantation. J Am Coll Cardiol 2004; 44: 18 Chen JM, Davies RR, Mital SR et al. al. Waiting list mortality among children 391–397. Trends and outcomes in transplantation for listed for heart transplantation in the United 8 Lee KJ, McCrindle BW, Bohn DJ. Clinical complex congenital heart disease: 1984 to States. Circulation 2009; 119: 717–727. outcomes of acute myocarditis in childhood. 2004. Ann Thorac Surg 2004; 78: 1352–1361. 29 Mulla NF, Johnston JK, Vander Dussen L Heart 1999; 82: 226–233. 19 Bernstein D, Naftel D, Chin C et al. Out- et al. Late rejection is a predictor of trans- 9 McCarthy RE 3rd, Boehmer JP, Hruban come of listing for cardiac transplantation plant in children. J RH et al. Long-term outcome of fulminant for failed Fontan: a multi-institutional Am Coll Cardiol 2001; 37: 243–250. myocarditis as compared with acute study. Circulation 2006; 114: 273–280. 30 Nicolas RT, Kort HW, Balzer DT et al. (nonfulminant) myocarditis. N Eng J Med 20 Chin C, Naftel D, Pahl E et al. Cardiac Surveillance for transplant coronary artery 2000; 342: 690–695. re-transplantation in pediatrics: a multi- disease in infant, child and adolescent heart

602 Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd A.Y. Schure and B.D. Kussman Pediatric heart transplantation

transplant recipients: an intravascular ultra- 40 Walsh RC, Everly JJ, Brailey P et al. Pro- 52 Gamba A, Merlo M, Fiocchi R et al. Heart sound study. J Heart Lung Transplant 2006; teasome inhibitor-based primary therapy for transplantation in patients with previous 25: 921–927. antibody-mediated renal allograft rejection. Fontan operations. J Thorac Cardiovasc 31 Ameduri R, Canter CE. Current practice in Transplantation 2010; 89: 277–284. Surg 2004; 127: 555–562. immunosuppression in pediatric cardiac 41 Stegall MD, Gloor JM. Deciphering anti- 53 Jayakumar KA, Addonizio LJ, Kichuk- transplantation. Prog Pediatr Cardiol 2009; body-mediated rejection: new insights into Chrisant MR et al. Cardiac transplantation 26: 31–37. mechanisms and treatment. Curr Opin after the Fontan or . JAm 32 Pollock-Barziv SM, Dipchand AI, McCrin- Organ Transplant 2010; 15: 8–10. Coll Cardiol 2004; 44: 2065–2072. dle BW et al. Randomized clinical trial of 42 Kirk AD. What should work, may not. Am 54 Cotts WG, Oren RM. Function of the tacrolimus- vs cyclosporine-based immuno- J Transplant 2010; 10: 447–448. transplanted heart: unique physiology and suppression in pediatric heart transplanta- 43 Pham MX, Teuteberg JJ, Kfoury AG, et al. therapeutic implications. Am J Med Sci tion: preliminary results at 15-month follow- Gene-expression profiling for rejection sur- 1997; 314: 164–172. up. J Heart Lung Transplant 2005; 24: 190– veillance after cardiac transplantation. N 55 Beebe DS, Shumway SJ, Maddock R. Sinus 194. Eng J Med 2010; 362: 1890–1900. NEJ- arrest after intravenous neostigmine in two 33 Kobashigawa JA, Patel J, Furukawa H Moa0912965. heart transplant recipients. Anesth Analg et al. Five-year results of a randomized, sin- 44 Jarcho JA. Fear of rejection – monitoring 1994; 78: 779–782. gle-center study of tacrolimus vs microemul- the heart-transplant recipient. N Eng J Med 56 Backman SB, Stein RD, Ralley FE et al. sion cyclosporine in heart transplant 2010; 362: 1932–1933. Neostigmine-induced bradycardia following patients. J Heart Lung Transplant 2006; 25: 45 Kirk R, Edwards LB, Aurora P et al. Regis- recent vs remote cardiac transplantation in 434–439. try of the International Society for Heart the same patient. Can J Anaesth 1996; 43: 34 English RF, Pophal SA, Bacanu SA et al. and Lung Transplantation: eleventh official 394–398. Long-term comparison of tacrolimus- and pediatric heart transplantation report – 57 Bjerke RJ, Mangione MP. Asystole after cyclosporine-induced nephrotoxicity in 2008. J Heart Lung Transplant 2008; 27: intravenous neostigmine in a heart trans- pediatric heart-transplant recipients. Am J 970–977. plant recipient. Can J Anaesth 2001; 48: Transplant 2002; 2: 769–773. 46 West LJ, Pollock-Barziv SM, Dipchand AI 305–307. 35 Mancini D, Pinney S, Burkhoff D et al. Use et al. ABO-incompatible heart transplanta- 58 Sawasdiwipachai P, Laussen PC, McGowan of rapamycin slows progression of cardiac tion in infants. N Eng J Med 2001; 344: FX et al. Cardiac arrest after neuromuscular transplantation vasculopathy. Circulation 793–800. blockade reversal in a heart transplant 2003; 108: 48–53. 47 West LJ, Karamlou T, Dipchand AI et al. infant. Anesthesiology 2007; 107: 663–665. 36 Boucek MM, Mathis CM, Boucek RJ Jr Impact on outcomes after listing and trans- 59 Backman SB. Anticholinesterase drugs and et al. Prospective evaluation of echocardio- plantation, of a strategy to accept ABO the transplanted heart. Anesthesiology 2008; graphy for primary rejection surveillance blood group-incompatible donor hearts for 108: 965; author reply 965. after infant heart transplantation: compari- neonates and infants. J Thorac Cardiovasc 60 Bengel FM, Ueberfuhr P, Schiepel N et al. son with endomyocardial biopsy. J Heart Surg 2006; 131: 455–461. Effect of sympathetic reinnervation on car- Lung Transplant 1994; 13: 66–73. 48 Singhal AK, Abrams JD, Mohara J et al. diac performance after heart transplanta- 37 Pahl E, Naftel DC, Canter CE et al. Death Potential suitability for transplantation of tion. N Eng J Med 2001; 345: 731–738. after rejection with severe hemodynamic hearts from human non-heart-beating 61 Toivonen HJ. Anaesthesia for patients with compromise in pediatric heart transplant donors: data review from the Gift of Life a transplanted organ. Acta Anaesthesiol recipients: a multi-institutional study. J Donor Program. J Heart Lung Transplant Scand 2000; 44: 812–833. Heart Lung Transplant 2001; 20: 279–287. 2005; 24: 1657–1664. 62 Ashary N, Kay AD, Hegazi AR et al. Anes- 38 Mills RM, Naftel DC, Kirklin JK et al. 49 Bernat JL, D’Alessandro AM, Port FK thetic considerations in the patient with a Heart transplant rejection with hemody- et al. Report of a National Conference on heart transplant. Heart Dis 2002; 4: 191– namic compromise: a multiinstitutional Donation after cardiac death. Am J Trans- 198. study of the role of endomyocardial cellular plant 2006; 6: 281–291. 63 Eltzschig HK, Zwissler B, Felbinger TW. infiltrate. Cardiac Transplant Research 50 Boucek MM, Mashburn C, Dunn SM et al. Perioperative implications of heart trans- Database. J Heart Lung Transplant 1997; Pediatric heart transplantation after declara- plant. Der Anaesthesist 2003; 52: 678–689. 16: 813–821. tion of cardiocirculatory death. N Eng J 64 Stobierska-Dzierzek B, Awad H, Michler 39 Everly JJ, Walsh RC, Alloway RR et al. Med 2008; 359: 709–714. RE. The evolving management of acute Proteasome inhibition for antibody- 51 Davies R, Chen JM, et al. Transplantation right-sided heart failure in cardiac trans- mediated rejection. Curr Opin Organ Trans- for the ‘‘failed’’ Fontan. Prog Pediatr Car- plant recipients. J Am Coll Cardiol 2001; 38: plant 2009; 14: 662–666. diol 2009; 26: 21–29. 923–931.

Pediatric Anesthesia 21 (2011) 594–603 ª 2010 Blackwell Publishing Ltd 603 Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Cardiomyopathy and heart failure in children: anesthetic implications David N. Rosenthal1 & Gregory B. Hammer2

1 Pediatric Cardiology, Division of Pediatric Cardiology, Stanford University School of Medicine, Palo Alto, CA, USA 2 Department of Anesthesia, Stanford University School of Medicine, Stanford, CA, USA

Keywords Summary cardiomyopathy; heart failure; pediatric anesthesia The purpose of this article is to provide a brief but systematic overview of heart failure and cardiomyopathy in children and the anesthetic manage- Correspondence ment of these patients. We will begin with disease definitions and descrip- Gregory B. Hammer, tions of the disorders. Our review will include the epidemiology and Professor of Anesthesia and Pediatrics, etiology of the more prevalent underlying causes of heart failure, the prin- Department of Anesthesia, Stanford cipal pathophysiology of the specific cardiomyopathies, as well as the com- University School of Medicine, 300 Pasteur Drive, Stanford, CA 94305, USA mon therapies in use today in both inpatient and outpatient settings. Email: [email protected] Important implications for anesthetic management will be highlighted.

Section Editor: Chandra Ramamoorthy

Accepted 18 February 2011 doi:10.1111/j.1460-9592.2011.03561.x

more specific etiologies such as the viral inflammatory Introduction DCM more typically referred to as myocarditis This review will include definitions and descriptions of (Figure 1). cardiomyopathy and heart failure in children. The epi- Heart failure is characterized by reduced cardiac demiology and etiology of the more prevalent causes output and increased venous pressure in the systemic of heart failure, pathophysiology of specific cardiomy- and/or pulmonary venous system. Progression of the opathies, and common therapies in use today will be disease occurs via myocyte dysfunction, myocyte described. Important implications for anesthetic man- death, and pathologic changes in the extracellular agement will be highlighted. matrix. Heart failure, in turn, causes both circulatory and endocrine abnormalities. While congestive heart failure is frequently associated with cardiomyopathy, Classification of cardiomyopathy and heart the two terms are not synonymous. Cardiomyopathy failure may exist without overt heart failure, and heart failure Cardiomyopathy is a condition in which there is an may occur in the absence of cardiomyopathy. abnormality of the myocardium. The most commonly While the categorization of cardiomyopathy is aimed used nomenclature used to describe the types of car- at distinguishing important anatomic and etiologic diomyopathy was developed by the World Health subtypes, the classification of heart failure focuses pri- Organization (1). Accordingly, cardiomyopathies are marily on the severity of illness. One of the oldest clas- classified according to etiology and physiology as (i) sification systems, the New York Heart Association dilated cardiomyopathy (DCM), (ii) restrictive cardio- Class (NYHA), remains useful even today. The levels myopathy (RCM), (iii) hypertrophic cardiomyopathy, of severity in NYHA correlate moderately well with (iv) arrhythmogenic right ventricular cardiomyopathy, both prognosis and biomarkers such as circulating lev- and (v) unclassified. Within these broad categories are els of norepinephrine (2,3). The NYHA classification

Pediatric Anesthesia 21 (2011) 577–584 ª 2011 Blackwell Publishing Ltd 577 Heart failure in children: anesthetic implications D.N. Rosenthal and G.B. Hammer

Cardiomyopathy

DCM ARVC HCMRCM Unclassified Figure 1 World Health Organization Cardio-

Idiopathic myopathy Classification. DCM, dilated car- diomyopathy; ARVC, arrhythmogenic right Inflammatoryy ventricular cardiomyopathy; HCM, hypertro- phic cardiomyopathy; RCM, restrictive car- Infectious diomyopathy.

Table 1 Classification of heart failure severity (5,6) of heart failure. Stage C is marked by a present or past NYHA classification Ross classification state of symptomatic heart failure. Stage C includes patients with NYHA class II and III disease as well as I No symptoms No limitations or symptoms patients who have had signs or symptoms but are II Symptoms with Mild tachypnea or diaphoresis with asymptomatic by virtue of heart failure treatment. moderate exertion feeding in infants; dyspnea on exertion in older children Stage D refers to patients with advanced heart failure III Symptoms with mild Marked tachypnea or diaphoresis requiring specialized therapy such as intravenous exertion with feeding or exertion inotropes or mechanical circulatory support. IV Symptoms at rest Symptomatic at rest with tachypnea, retractions, grunting, or diaphoresis Epidemiology A number of large studies describe the prevalence of DCM and of heart failure in children. Arola et al. (8) ranges from I (asymptomatic) to IV (heart failure described a prevalence of idiopathic DCM of 2.6 per symptoms at rest) (Table 1). Designed for adult 100 000 children in Finland, with an annual incidence patients, this scale has also been applied to children of 0.34 cases per 100 000 population. Nugent et al. (9) (4). The Ross Heart Classification Scale, with severity cited an incidence of 1.24 cases of new cardiomyopa- also ranging from I to IV, may be used in infants (5). thy per 100 000 children <10 years of age in Austra- An updated classification scheme that incorporates lia. Of these cases, 58% were DCM and 26% were intensity of medical therapy was developed in 2001 hypertrophic cardiomyopathy. Towbin et al. (10) pub- and adapted for children in 2004 (Table 2) (6,7). Heart lished similar data in the United States, with an overall failure Stage A refers to the subset of the population incidence of DCM of 0.57 cases per 100 000. This with grossly intact function but recognized to have a study demonstrates a markedly increased incidence in risk factor for the development of heart failure, e.g., infants (4.4 cases per 100 000). Hypertrophic cardio- exposure to cardiotoxic chemotherapy or a valvular myopathy in children has a lower disease incidence of abnormality that imposes a volume overload on the 0.32 cases per 100 000. RCM is rarer still, with an esti- ventricle. Stage B is a presymptomatic state in which mated prevalence of 0.03 per 100 000 children. there is discernible myocardial dysfunction, including Another way to determine the population frequency chamber enlargement or reduced systolic function. of heart failure is to examine the frequency with which Patients in Stage B have never had signs or symptoms heart transplantation is performed in children. These data are tracked by both United Network for Organ

Table 2 Heart failure staging for infants and children (7) Sharing (UNOS) and a multicenter research database, the Pediatric Heart Transplant Study (PHTS). The Stage Interpretation PHTS annual report shows that approximately 350

A Increased risk of developing heart failure, but with heart transplants are performed annually in children in normal cardiac function and size the United States (11). The two most common groups B Abnormal cardiac morphology or function, with no heart of diagnoses leading to heart transplantation are car- failure symptoms or history of symptoms in past diomyopathy and congenital heart disease (all types C Underlying structural or functional heart disease and combined) (Table 3). The frequency of these two diag- heart failure symptoms past or present noses varies by age but together they account for over D End-stage heart failure 90% of all pediatric heart transplants.

578 Pediatric Anesthesia 21 (2011) 577–584 ª 2011 Blackwell Publishing Ltd D.N. Rosenthal and G.B. Hammer Heart failure in children: anesthetic implications

physiology. In RCM, the systolic function of the heart Pathophysiology of heart failure is typically preserved. Diastolic function is impaired, The sequence of heart failure begins with myocyte leading to elevated atrial pressure. Pulmonary conges- injury because of a congenital cardiomyopathy or tion occurs when the left atrial pressure is elevated, underlying structural disease. The initial injury results and hepatic and renal dysfunction occurs when the in a reduction in cardiac output that is, in turn, coun- right atrial and systemic venous pressures are tered by two major defense mechanisms. The first of increased. Cardiac output is quite limited in the setting these is intrinsic catecholamine stimulation. Catechol- of advanced RCM, in which there is a limitation in amine secretion has the early effect of increasing stroke volume because of the impaired diastolic filling. myocardial contractility and, via peripheral vasocon- The major clinical problems with hypertrophic cardio- striction, redirecting systemic blood flow to critical myopathy are not generally those related to heart fail- organs, including the heart. Enhanced catecholamine ure per se, but rather dysrhythmias. The reader is secretion, however, leads to further myocyte injury, referred to other reviews for further details regarding dysfunctional intra-cellular signaling in response to hypertrophic cardiomyopathy (12). beta-adrenergic stimulation, and ultimately myocyte death. The second important ‘compensatory’ mecha- Outpatient treatment nism elicited by circulatory insufficiency is the stimula- tion of the renin–aldosterone–angiotensin system In the current era, the outpatient treatment of heart (RAAS). The elevation of both aldosterone and angio- failure is based on several clinical endpoints. The tensin II promotes cardiac fibrosis and apoptosis. guidelines for heart failure management in children These mechanisms may temporarily contribute to cir- published by the International Society of Heart and culatory stability, but over time become maladaptive Lung Transplantation offer detailed assessment of the and promote the progression of heart failure. evidence base for the practices described later (11). On a local organ level, diminished myocardial con- The first treatment strategy for heart failure is to traction in cardiomyopathy leads to reduced ejection establish a euvolemic state through a combination of fraction and chamber enlargement. As a consequence fluid restriction and diuretic therapy. This achieves of the increased diameter of the heart, coupled with symptom relief and also promotes good function of approximately stable wall thickness, there is a dramatic the major organ systems. This treatment, however, has increase in systolic wall stress that leads to oxygen sup- no direct influence on the rate of progression of heart ply–demand mismatch in the myocardium. In some failure. Furosemide is the diuretic most commonly patients, enlargement of the ventricle leads to insuffi- used for this purpose in children because of its rela- ciency of the atrioventricular valve and imposes a vol- tively predictable action. ume load on the already stressed myocardium. This, The second therapeutic option is to inhibit the too, is maladaptive and leads to disease progression. RAAS pathway. This is critical to slowing disease pro- The mechanisms described earlier apply to all forms gression. In adults, the efficacy of this approach has of DCM and do not appear to depend upon the been proven in multiple large randomized clinical trials underlying etiology (infectious, genetic, toxic, etc). It is (13–15). No such randomized trials have been made in likely that these same mechanisms operate in the children, although it is generally believed that the setting of congenital heart disease with reduced systolic RAAS pathway plays an important role in heart fail- function, although research in this area is incomplete. ure in children as well. The class of medications most The diseases of RCM and hypertrophic cardiomyopa- commonly used for RAAS blockade is the angiotensin thy, however, are associated with very different patho- converting enzyme (ACE) inhibitors. These medica- tions also cause arteriolar vasodilation and afterload reduction for the diseased myocardium. Afterload Table 3 Diagnosis of heart transplant recipients (16) reduction, however, is not the major benefit from ACE inhibitor therapy, as was shown in the V-HEFT II trial Recipient Recipient Recipient (14). In this study, enalapril was compared with hydra- age <1 age 1–10 age 11–17 Etiology year (%) years (%) years (%) zine, a nonspecific vasodilator. Despite the comparable degrees of vasodilation, the group treated with enalap- Cardiomyopathy 31 55 24 ril had better outcomes, with a reduction in mortality Congenital heart disease 63 36 64 of 30%. As an alternative to ACE inhibitors, angioten- Retransplantation 1 6 7 sin receptor blockers may produce similar effects. Other 5 3 4 Finally, the use of medications such as spironolactone

Pediatric Anesthesia 21 (2011) 577–584 ª 2011 Blackwell Publishing Ltd 579 Heart failure in children: anesthetic implications D.N. Rosenthal and G.B. Hammer or eplerenone to produce blockade of aldosterone has with mechanical circulatory support either as ‘bridge been proven efficacious, including in patients treated to transplant’ or ‘destination’ therapy, i.e., with no with ACE inhibitors (16). plan of performing heart transplantation. This trend A third option for outpatient heart failure therapy is also applies to pediatric patients with severe heart fail- the use of beta-receptor antagonists. Chronic eleva- ure. Such patients may have apparently stable circula- tions in plasma catecholamine concentrations cause tory dynamics but are anticoagulated and at risk of heart failure progression via direct myocyte toxicity. severe bleeding associated with surgical procedures. Accordingly, circulating norepinephrine is a potent Additionally, cardiopulmonary resuscitation in such predictor of patient mortality. Beta-receptor antago- patients is hazardous and poses a risk of dislodging nists afford the opportunity to effectively block this the intra-cardiac and intra-aortic cannulae. Children mechanism and thereby retard disease progression, or treated with mechanical circulatory support devices in some cases, to allow for significant degrees of car- constitute a novel group of patents with unique treat- diac recovery (17–19). However, implementation of ment requirements. this therapy may worsen heart failure by elimination of a critical element of circulatory support. It is, there- fore, necessary to begin treatment with low doses of Anesthetic management beta-blockers (e.g., propranolol, carvedilol) and titrate upwards slowly over a period of weeks to months. Preoperative considerations During heart failure exacerbations and inpatient treat- When evaluating the outpatient with heart failure, ment, it may be necessary to taper or discontinue beta- anesthetic concerns arise both from heart failure blockade. itself and its treatment. Even ambulatory heart fail- ure patients can have low blood pressure at baseline as a result of the combined effects of disease, diure- Inpatient treatment tic therapy, ACE inhibition, and beta-blockade. Children who require outpatient therapy for heart fail- These conditions may predispose to severe hypoten- ure often require hospitalization owing to disease pro- sion following the administration of anesthetic gression or temporary exacerbations of heart failure. agents. In addition, carvedilol, which is one of the Intravenous inotropes and vasodilators (or inodilators) commonly employed beta-blockers, has been shown are the mainstay of initial therapy, serving to restore to reduce the endogenous chronotropic and vasodila- adequate perfusion pressure and systemic oxygen tory responses to anesthesia (21). Patients who are delivery. The agents selected are often specific to receiving comprehensive outpatient therapy are likely institutional practice patterns: milrinone (a phosphodi- to have intravascular hypovolemia at baseline and esterase inhibitor), dopamine (a mixed alpha- and this will be exacerbated by any period of preanes- beta-agonist), and dobutamine (a beta-agonist) are thetic fasting. These patients may become acutely among the more frequently used medications. Reduc- hypotensive as a result of anesthetic-induced venodi- tion in metabolic demand via mechanical ventilation is lation, which can produce unexpectedly large decre- another method of support. Several types of mechani- ments in cardiac preload. Patients with poor systolic cal circulatory support are employed to ‘bridge’ the function and marked degrees of ventricular enlarge- heart failure patient to transplantation when more ment may have relatively normal vital signs and conventional therapies have failed. physical examination but have acute deterioration Mechanical circulatory support therapies have following anesthetic exposure. Initiation of low-dose evolved rapidly over the past decade for both adults inotropic therapy prior to anesthetic induction may and children. In 2006, Blume described the use of be warranted; this facilitates an increase in adminis- mechanical circulatory support in a 10-year period tration rate in the event of decreased cardiac output across the United States, including 99 of 2300 patients during the anesthetic. who subsequently had cardiac transplantation (20). It is crucial that the preprocedure anesthetic assess- More recent UNOS data are that 12% of children ment include a detailed history, including exercise tol- under age 12 years and 20% of teenagers have been erance, prior anesthetic records, medications, and most treated with mechanical circulatory support while recent laboratory assessments, including echocardiog- awaiting heart transplantation. This percentage contin- raphy. Direct communication between the treating car- ues to increase with the availability of devices such as diologist and the anesthesiologist should transpire the Berlin Heart EXCOR, which is suitable for chil- proximate to the time of administration of anesthetic dren as small as 3 kg. Adults are increasingly treated agents. Through this communication, the patient’s clin-

580 Pediatric Anesthesia 21 (2011) 577–584 ª 2011 Blackwell Publishing Ltd D.N. Rosenthal and G.B. Hammer Heart failure in children: anesthetic implications ical status can be adequately conveyed, including resistance [SVR]). Both agents may depress myocardial details of overall functional status as well as a detailed contractility to some degree, although in infants with interpretation of results. For example, normal hearts the cardiac index did not change after it is not uncommon for echocardiographic ejection induction of anesthesia with either agent in one echo- fraction to vary from study to study, and this may lead cardiographic study (22). Ketamine administration for to errors in understanding the patient’s severity of anesthetic induction is often associated with sympat- illness. Of equal importance, the anesthesiologist is homimetic effects, i.e., increased heart rate and blood often familiar with the aspects of both the procedure pressure, in patients with normal cardiac function. Ke- and the anesthetic requirements; this can be invaluable tamine, however, does cause direct myocardial depres- in modifying the selected procedure (e.g., open vs lapa- sion in the isolated neonatal rat myocyte (23). In roscopic gastrostomy and fundoplication). Communi- patients with cardiomyopathy and end-stage heart fail- cation regarding patient management in advance of ure, ketamine administration may cause decreased car- the scheduled procedure is of paramount importance. diac output and hypotension. As with propofol and Generally, inpatients represent a more severely thiopental, hypotension may, in turn, lead to myocar- affected group than outpatients, and the risks of dial ischemia that exacerbates the initial circulatory anesthesia are even greater. When patients are main- depression; arrhythmias and cardiac arrest may ensue. tained on intravenous inotropes, the potential for Combinations of ketamine and other anesthetics, dysrhythmias is significant and the ability to escalate including sevoflurane and propofol, have been used for support to maintain physiologic cardiac output and maintenance of anesthesia during cardiac catheteriza- blood pressure is decreased. It is important in these tion in children with good results (24,25). Etomidate is cases that the child’s treatment plan and strategies an induction drug that appears to cause less myocar- and decisions as to the escalation of care are articu- dial depression than the other three agents (26). In a lated prior to administration of anesthesia. This study of normal and diseased human cardiac tissue, might include discussions to ensure suitability and etomidate caused no decrease in myocardial contractil- immediate availability of extracorporeal membrane ity; ketamine caused slight negative inotropic effects; oxygenation (ECMO) or other circulatory support and thiopental significantly depressed myocardial func- device prior to anesthetic induction as well as tion (27). In children with congenital heart defects, whether the child is ultimately a candidate for car- etomidate (0.3 mgÆkg)1 IV) and ketamine (4 mgÆkg)1 diac transplantation. IV) did not cause a change in mean arterial pressure. In another study of a small group of children with normal cardiac function, the same dose of etomidate Anesthetic agents was not associated with any measurable circulatory Children with heart failure ± cardiomyopathy com- changes (28). Etomidate causes a decrease in serum monly undergo diagnostic and therapeutic procedures cortisol because of adrenal suppression and its use for which anesthesia is required. Diagnostic proce- remains somewhat controversial (29,30). Etomidate is dures include CT scan, MRI, and cardiac catheteri- used in children with end-stage heart failure at the zation. Therapeutic procedures include placement of authors’ institution because the relative risk of adrenal a central venous catheter, gastrostomy tube, and suppression appears to be less than circulatory collapse pacemaker or implantable cardioverter-defibrillator, associated with the use of alternative anesthetic as well as cardiac transplantation. These patients are induction agents. at an increased risk of anesthetic complications com- mensurate with their severity of disease. A brief Volatile anesthetics description of commonly used anesthetic agents and their effects on cardiac function follows. All of the volatile anesthetic gases depress myocardial contractility, especially in the newborn (31). Isoflurane and sevoflurane lower SVR and, therefore, blood pres- Intravenous induction agents sure. The net effect on cardiac output may depend on Intravenous drugs used for induction of anesthesia in preload and concomitant use of other cardio- and children with cardiac disease include propofol, thio- vasoactive agents. These agents do not appear to sig- pental, ketamine, and etomidate. Propofol and thio- nificantly change the ratio of pulmonary-to-systemic pental exert a dose-dependent hypotensive effect owing blood flow in patients with shunt lesions. In children to a reduction in preload (i.e., dilation of venous with single ventricle physiology and stable ventricular capacitance vessels) and afterload (systemic vascular function, sevoflurane at concentrations up to 1.5

Pediatric Anesthesia 21 (2011) 577–584 ª 2011 Blackwell Publishing Ltd 581 Heart failure in children: anesthetic implications D.N. Rosenthal and G.B. Hammer minimum alveolar concentration does not decrease Dexmedetomidine myocardial performance index (32,33). The volatile The role of dexmedetomidine in anesthetizing children agents may produce an increase in heart rate in normal with heart disease, especially those with heart failure, children, especially during induction and during light remains unclear (35). Although this drug may allow anesthesia with or without surgical stimulation. The for reduced doses of other maintenance agents, includ- effects of volatile agents have not been well studied in ing inhaled agents, propofol, ketamine, and opioids, children with heart failure. Even low concentrations of undesired slowing of the heart rate may occur because volatile anesthetics may cause severe hypotension in of the effects on cardiac conduction (36). This may be this population because of myocardial depression and particularly deleterious in children with limited stroke reduction in SVR, both of which may be associated volume in whom cardiac output may be relatively with decreased coronary perfusion and progressive dependent on heart rate. The efficacy of adding dex- compromise in cardiac function. medetomidine to a propofol-based regimen for mainte- nance of anesthesia has been questioned (37). Kipps et al. (38) reviewed records from 26 children Nitrous oxide with heart failure undergoing 34 procedures between There is little information regarding the effects of 2002 and 2005. Only children with fractional shorten- nitrous oxide in children with congenital heart disease ing £28% were included, so that few patients with and those with heart failure. Adverse effects of the RCM and arrhythmogenic right ventricular cardio- addition of nitrous oxide to volatile anesthetics may myopathy were studied. Patients were categorized as well be attributable to the increased depth of anesthe- having mild (fractional shortening [FS] 23–28%), sia, i.e., equivalent increases in the concentrations of moderate (FS 16–22%), or severe (FS <16%) systolic inhaled anesthetics have at least as much cardiodepres- dysfunction. All patients were receiving one or more sant effect. In general, nitrous oxide appears to have medications for heart failure. The pediatric cardiac little effect on heart rate, cardiac contractility, pulmo- anesthesia team managed 85% of the patients, includ- nary vascular resistance, and SVR compared with ing all of those with structural heart disease or other agents administered in equi-anesthetic doses. idiopathic cardiomyopathy. The anesthetic manage- ment varied according to the severity of disease. The majority of children had anesthesia induced with Opioids ketamine or etomidate. Maintenance was achieved with propofol and ketamine or a volatile anesthetic Several publications have shown that high-dose opi- agent. While 5 of the 13 procedures in patients with oids, including fentanyl and sufentanil, have minimal mild or moderate ventricular dysfunction were adverse effects on cardiac function and blunt the discharged home on the day of the procedure, all of adverse stress response to surgery in infants. Remifen- the procedures in those with severe dysfunction were tanil may be useful because its short onset and offset hospitalized afterward. Fourteen of these 21 patients allow careful titration. In children without end-stage ) were admitted to the intensive care unit, including 10 heart disease, remifentanil doses up to 0.1 lgÆkg 1Æ ) who were outpatients or admitted to the regular inpa- min 1 IV may be administered to spontaneously tient care unit prior to the procedure. Complications breathing patients during painful procedures (e.g., car- were noted in 38% of the 34 procedures, primarily diac catheterization) with maintenance of circulatory (83%) in those with severe dysfunction. The most function and without excessive respiratory depression common complication was hypotension requiring (34). Histamine release associated with morphine vasoactive drug administration. One death occurred administration in some patients may cause undesired in a 16-year-old boy with severe DCM after the decrease in preload and/or SVR. Caution must be development of hypotension followed by cardiac applied to the use of large doses of all opioids in chil- arrest during placement of a biventricular pacemaker. dren with severe heart failure. Adverse effects may include slowing of the heart rate and a reduction in sympathetic tone in general, either of which may pro- Conclusions voke exacerbation of heart failure in this population. Infants and children with heart failure or cardiomyop- The concomitant use of sedative agents, e.g., ben- athy represent one of the most challenging groups zodiazepines, increases the risk of circulatory depres- presenting to pediatric anesthetists. Some controversy sion. Studies of combination drug therapy in children exists as to whether pediatric cardiac anesthetists with heart failure have not been carried out.

582 Pediatric Anesthesia 21 (2011) 577–584 ª 2011 Blackwell Publishing Ltd D.N. Rosenthal and G.B. Hammer Heart failure in children: anesthetic implications should care for all of these patients. Attention must be utilizing local anesthesia with or without a small dose focused on preparation of the anesthetizing location, of an anxiolytic drug. Prior to the administration of ensuring that resuscitation drugs are prepared in any anesthetic agents, there should be discussion with appropriate concentrations prior to the arrival of the the physician performing the procedure aimed at patient. For all but the mildest of cases, intravenous confirming the need for anesthesia (and the procedure) induction should be undertaken with consideration of and minimizing the duration of the procedure. When administration of small doses of hypnotic agents to appropriate, arrangements should be made for cardio- ascertain a ‘dose–response’ with respect to blood pres- pulmonary bypass (or ECMO) back-up in the event of sure. In the highest risk patients, placement of an arte- development of a life-threatening low cardiac output rial catheter prior to induction should be considered state.

References 1 Richardson P, McKenna W, Bristow M myopathy in Australia. N Engl J Med 2003; 19 Packer M, Fowler MB, Roecker EB et al. et al. Report of the 1995 World Health 348: 1639–1646. Effect of carvedilol on the morbidity of Organization/International Society and Fed- 10 Towbin JA, Lowe AM, Colan SD et al. patients with severe chronic heart failure: eration of Cardiology Task Force on the Incidence, causes, and outcomes of dilated results of the carvedilol prospective random- Definition and Classification of cardiomyop- cardiomyopathy in children. JAMA 2006; ized cumulative survival (COPERNICUS) athies. Circulation 1996; 93: 841–842. 296: 1867–1876. study. Circulation 2002; 106: 2194–2199. 2 Swedberg K, Eneroth P, Kjekshus J et al. 11 Boucek MM, Waltz DA, Edwards LB et al. 20 Blume ED, Naftel DC, Bastardi HJ et al. Hormones regulating cardiovascular func- Registry of the International Society for Outcomes of children bridged to heart tion in patients with severe congestive heart Heart and Lung Transplantation: ninth offi- transplantation with ventricular assist failure and their relation to mortality. CON- cial pediatric heart transplantation report- devices: a multi-institutional study. Circula- SENSUS Trial Study Group. Circulation 2006. J Heart Lung Transplant 2006; 25: tion 2006; 113: 2313–2319. 1990; 82: 1730–1736. 893–903. 21 Bollano E, Tang MS, Hjalmarson A et al. 3 Benedict CR, Shelton B, Johnstone DE 12 Colan SD. Hypertrophic cardiomyopathy in Different responses to dobutamine in the et al. Prognostic significance of plasma nor- childhood. Heart Fail Clin 2010; 6: 433–444. presence of carvedilol or metoprolol in epinephrine in patients with asymptomatic 13 Effect of enalapril on survival in patients patients with chronic heart failure. Heart left ventricular dysfunction. SOLVD Investi- with reduced left ventricular ejection frac- 2003; 89: 621–624. gators. Circulation 1996; 94: 690–697. tions and congestive heart failure. The 22 Wodey E, Chonow L, Beneux X et al. Hae- 4 Bruns LA, Chrisant MK, Lamour JM et al. SOLVD Investigators. N Engl J Med 1991; modynamic effects of propofol vs. thiopen- Carvedilol as therapy in pediatric heart fail- 325: 293–302. tal in infants: an echocardiographic study. ure: an initial multicenter experience. J Pedi- 14 Cohn JN, Johnson G, Ziesche S et al. A Br J Anaesth 1999; 82: 516–520. atr 2001; 138: 505–511. comparison of enalapril with hydralazine- 23 Kudoh A, Matsuki A. Ketamine inhibits 5 Ross RD, Daniels SR, Schwartz DC et al. isosorbide dinitrate in the treatment of inositol 1,4,5-triphosphate production Plasma norepinephrine levels in infants and chronic congestive heart failure. N Engl J depending on the extracellular Ca2+ concen- children with congestive heart failure. Am J Med 1991; 325: 303–310. tration in neonatal rat cardiomyocytes. Cardiol 1987; 59: 911–914. 15 Effects of enalapril on mortality in severe Anesth Analg 1999; 89: 1417–1422. 6 Hunt SA, Baker DW, Chin MH et al. congestive heart failure. Results of the 24 Williams GD, Philip BM, Chu LF et al. ACC/AHA guidelines for the evaluation Cooperative North Scandinavian Enalapril Ketamine does not increase pulmonary and management of chronic heart failure in Survival Study (CONSENSUS). The CON- vascular resistance in children with pulmo- the adult: executive summary. A report of SENSUS Trial Study Group. N Engl J Med nary hypertension undergoing sevoflurane the American College of Cardiology/Ameri- 1987; 316: 1429–1435. anesthesia and spontaneous ventilation. can Heart Association Task Force on Prac- 16 Pitt B, Zannad F, Remme WJ et al. The Anesth Analg 2007; 105: 1578–1584. tice Guidelines (Committee to revise the effect of spironolactone on morbidity and 25 Singh A, Girotra S, Mehta Y et al. Total 1995 Guidelines for the Evaluation and mortality in patients with severe heart fail- intravenous anesthesia with ketamine for Management of Heart Failure). J Am Coll ure. Randomized Aldactone Evaluation pediatric interventional cardiac procedures. Cardiol 2001; 38: 2101–2113. Study Investigators. N Engl J Med 1999; Anesth Analg 2000; 14: 36–39. 7 Rosenthal D, Chrisant MR, Edens E et al. 341: 709–717. 26 Sprung J, Ogletree-Hughes ML, Moravec International Society for Heart and Lung 17 Bristow MR, Gilbert EM, Abraham WT CS. The effects of etomidate on the Transplantation: practice guidelines for et al. Carvedilol produces dose-related contractility of failing and nonfailing human management of heart failure in children. J improvements in left ventricular function heart muscle. Anesth Analg 2000; 91: 68–75. Heart Lung Transplant 2004; 23: 1313–1333. and survival in subjects with chronic heart 27 Nguyen NK, Magnier S, Georget G et al. 8 Arola A, Jokinen E, Ruuskanen O et al. failure. MOCHA Investigators. Circulation Anesthesia for heart catheterization in chil- Epidemiology of idiopathic cardiomyopa- 1996; 94: 2807–2816. dren. Comparison of 3 techniques. Ann Fr thies in children and adolescents. A nation- 18 Packer M, Bristow MR, Cohn JN et al. The Anesth Reanim 1991; 10: 522–528. wide study in Finland. Am J Epidemiol effect of carvedilol on morbidity and mor- 28 Sarkar M, Laussen PC, Zurakowski D et al. 1997; 146: 385–393. tality in patients with chronic heart failure. Hemodynamic responses to etomidate on 9 Nugent AW, Daubeney PE, Chondros P U.S. Carvedilol Heart Failure Study Group. induction of anesthesia in pediatric patients. et al. The epidemiology of childhood cardio- N Engl J Med 1996; 334: 1349–1355. Anesth Analg 2005; 10: 522–528.

Pediatric Anesthesia 21 (2011) 577–584 ª 2011 Blackwell Publishing Ltd 583 Heart failure in children: anesthetic implications D.N. Rosenthal and G.B. Hammer

29 den Brinker M, Hokken-Koelega ACS, children with congenital heart disease: an cardiac surgery. J Cardiothorac Vasc Anesth Hazelzet JA et al. One single dose of etomi- echocardiographic study of myocardial con- 2008; 22: 152–154. date negatively influences adrenocortical tractility and hemodynamics. Anesthesiology 36 Hammer GB, Drover DR, Cao H et al. The performance for at least 24 hours in chil- 2001; 94: 223–229. effects of dexmedetomidine on cardiac elec- dren with meningococcal sepsis. Int Care 33 Ikemba CM, Su JT, Stayer SA et al. trophysiology in children. Anesth Analg Med 2008; 34: 163–168. Myocardial performance index with 2008; 106: 79–83. 30 Donmez A, Kaya H, Haberal A et al. The sevoflurane-pancuronium versus fentanyl- 37 Hammer GB, Sam WJ, Chen MI et al. effect of etomidate induction on plasma cor- midazolam-pancuronium in infants with a Determination of the pharmacodynamic tisol levels in children undergoing cardiac functional single ventricle. Anesthesiology interaction of propofol and dexmedetomi- surgery. J Cardiothorac Vasc Anesth 1998; 2004; 101: 1298–1305. dine during esophagogastroduodenoscopy 12: 182–185. 34 Donmez A, Kizilkan A, Berksun H et al. in children. Pediatr Anesth 2009; 19: 138– 31 Hanouz JL, Massetti M, Guesne G. In vitro One center’s experience with remifentanil 144. effects of desflurane, sevoflurane, isoflurane, infusions for pediatric cardiac catheteriza- 38 Kipps AK, Ramamoorthy CR, Rosenthal and halothane in isolated human right atria. tion. J Cardiothorac Vasc Anesth 2001; 15: DN et al. Children with cardiomyopathy: Anesthesiology 2000; 92: 116–124. 736–739. complications after noncardiac procedures 32 Rivenes SM, Lewin MB, Stayer SA et al. 35 Hammer GB. Con: dexmedetomidine should with general anesthesia. Pediatr Anesth Cardiovascular effects of sevoflurane, isoflu- not be used for infants and children during 2007; 17: 775–781. rane, halothane, and fentanyl-midazolam in

584 Pediatric Anesthesia 21 (2011) 577–584 ª 2011 Blackwell Publishing Ltd Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Perioperative management of pediatric patients on mechanical cardiac support Emad B. Mossad1, Pablo Motta1, Joseph Rossano2, Brittani Hale1 & David L. Morales3

1 Division of Pediatric Cardiovascular Anesthesia, Texas Children’s Hospital, Baylor College of Medicine, Houston, TX, USA 2 Division of Pediatric Cardiology, Texas Children’s Hospital, Baylor College of Medicine, Houston, TX, USA 3 Division of Congenital Heart Surgery, Texas Children’s Hospital, Baylor College of Medicine, Houston, TX, USA

Keywords Summary end-stage heart failure; congenital heart disease; ventricular assist devices The population of children with end-stage heart failure requiring mechani- cal circulatory support is growing. These children present for diagnostic Correspondence imaging studies, various interventions and noncardiac surgical procedures Emad B. Mossad, that require anesthetic care. This article is a review of the population Division of Pediatric Cardiovascular demographics of children on mechanical cardiac support, the alternative Anesthesia, Texas Children’s Hospital, 6621 devices available, and the important concepts for safe perioperative man- Fannin Street, WT 17-417B, Houston, TX 77030, USA agement of these patients. The discussion will be limited to devices for E-mail: ebmossad@ short- and long-term cardiac support, excluding extracorporeal membrane texaschildrenshospital.org oxygenation (ECMO) for respiratory support.

Section Editor: Chandra Ramamoorthy

Accepted 19 January 2011 doi:10.1111/j.1460-9592.2011.03540.x

children, and endoventricular circular patch plasty in Introduction adults (3). Heart failure in children is a diverse condition that Only the minority of patients fail medical or surgical can result from a variety of disease states including intervention and requires the use of extracorporeal ‘simple’ and ‘complex’ congenital heart defects, vari- membrane oxygenation (ECMO) (2%) or ventricular ous forms of cardiomyopathy, and myocarditis. The assist devices (VADs) (0.7%) performed in patients condition is not rare with over 12 000 pediatric hos- with the most severe disease. The first description of pital admissions for heart failure in the United VAD use in children was in 1989 (4). In the 1990s, States annually (1). Approximately 20% of these VAD use in children was sporadic. However, owing to children have a cardiomyopathy and almost 60% improvements and miniaturization of equipment, VAD have some form of congenital heart disease implantation in children has grown exponentially in (Table 1). Among patients with congenital heart dis- recent years (5). The use of VADs is increasing in chil- ease, defects range from ‘simple’ left to right shunt- dren (6,7) and likely will continue to progress as the ing lesions found in two-thirds of patients to single availability and success of suitable devices increases ventricles found in approximately 10% of patients. (8). Similar to adults, VAD support in children could This diversity of diseases accounting for heart failure be used as a bridge to recovery, cardiac transplanta- admissions increases the complexity of care for these tion, and, eventually, as destination therapy in the patients. While medical therapy is sufficient for many foreseeable future. As these children survive longer of these patients, some form of cardiac surgical or with VAD support, awaiting organ availability, they interventional procedure is performed in one-third of will undergo multiple diagnostic imaging studies, inter- patients. These include late Fontan conversions (2), ventions and noncardiac surgical procedures that will mitral valve annuloplasty, partial in require perioperative sedation and anesthetic care (9).

Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd 585 Perioperative management of pediatric patients on mechanical cardiac support E.B. Mossad et al.

Table 1 Characteristics of children admitted for heart failure in the Table 2 Criteria for consideration of VAD as a bridge to transplan- United States tation

Heart failure patent characteristics n (%) A pediatric patient with advanced heart failure should be considered for implantation of a VAD as bridge to transplant Total pediatric admissions 2003 12 683 when the patient is listed for transplant or presumed to be a Female 6319 (50) potentially acceptable candidate for transplantation and: Congenital heart disease 7418 (59) 1. the patient has demonstrated inotrope dependency ASD, VSD, and/or PDA 4911 (66) (worsening signs/symptoms of heart failure when inotropic Single ventricle, HLHS, tricuspid atresia 848 (12) support is weaned or withdrawn) Other congenital heart disease 1660 (22) AND Cardiomyopathy 2253 (18) 2. there is evidence of compromise to at least one other organ Myocarditis 347 (3) system expected to recover with support: Arrhythmia 1827 (14) i. respiratory failure (e.g., requiring mechanical ventilation) Pulmonary hypertension 2058 (16) ii. worsening renal function (e.g., rise in serum creatinine by at Sepsis 1082 (9) least 0.3 mgÆdl)1) Acute renal failure 675 (5) iii. hepatic dysfunction (e.g., AST >50 UÆml)1, conjugated bilirubin >1.0 or INR >1.5) Data from the 2003 Kids’ Inpatient Database. iv. inability to tolerate enteral feeds ASD, atrial septal defect; VSD, ventricular septal defect; PDA, v. impaired mobility because of heart failure symptoms patent ductus arteriosus; HLHS, hypoplastic left heart syndrome. A resulting in confinement to bed person may have >1 diagnosis so the percentages will not add up OR to 100%. 3. the patient is in cardiogenic shock or impending cardiogenic shock.

Patient selection and timing of intervention VAD, ventricular assist device.

The successful management of patients on mechanical patient with acute myocarditis requiring MCS showed circulatory support (MCS) begins with patient selec- a 61% survival to discharge from the hospital (12). tion. Patients whose heart failure has progressed to Device selection is as important as candidate selec- cardiogenic shock with multi-organ system failure, tion and timing of cardiac support. If the decision for especially in the setting of chronic heart failure, have MCS has been made, then we follow the algorithm in a very poor prognosis with VADs (10). Thus, it is our Figure 1, which is based on the etiology of the refrac- strategy to place patients on support before this tory heart failure. occurs. We have developed criteria for consideration If the cause of heart failure is acute and expected to of VAD therapy (Table 2), which consist of the pres- have a transient course (i.e. myocarditis, cardiac graft ence of refractory heart failure with inotrope depen- rejection) with expected recovery in 7–14 days, then dence and at least one other organ system short-term devices such as the RotaFlow (Maquet, dysfunction. Getinge AB, Sweden) or Tandem Heart (CardiacAssist The most common cardiac indication for circulatory Inc., Pittsburgh, PA, USA) are used. Postcardiotomy support is failure to wean from cardiopulmonary bypass (CPB) after repair or palliation of complex con- genital heart disease. There is no single diagnosis asso- ciated with the need of postoperative ECMO; however, in some series, single ventricle physiology and cyanotic heart lesions more commonly require support (11). The reported frequency of ECMO use after CPB in children is 1–5%. Left ventricular assist device (LVAD) is preferable over ECMO for patients sus- pected to need support longer (>2 weeks), who do not have pulmonary hypertension or respiratory dysfunc- tion. However, for patient in acute decompensation and cardiac arrest, ECMO can be a lifesaving therapy. Viral myocarditis is the leading cause of acute heart failure in children without congenital heart disease and Figure 1 Algorithm for device selection at Texas Children’s Hospi- is the most common indication of nonsurgical MCS. A tal. ECMO, extracorporeal membrane oxygenation; BSA, body sur- recent review of the ELSO database in pediatric face area.

586 Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd E.B. Mossad et al. Perioperative management of pediatric patients on mechanical cardiac support shock can also be managed with these devices but North American Berlin EXCOR Pediatric Implant History recovery in this cohort should be expected within 90 85 77 5 days since it is well documented that support past 80 this timeframe for postsurgical dysfunction has very 70 60 poor outcomes. These short-term devices are a bridge 50 to recovery or a bridge to decision. Assist devices are 50 40 35 used as a bridge to recovery in children undergoing 30 repair of congenital heart defects where the systemic 30 ventricle may require a period of training, as following 20 the arterial switch operation in older infants with d- 10 3 0 transposition of the great arteries (13), or to allow res- <2004 2005 2006 2007 2008 2009 olution of myocardial edema or arrhythmias following IDE (Investigational Device Exemption) FDA sponsored study – May 2007 a prolonged bypass (14). The use of MCS as a bridge for postcardiotomy recovery varies significantly Figure 2 North American Berlin EXCOR Pediatric Implant history. between centers. We reported the rare use of MCS for this indication in our center over a 7-year period (6/ that presently is the most used LVAD in North Amer- 2847 bypass procedures, 0.2%) with a 50% survival ica because of its low morbidity, especially with respect (15). However, others have applied postcardiotomy to thromboembolic events. Another advantage is that MCS as a routine following Norwood stage I repairs it allows quick rehabilitation and is FDA approved for of hypoplastic left heart syndrome (average support of hospital discharge. We have begun to discharge our 3.1 days), with excellent hospital survival (89%) and patients from the hospital while they await transplant, near normal neurodevelopmental testing at 6-month which is a tremendous advantage socially and mentally follow-up (16). The bridge to decision occurs when a especially for teenagers. The device can not only be patient presents whose etiology of heart failure or its used as a bridge to transplant, but since early 2010, it time course is unknown or their neurological status is has been used as a destination therapy (bridge to desti- in question. The bridge to decision children either nation). We have implanted the device in adolescents recover so the VAD can be explanted or they do not ranging from 12 to 17 years old, all of whom have sur- recover and are transitioned to a long-term device if vived to transplant and discharge. they are a transplant candidate (bridge to bridge strat- egy) or weaned to maximal medical support if they are Types of devices available for pediatric cardiac not a transplant candidate. support If the cause of heart failure is chronic such as long- standing dilated cardiomyopathy or end-stage congeni- The use of MCS at Texas Children’s Hospital has tal heart disease, these patients are implanted with a increased dramatically over the past 15 years long-term device as a bridge to transplant (17). For (Figure 3) with a changing makeup from a predomi- children with a body surface area (BSA) < 1.3 m2, the nance of ECMO in the early era to an increase in the Berlin EXCOR VAD (Berlin Heart AG, Berlin, Ger- use of short- and long-term VADs such as the Tandem many) is implanted. This is an air actuated paracorpo- Heart and Berlin Heart EXCOR in the current era. real VAD that is the only VAD being used in North Our center has favored the use of LVADs and avoid- America developed specifically for children. It can uniquely support children of all sizes and ages from CV ECLS Patients at TCH: 1995-2009 2.8-kg neonates to adult size children because it is available in six sizes. There are over 700 implants worldwide and over 300 in North America alone (Figure 2). The EXCOR is towards the end of its FDA IDE trial as a bridge to transplant only and has completed cohort one that are patients with a BSA < 2

0.7 m . The Berlin EXCOR allows for children to be Number of Cases mobile and rehabilitate physically and nutritionally as well as continue their development (18). For patients with a BSA > 1.3 m2, we implant the Heart Mate II (Thoratec Corporation, Pleasanton, CA, USA), which Figure 3 A 15-year experience of pediatric extracorporeal support is an intracorporeal small, noiseless axial flow device in one institution.

Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd 587 Perioperative management of pediatric patients on mechanical cardiac support E.B. Mossad et al. ing the use of a right sided device (RVAD) or ECMO whenever possible. This is based on the concept that even poorly functioning right ventricles can allow for home discharge home discharge

filling of the LVAD if timing of support and manage- Ambulation ment is optimized, and pulmonary vascular resistance (PVR) is managed appropriately (19). Additionally, LVADs with adequate management of RV function ) and PVR appear to have a survival advantage over the 2 Body surface area range (m

use of a biventricular device (BiVAD) at least among ) 1 adult patients (10). Although we know from experience ) min in care for children with single ventricular physiology Æ Variable 0.2–1.3 Yes Flow range (L that patients can survive for years without a subpul- monic pumping chamber (e.g. Fontan patients), a sin- gle ventricle is anatomically and physiologically different from a failing, dilated RV in the VAD set- ting. Varying degrees of right ventricular dysfunction (60,80 in Europe only) 6000–15000 rpm >2.5 >1.4 Yes Possible are commonly present in patients with advanced heart SV (ml) or speed (rpm) failure and in the absence of mechanical support for the right ventricle, much of our medical therapy is aimed at support of the right ventricle (20). Devices range from short-term acute support devices to long-term support with varied technology and flow from the impeller characteristics. The most commonly used devices in North America are summarized in Table 3.

Short-term MCS ECMO The most commonly used device in acute postcardioto- my or cardiac arrest states, an ECMO circuit is com- posed of a centrifugal or roller pump with a hollow fiber membrane oxygenator, oxygen blender, pump console and heat exchanger. Additional information about the ECMO equipment components is available at the ELSO web site: http://www.elso.med.umich.edu/.

The RotaFlow A centrifugal pump without an oxygenator was sus- pended in three magnetic fields, allowing minimal fric- tion and heat production. This results in almost immediate laminar flow and minimal damage to blood cells or activation of blood components. It is placed cen- trally with the inflow cannula in the left atrium and the Short Rotational Continuous Electric Vacuum assisted 3000–7500 rpm < 5 >1.3 No outflow cannula in the aorta. This is our preferred device for patients needing temporary MCS who are <40 kg although it can be applied to patients of all sizes. extracorporeal Intracorporeal Long term Axial Non-pulsatile Electric Vacuum assisted 7500–12500 rpm 1–10 >0.7–<1.5 Yes Tandem heart

For patient >40 kg especially those with prior sternot- Most commonly used pediatric ventricular assist devices in North America omy, the Tandem Heart is the preferred device. The DeBakey MicroMed Thoratec ExtracorporealHeart Mate Long II term Intracorporeal Long term Pusher plate Axial Pulsatile Pneumatic Non-pulsatile Vacuum assisted Electric 65 ml Kinetic energy <6 >0.7 Yes Possible RotaFlowLevitronixTandem Extracorporeal Heart Percutaneous Extracorporeal Short <14 daysBerlin Short Heart Rotational/centrifugal Extracorporeal Continuous Long term Electric Rotational/centrifugal Continuous Pusher plate Vacuum Electromagnetic assisted Vacuum 0–4500 assisted rpm 0–5500 rpm Pulsatile 0.5–6 <1.5 No Pneumatic minimum No <0.5 Vacuum assisted 10,25,30,50 ml No Tandem Heart has favorable fluid dynamics because of Table 3 Device Position Duration of support Device type Flow type Power source Fill

588 Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd E.B. Mossad et al. Perioperative management of pediatric patients on mechanical cardiac support its hydrodynamic fluid bearing that supports the spin- ning rotor. The Tandem Heart is placed percutaneously through the femoral vessels. The Transseptal Extended Flow Cannula is placed through the femoral vein in the catheterization laboratory via a transseptal puncture into the left atrium. The arterial side is placed in the femoral artery. For children under 80 kg, we sew a vascular graft to the femoral artery and cannulate the graft so that the cannula never enters the artery itself, thus avoiding ischemia to the extremity.

Long-term MCS See Figure 4.

Figure 5 Berlin EXCOR device in situ. Reflective mirror to examine Berlin EXCOR chamber for fibrin strands or clots. Available in several pumping chamber sizes (10– 80 ml), Berlin Heart VAD or EXCOR provides pulsa- tion in the pump or cannulae, the pump must be tile flow delivered through a pneumatically driven thin exchanged to avoid systemic embolization. The blood- membrane pump (14). In , the pump moves contacting surfaces of the pump including the polyure- compressed air into the diaphragm causing the ejec- thane valves are heparin coated through the Carmeda tion. In diastole, negative pressure is generated to aid process, reducing anticoagulation requirements. It also in the filling of the chamber. The maximum systolic has silicon cannulae with a Dacron covering that positive pressure generated is 350 mmHg and the max- works as a biologic barrier against ascending infections imum negative driving pressure is minus 100 mmHg. It (Figure 5). When using this, VAD patients are not is a fixed volume, variable rate device. The pump rate dependent on mechanical ventilation, can resume ent- can be adjusted between 30 and 150 beatsÆmin)1, and eral feeding, are ambulatory and can be discharged the systolic time can be adjusted between 20% and from the ICU. The rate of neurologic complications 70% of the cycle. The blood pump is transparent, of 5% is lower than ECMO and other adult-sized allowing visual inspection of filling, emptying and VADs (21). thrombus formation. If there is any thrombus forma- Heart Mate II Heart Mate II is an axial-flow LVAD. Axial flow (a) (c) devices are smaller and simpler than pulsatile pumps. (b) They have only one moving component, with no valves, vent or compliance chamber reducing the com- plexity. Because of their small size, axial-flow LVADs can be used in smaller patients, BSA ‡ 1.4 m2. Frazier published the first series using Heart Mate II in adult and teenage patients with an 81% survival rate (22).

Thoratec Thoratec VAD (Thoratec Corporation) is a paracorpo- real pneumatic VAD. The prosthetic ventricle has a 65-ml stroke volume chamber with a maximum output flow of 7 LÆmin)1. It works on three different modes: fixed rate, synchronous (for weaning) or fill-to-empty mode. Like other paracorporeal pneumatic devices, it Figure 4 Pediatric long-term VAD support; a, Thoratec; b, Heart can be used as a LVAD or Biventricular VAD. The Mate II; c, Berlin EXCOR. lower limits for implantation are a BSA of 0.7 m2 and

Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd 589 Perioperative management of pediatric patients on mechanical cardiac support E.B. Mossad et al. an age of 7 year or older. Reinhartz reported a 68% procedure (25,26). In a study of 42 survivors of MCS survival to transplant or recovery in 209 pediatric from our institution, 36% had moderate or severe patients (range 5–18 years, BSA 0.7–2.3 m2) (23). The motor or cognitive impairment at a median follow-up risk of thromboembolism in children is higher than in of 33 months after weaning from support (26). adults because of the reduced flow velocities and blood Almost all children on MCS are at increased risk of stasis in the device owing to mismatch of the pump consumption coagulopathy, and many are on a regi- size to patient. men of anticoagulant therapy ranging from unfraction- ated heparin, Enoxaparin, anti-platelet therapy (Dipyridamole or Aspirin) or Coumadin (27). Coagula- Noncardiac procedures in children on mechanical tion studies, medications, and blood transfusion cardiac support requirements should be carefully considered preopera- Children on MCS have a similar survival to transplan- tively. tation and improved functional class similar to those transplanted without need for a bridge device (24). As Intraoperative considerations the survival and duration of support improves, these patients require noncardiac interventions (9). In our The response of children on MCS to various induction series of Berlin EXCOR implants, 21 patients (mean and maintenance anesthetic regimens has been exam- age 42.7 months) underwent 62 noncardiac procedures ined in several case series (28–30). Hypotension (range 1–10 per patient), including imaging studies, ot- occurred in 69% of a series of 11 Berlin EXCOR olaryngologic, general surgical and neurosurgical pro- supported children, regardless of medication used and cedures and placement of PICC lines. The most even at subtherapeutic doses (30). Preoperative stability common procedure was device exchange owing to was not predictive of intraoperative hypotension. In detection of fibrin strands or clots. The risk of requir- fact, only 17% of patients on preoperative inotropes ing a procedure correlated with the duration of sup- experienced hypotension on induction of anesthesia port and not the device size. compared o 41% of device supported children on no inotropes. The most important principle is the narrow therapeutic window of these patients’ tolerance to Anesthetic considerations changes in venous or arterial capacitance using any The perioperative care and management strategy are induction agent. Ketamine induction and remifentanil dependent on the patient’s preoperative condition and infusion maintenance had the lowest incidence of hypo- comorbidities, intraoperative considerations and type tension. Hypotension was corrective with a fluid bolus of device placed. (10 mlÆkg)1) in 60% of patients and an alpha receptor agonist (phenylephrine 1–5 mcgÆkg)1 bolus, norepi- nephrine, or vasopressin infusions). Patient condition Similar to the Thoratec, the Berlin EXCOR has a These patients present for procedures with their under- fixed chamber size, and both are considered fixed vol- lying cardiac condition and its impact on end organ ume, variable rate devices. The rate of these devices is function, most importantly hepatic and renal impair- set (fixed) but can be adjusted and changed to improve ment. Although the use of mechanical support has perfusion. The Thoratec device will slow down in the clearly been successful in improving the medical condi- presence of hypovolemia to allow the device chamber tion of these children, the use is not without the risk (65 ml) to fill. As stroke volume is significantly volume of significant morbidities that will affect the delivery of dependent in these devices, the most effective therapy an anesthetic. The main risks appear to be central ner- for hypotension is fluid bolus and alpha-receptor agon- vous system thromboembolic and bleeding; though, ists (31). Consideration for the effect of temperature, infections, abdominal complications and mechanical regional anesthetic techniques, and patient positioning issues with the device are not trivial. While some stud- on volume status are essential intraoperatively. ies have demonstrated good neurological outcomes in Increased pulmonary vascular resistance and right selected populations early after mechanical support the ventricular failure should always be considered and overall long-term neurocognitive outcomes of these treated promptly in the presence of unexplained hypo- patients is still largely unknown. However, limited data tension (32,33). Interventions to improve RV function from ours and other centers has demonstrated a high include maintaining a mildly alkalotic environment, incidence of impairments that should be considered use of phosphodiesterase inhibitors (milrinone), diuret- before the delivery of an anesthetic for a noncardiac ics, and inhaled nitric oxide (iNO) or other pulmonary

590 Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd E.B. Mossad et al. Perioperative management of pediatric patients on mechanical cardiac support

Table 4 Management of perioperative hemodynamic changes

Hemodynamic Device change Possible etiology Device inspection Possible intervention

Berlin Heart Hypotension Hypovolemia Inspect chamber (reflective mirror) Fluid bolus (10 mlÆkg)1) EXCOR for wrinkling and incomplete fill Consider and treat promptly in diastole RV failure or › pulmonary vascular resistance (PVR) Increasing device rate not recommended as it will shorten fill time Hypotension Systemic vasodilation Chamber may be filling fully Alpha agonists (phenylephrine, (e.g. with induction norepinephrine) or vasopressin of anesthesia) Hypertension Pain Inspect chamber for wrinkling Avoid increasing systolic driving during systole pressure >195 mmHg Awareness under Sedation, analgesia anesthesia Anxiety Titrate vasodilators (nitroprusside) Heart Mate II Hypotension Hypovolemia Device flow will decrease (lowest Heart Mate II flow dependent on 3LÆmin)1) and rpm will slow rpm and DP (Aortic-LV pressure), (lowest 8000) beyond which highly preload sensitive device will shut until volume Fluid bolus (10 mlÆkg)1) status corrected Consider and treat promptly RV failure or › PVR Hypotension Systemic vasodilation Lower 4P (Aortic-LV pressure) Alpha agonists (phenylephrine, (e.g. with induction will increase device flow with norepinephrine) or vasopressin of anesthesia) no change in rpm to avoid excess suction and collapse of LV around inflow cannula Hypertension Pain flfl device flow with M in rpm Sedatives, analgesics Awareness, anxiety Titrate vasodilators (nitroprusside) . Spontaneous ventilation lowers alveolar pres- evaluate the presence of residual intracardiac shunting sure and PVR, improves venous return and hemody- (a PFO) and assess the risk of systemic embolism. namic stability, and should be considered if the Strict sterile techniques and bacterial endocarditis procedure allows. prophylaxis should be adhered to for all invasive pro- Transesophageal echocardiography (TEE) is a useful cedures and whenever handling the devices. Blood guide not only during the initial device insertion, but stream infections can be fatal in these patients. Given also in other noncardiac procedures if applicable (34). the profound effects that nosocomial infections have Valvular disease can significantly affect device func- on morbidity, mortality, and cost (35,36), all attempts tion. Mitral stenosis will limit device filling, while aor- are made to avoid central venous access if not needed tic or mitral insufficiency will cause back flow to the for the procedure and to remove invasive lines as soon heart, thus limiting forward output and impeding LV as possible. Placement of central venous lines must be unloading, resulting in LV distension, ventricular sep- carried out with careful positioning and caution to tal shifting, and RV failure. An adequately functioning avoid the risk of venous air embolism from entrained device should result in minimal opening of the aortic air especially in patients with a vacuum-assisted dia- valve, a decompressed LV chamber, and a midline sep- stolic filling device. tal position. TEE can also identify inflow cannula posi- tion, which should be aligned with the mitral inflow to Device type produce laminar flow into the cannula. A peak mitral inflow velocity >2.3 mÆs)1 is indicative of possible Understanding device technology is essential for man- inflow cannula obstruction. RV function and PVR can aging children on MCS for noncardiac procedures. An also be assessed using TEE evaluation of septal posi- appreciation of the various changes in the device out- tion, chamber size, and degree of tricuspid insuffi- put and the manipulations that can be made will help ciency. Finally, TEE can also detect air entrainment in treating hemodynamic changes during the delivery through suture lines if the LV completely collapses and of a sedative or anesthetic. A perfusionist should

Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd 591 Perioperative management of pediatric patients on mechanical cardiac support E.B. Mossad et al. accompany the patient during transport and remain gists will become more involved in the care for these accessible during the procedures. Table 4 describes patients not only for device insertion or transplants some of the changes and interventions with the two but also for more noncardiac procedures, the longer most common long-term pediatric MCS devices in our and better they survive. A better understanding of practice. patient demographics, device selections, and response In conclusion, the population of children with end- to anesthetics is essential for the safe care for these stage heart failure on MCS is growing. Anesthesiolo- children.

References 1 Rossano JW, Zafar F, Graves DE et al. 12 Rajagopal SK, Almond CS, Laussen PC mate II axial-flow left ventricular assist Prevalence of heart failure related hospital- et al. Extracorporeal membrane oxygenation device. Tex Heart Inst J 2007; 34: 275–281. izations and risk factors for mortality in for the support of infants, children, and 23 Reinhartz O, Keith FM, El-Banayosy A pediatric patients: an analysis of a nation- young adults with acute myocarditis: a et al. Multicenter experience with the wide sampling of hospital discharges. Circu- review of the Extracorporeal Life Support Thoratec ventricular assist device in children lation 2009;120:S586 (Abstract). Organization registry. Crit Care Med 2010; and adolescents. J Heart Lung Transplant 2 Mavroudis C, Backer CL, Deal BJ. Late 38: 382–387. 2001; 20: 439–448. reoperations for Fontan patients; state of 13 Naughton P, Mossad E. Retraining the left 24 Lin MH, Chou NK, Chen YS et al. Out- the art invited review. Eur J Cardiothorac ventricle after arterial switch operation: come in children bridged and nonbridged to Surg 2008; 34: 1034–1040. emerging uses for the Left Ventricular Assist cardiac transplantation. Transplant Proceed- 3 Suma H, Isomura T, Horii T et al. Non- Device in Pediatric cardiac surgery. J Car- ings 2010; 42: 916–919. transplant cardiac surgery for end-stage car- diothorac Vasc Anesth 2000; 14: 454–456. 25 Lequier L, Joffe AR, Robertson CM et al. diomyopathy. J Thorac Cardiovasc Surg 14 Zhu DM, Wang W, Chen H et al. Left ven- Two-year survival, mental, and motor 2000; 119: 1233–1244. tricular assist devices for pediatric postcardi- outcomes after cardiac extracorporeal life 4 Frazier OH, Bricker JT, Macris MP et al. otomy cardiac failure. ASAIO J 2006; 52: support at less than five years of age. Use of a left ventricular assist device as a 603–604. J Thorac Cardiovasc Surg 2008; 136: 976– bridge to transplantation in a pediatric 15 Undar A, McKenzie ED, McGarry MC 983. patient. Tex Heart Inst J 1989; 16: 46–50. et al. Outcomes of congenital heart surgery 26 Morris SA, Noll LM, Rossano JW et al. 5 Blume ED, Naftel DC, Bastardi HJ et al. patients after extracorporeal life support at Neurodevelopmental outcomes in pediatric Outcomes of children bridged to heart Texas Children’s Hospital. Artif Organs cardiac survivors of mechanical circulatory transplantation with ventricular assist 2004; 28: 963–966. support. J Am Coll Cardiol 2009; 53: A359 devices: a multi-institutional study. Circula- 16 Ungerleider RM, Shen I, Yeh T et al. Rou- (Abstract). tion 2006; 113: 2313–2319. tine mechanical ventricular assist following 27 Stiller B, Lemmer J. Consumption of blood 6 Hetzer R, Potapov EV, Stiller B et al. the : improved neuro- products during mechanical circulatory Improvement in survival after mechanical logic outcome and excellent hospital sur- support in children: comparison between circulatory support with pneumatic pulsatile vival. Ann Thorac Surg 2004; 77: 18–22. ECMO and a pulsatile ventricular assist ventricular assist devices in pediatric 17 Morales DLS, Zafar F, Rossano JW et al. device. Intensive Care Med 2004; 30: 1814– patients. Ann Thorac Surg 2006; 82: 917– Use of ventricular assist devices in children 1820. 924. across the US: analysis of 7.5 million pediat- 28 El-Magharbel I. Ventricular assist devices 7 Duncan BW. Pediatric mechanical circula- ric hospitalizations. Ann Thorac Surg 2010; and anesthesia. Semin Cardiothorac Vasc tory support in the United States: past, 90: 1313–1319. Anesth 2005; 9: 241–249. present, and future. ASAIO J 2006; 52: 18 Imamura M, Dossey AM, Prodhan P et al. 29 Stone M, Soong W, Krol M et al. The anes- 525–529. Bridge to cardiac transplant in children: thetic considerations in patients with 8 Hetzer R, Stiller B, Potapov E et al. Ven- Berlin Heart versus extracorporeal mem- ventricular assist devices presenting for non- tricular assist device and mechanical circula- brane oxygenation. Ann Thorac Surg 2009; cardiac surgery: a review of eight cases. tory support for children. Current Opin 87: 1894–1901. Anesth Analg 2002; 95: 42–49. Organ Transplant 2007; 12: 522–528. 19 Gandhi SK, Huddleston CB, Balzer DT 30 Cave D, Fry K, Buchholz H. Anesthesia for 9 Stehlik J, Nelson DM, Kfoury AG et al. et al. Biventricular assist devices as a bridge noncardiac procedures for children with Outcome of noncardiac surgery in patients to heart transplantation in children. Circula- Berlin Heart EXCOR Pediatric Ventricular with ventricular assist devices. Am J Cardiol tion 2008; 118: S89–S93. Assist Device: a case series. Pediatr Anesth 2009; 103: 709–712. 20 Ochiai Y, McCarthy PM, Smedira NG et al. 2010; 20: 647–659. 10 Holman WL, Kormos RL, Naftel DC et al. Predictors of severe right ventricular failure 31 Mudge G, Fang J, Smith C et al. The physi- Predictors of death and transplant in after implantable left ventricular assist ologic basis for the management of ventricu- patients with a mechanical circulatory sup- device insertion: analysis of 245 patients. lar assist devices. Clin Cardiol 2006; 29: port device: a multi-institutional study. J Circulation 2002; 106: I198–I202. 285–289. Heart Lung Transplant 2009; 28: 44–50. 21 Rockett SR, Bryant JC, Morrow WR et al. 32 Kavarana MN, Pessin-Minsley MS, Urtecho 11 Aharon AS, Drinkwater DC, Churchwell Preliminary single center North American J et al. Right ventricular dysfunction and KB et al. Extracorporeal membrane oxygen- experience with the Berlin Heart pediatric organ failure in left ventricular assist device ation in children after repair of congenital EXCOR device. ASAIO 2008; 54: 479–482. recipients: a continuing problem. Ann Tho- cardiac lesions. Ann Thorac Surg 2001; 72: 22 Frazier OH, Gemmato C, Myers TJ et al. rac Surg 2002; 73: 745–750. 2095–2102. Initial clinical experience with the Heart-

592 Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd E.B. Mossad et al. Perioperative management of pediatric patients on mechanical cardiac support

33 Kormos RL, Teuteberg JJ, Pagani FD et al. examination for ventricular assist device 36 Holman WL, Kirklin JK, Naftel DC et al. Right ventricular failure in patients with the implantation. Anesth Analg 2007; 105: 583– Infection after implantation of pulsatile Heart Mate II continuous flow left ventricu- 601. mechanical circulatory support devices. J lar assist device; incidence, risk factors and 35 Grisaru-Soen G, Paret G, Yahav D et al. Thorac Cardiovasc Surg 2010; 139: 1632– effect on outcomes. J Thorac Cardiovasc Nosocomial infections in pediatric 1636. Surg 2010; 139: 1316–1324. cardiovascular surgery patients: a 4 year 34 Chumnanvej S, Wood MJ, MacGillivray TE survey. Pediatr Crit Care Med 2009; 10: et al. Perioperative echocardiographic 202–206.

Pediatric Anesthesia 21 (2011) 585–593 ª 2011 Blackwell Publishing Ltd 593 Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Perioperative management of low birth weight infants for open-heart surgery Glyn D. Williams1 & Ronald S. Cohen2

1 Department of Anesthesia, Lucile Packard Children’s Hospital, Stanford University, Stanford, CA, USA 2 Department of Pediatrics, Intermediate and Special Care Nurseries, Lucile Packard Children’s Hospital, Stanford University, Stanford, CA, USA

Keywords Summary low birth weight; prematurity; heart surgery; cardiopulmonary bypass; infants Infants of birth weight £2500 g are termed low birth weight (LBW). These children often have considerable morbidity from prematurity and intra- Correspondence uterine growth restriction. Additionally, LBW infants have increased risk Glyn D. Williams, for cardiac and noncardiac congenital anomalies and may require surgery. Associate Professor, Department of Primary rather than palliative surgical repair of cardiac lesions has been Anesthesia, Stanford University Medical preferred in recent years. However, LBW remains a risk factor for Center, 300 Pasteur Drive, H3587, Stanford, CA 94305-5740, USA increased mortality and morbidity after open-heart surgery (OHS). There is Email: [email protected] a paucity of information about the anesthetic challenges presented by LBW infants undergoing OHS. This review summarizes the perioperative issues Section Editor: Chandra Ramamoorthy of relevance to anesthesiologists who manage these high-risk patients. Emphasis is placed on management concerns that are unique to LBW Accepted 23 October 2010 infants. Retrospective data from the authors’ institution are provided for those aspects of anesthetic care that lack published studies. Successful doi:10.1111/j.1460-9592.2011.03529.x outcome often requires substantial hospital resources and collaborative multi-disciplinary effort.

and have CHD have associated extracardiac anomalies, Low birth weight infants syndromes, or chromosomal anomalies (7,8). The term low birth weight (LBW) refers to infants When compared to term infants, LBW/premature born at a weight less than or equal to 2500 g. This infants were more likely to have pulmonary atresia group includes infants born preterm (gestational age with ventricular septal defect, complete atrioventricular <37 weeks) and those born at term but who are small septal defect, coarctation of the aorta, tetralogy of Fal- for gestational age (SGA, defined as <3rd percentile lot, and pulmonary valve stenosis. Fewer LBW/pre- for weight). Many SGA babies suffer from intrauterine term infants had pulmonary atresia and intact growth restriction (IUGR). Risk factors for prematu- ventricular septum, transposition of the great arteries, rity and SGA are similar, and up to 50% of extremely and single ventricle (2). preterm neonates are SGA (1). Survival rates for open-heart surgery (OHS) in neo- Premature babies are more that twice as likely to nates have improved substantially over the last have cardiovascular malformations than infants born 10 years (9–12). However, LBW remains a risk factor at term (2), and infants with congenital heart disease for increased mortality after OHS (8,13–20), and hos- (CHD) have a high incidence (approximately 16%) of pital survival rates of LBW infants undergoing heart LBW (2–5). surgery range from 73% to 90% (8,14–21). Infants with CHD have an incidence of 13–20% of Perioperative management of these high-risk infants associated congenital anomalies or genetic syndromes must take into account the patient’s small size, the clin- (4–6), and many of these conditions are associated with ical implications of prematurity, IUGR and associated IUGR. Thus, up to 30% of infants who are of LBW extracardiac anomalies, the cardiac pathophysiology

538 Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd G.D. Williams and R.S. Cohen Open-heart surgery for low birth weight infants and issues related to anesthesia, surgery and cardiopul- Table 1 Patient demographics (n = 65) monary bypass (CPB) (22). Variable Valuea We were unable to find reports of the anesthetic challenges presented by LBW infants undergoing OHS. Demographics Therefore, we describe our approach at Stanford Uni- LBW infants (>1.5–2.5 kg) 56 (86) versity’s Lucile Packard Children’s Hospital and pres- Very LBW infants (>1.0–1.5 kg) 6 (9) Extremely LBW infants (£1.0 kg) 3 (5) ent data from a retrospective review of our experience. Female gender 38 (58) The medical records of all children who underwent Weight at birth (kg, mean ± SD) 1.92 ± 0.51 OHS during the period January 1, 2002 to April 30, Weight at surgery (kg, mean ± SD) 2.02 ± 0.44 2006 were examined after obtaining Institutional Born at study institution 13 (20) Review Board approval and waiver of consent. Age at admission to study institution 8±14 Patients weighing 2.5 kg or less at the time of surgery (days, mean ± SD) were included in the study. Data concerning patient Age at surgery (days, mean ± SD)18±20 Pregnancy and birth demographics, antenatal, and hospital course were Maternal age (years, mean ± SD) 28.7 ± 6.4 recorded, with special attention to the perioperative Parity (mean ± SD)2±1 period. Prenatal maternal complications 28 (43) Fetal diagnosis of congenital heart disease 19 (29) Premature infants (<37 weeks gestation) 37 (57) Timing and type of surgery Gestational age (weeks, mean ± SD)35±3 In the past, OHS was intentionally delayed for LBW Small for gestational age (SGA) 28 (43) Premature + SGA 11 (17) infants. If surgical intervention was necessary, pallia- Twin or triplet 18 (28) tive procedures were performed rather than more Apgar score at 1 min (mean ± SD)7±2 extensive, complete repairs. The intention was to wait Apgar score at 5 min (mean ± SD)8±1 for the infant to attain greater weight and a more Genetic or congenital anomaly 16 (24) mature gestational age. These approaches have not LBW, low birth weight. been shown to improve outcome, and some studies aValues expressed as n (%) unless stated otherwise. suggest that intervening mortality is higher (8), and significant morbidities are associated with the delay (20). Nowadays, early primary repair is often preferred because an unfavorable in utero environment further (13–17,22,23). compromises the patient’s immature organs (1). IUGR increases the incidence of respiratory distress syndrome (RDS), bronchopulmonary dysplasia (BPD), and reti- Preoperative concerns nopathy of prematurity (ROP) in premature babies Therapeutic strategies to optimize the patient’s cardio- (24–28). Tables 2 and 3 list the pathologies encoun- vascular status are lesion-specific and are as relevant to tered commonly in LBW infants. LBW infants, as they are to term infants of normal weight. The preoperative challenges unique to LBW Pulmonary issues infants are largely because of comorbidity. A retrospec- tive review of 105 LBW infants undergoing cardiac sur- Preterm infants have arrested lung development with gery with or without CPB (patients with isolated patent impaired alveolarisation and dysmorphic vasculogenesis ductus arteriosus were excluded) found median birth (29). Reduction in functional residual capacity (FRC) is weight was 2130 g (range 431–2500 g), and median independently associated with prematurity, IUGR, and gestational age was 36 weeks (range 25–41.1 weeks). Of severity of BPD (30). Premature infants with surfactant the group, 64% had been born prematurely, 27% were deficiency develop RDS; BPD is the end result of a vari- SGA, and 32% had associated major congenital anom- ety of insults to the lungs, including lung immaturity, alies (22). The Stanford series (all patients undergoing surfactant deficiency, oxygen toxicity, infection, poor OHS) was similar (Table 1). nutrition, and mechanical ventilation. Improvements in ventilator management strategies, widespread use of sur- factant, and prenatal administration of glucocorticoids Concerns related to prematurity and SGA have improved survival of premature infants and hence Infants who are premature or SGA have increased risk the prevalence of BPD (31,32). CHD results in a slower of morbidity and mortality. The risk is even greater rate of recovery of pulmonary function in patients with for those who are premature and SGA, presumably BPD (33), and patients with CHD and BPD have signifi-

Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd 539 Open-heart surgery for low birth weight infants G.D. Williams and R.S. Cohen

Table 2 Premature infants: postnatal clinical concerns Table 3 Small for gestational age infants: Postnatal clinical con- cerns Small size Procedures technically challenging, specialized equipment Perinatal asphyxia; sequelae include required Hypoxic-ischemic encephalopathy, seizures Respiratory system Aspiration pneumonia Birth asphyxia Myocardial dysfunction Respiratory distress syndrome Acute tubular necrosis of the kidney Chronic lung disease (bronchopulmonary dysplasia) Necrotizing enterocolitis Apnea of prematurity Liver dysfunction Subglottic stenosis Adrenal insufficiency because of hemorrhage Increased incidence of congenital respiratory system anomalies Disseminated intravascular coagulation Cardiovascular system Meconium aspiration syndrome Persistent fetal circulation Persistent pulmonary hypertension of the newborn Patent ductus arteriosus Pulmonary hemorrhage Persistent pulmonary hypertension Hypoglycemia Increased incidence of congenital cardiovascular anomalies Hyperglycemia Central nervous system Polycythemia/hyperviscosity Intraventricular hemorrhage Temperature instability Hypoxic-ischemic injury before, during, and after birth Retinopathy of prematurity Modified from Rosenberg A (1). Neurological injury at lower bilirubin levels Gastrointestinal system period until their pulmonary arterial and alveolar func- Jaundice from reduced hepatic conjugation of bilirubin tion has had an opportunity to recover is preferable to Necrotizing enterocolits, possibly resulting in short gut syndrome an early corrective procedure. However, palliation, Hepatotoxic agents (e.g., parenteral nutrition) including banding of the pulmonary trunk or shunting Increased risk of congenital gastrointestinal anomalies Genitourinary system the pulmonary artery in the face of labile pulmonary Fluid and electrolyte abnormalities vasculature, can pose serious problems in regulating pul- Increased exposure to hypoxia, hypoperfusion and nephrotoxic monary blood flow. Patients with BPD and univentricu- agents lar hearts have a high risk of mortality (34). Abnormal Increased risk of congenital genitourinary anomalies capillary distribution and the presence of aortopulmo- Infection nary collaterals in children with BPD may have signifi- Increased risk of infection from mother, indwelling catheters, cant implications for regulating pulmonary blood flow endogenous gut flora, hospital environment Nutrition and metabolic in palliated univentricular hearts (34). Increased risk of hypo- and hyperglycemia Lung disease complicates the perioperative manage- Vitamin K, iron deficient ment of the infant with CHD as follows: Reduced serum bicarbonate, therefore reduced buffering of acid Temperature and humidity Surfactant Increased risk of hypothermia and hyperthermia Premature infants with RDS and ongoing ventilator Prone to dehydration from transdermal water loss and oxygen requirements may require repeat doses of Hematology Anemia (from prematurity, blood sampling, iron deficiency, surfactant. In some cases, it may be difficult to deter- disease-e.g., sepsis) mine whether the continued respiratory dysfunction is Polycythemia (cyanotic heart disease) primarily pulmonary or cardiac in origin. Adverse Thrombosis (intravascular catheters, dehydration, low cardiac effects of surfactant administration include pulmonary output) hemorrhage (particularly in patients with left-to-right Bleeding (vitamin K deficient, thrombocytopenia, disease-e.g., shunt via a ductus arteriosus), endotracheal tube sepsis) (ETT) obstruction, sinus bradycardia, blood pressure Endocrinology Insufficiency of thyroid and adrenal hormones swings, and desaturation (35). The benefits from sur- Pharmacology factant therapy need to be weighed against the risks of Polypharmacy with potential for drug toxicity, adverse reactions, treatment. Premature infants are at risk for surfactant drug interactions, and medication errors dysfunction, even after the initial RDS has resolved (36). Additionally, CPB may result in surfactant dys- cant morbidity and mortality compared with CHD function in infants (37,38). It is unclear whether peri- patients without BPD (34). It is unclear whether palliat- operative surfactant therapy improves patient outcome ing high-risk patients [increased pulmonary vascular (35). At Stanford, it is reserved for instances of severe resistance (PVR), severe BPD] in the early postnatal perioperative RDS.

540 Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd G.D. Williams and R.S. Cohen Open-heart surgery for low birth weight infants

Lung compliance permit (35). PEEP or continuous positive airway pres- Surfactant therapy, resolving RDS, and evolving BPD sure helps maintain FRC, preventing atelectasis, alter pulmonary compliance and thereby affect pulmo- improve ventilation/perfusion matching, and stabilize nary blood flow. This complicates the maintenance of the premature’s highly compliant chest wall. However, balanced pulmonary and systemic circulations in caution is necessary because overdistension of the lung patients with cardiac shunt. The inflammatory response will impede venous return, increase PVR, and potenti- elicited by OHS may adversely affect a previously dam- ate lung injury. Unfortunately, the prolonged ventila- aged lung (39). LBW infants may have limited defenses tor support required by patients with BPD increases to free radical damage sustained during CPB (40). the risk of subsequent morbidity (34). Apnea of prematurity is common in infants <32 Airway resistance gestational age and may complicate postoperative Elevation in airway resistance may be because of bron- withdrawal of mechanical ventilation. Of note, caf- chial vessel compression, bronchial compression by feine, initially used to treat apnea of prematurity, has enlarged pulmonary arteries, ischemic tracheal seg- been shown to decrease the incidence of BPD and ments after surgery to unifocalize aortopulmonary col- improve neurological outcome (58). laterals, accumulation of secretions and bronchial reactivity, particularly after CPB (39,41–49). The air- Carbon dioxide way may be compromised by congenital anomalies Permissive hypercapnia is a common mechanical ventila- (e.g., complete tracheal rings) or acquired subglottic tion strategy to limit BPD in LBW infants. However, stenosis from previous endotracheal intubations. hypercapnia may perturb hemodynamics by inducing pulmonary vasoconstriction, systemic vasodilation, and Mechanical ventilation decreased cardiac function. This may be especially rele- The stresses imposed on premature lungs by mechani- vant post-CPB, when there may already be elevated cal ventilation are implicated in the arrest of normal PVR and myocardial dysfunction. Preoperatively, per- lung growth seen with RDS and the severity of subse- missive hypercapnia is preferentially employed at our quent BPD. For neonatologists, a primary goal during institution to balance systemic and pulmonary blood mechanical ventilation is the avoidance of ventilator- flow in patients with mixing lesions because it augments induced lung injury (VILI). The strategies employed in cerebral blood flow (59). Most investigators believe that most NICUs emphasize stabilizing FRC with positive permissive hypercapnia within reasonable limits (e.g., end-expiratory pressure (PEEP), minimizing ‘volutrau- PaCO2 < 55 initially, slightly higher if adequate pH ma’ by using small tidal volumes (e.g., 4–5 mlÆkg)1), compensation) (60,61) does not increase the risk of minimizing inspiratory times (0.35 s or less) to avoid intraventricular hemorrhage (IVH). However, extreme an I : E-ratio >1 : 1, and increasing rate as needed to hypercapnia, and marked swings in PaCO2, may be asso- maintain a PaCO2 consistent with mild ‘permissive ciated with increased IVH rates (62,63). Both extremes hypercapnia’ (50–53). The ultimate expression of this of arterial carbon dioxide pressure and the magnitude of strategy is high-frequency ventilation; FRC is main- fluctuations in arterial carbon dioxide pressure are asso- tained by PEEP or mean airway pressure, and rapid ciated with severe IVH in preterm infants (64,65). tiny volume breaths are used for ventilation. Even a few large tidal volume breaths have been shown to be Oxygen injurious to newborn animal lungs (53–56). For these Oxygen administration to the preterm infant is sus- reasons, neonatologists commonly are averse to bag- pected to cause tissue damage via oxygen-free radicals mask ventilation, especially given the inaccuracy of and is implicated in BPD and ROP (66). The antioxi- analog manometers (57). At Stanford, we are moving dant system is compromised in prematurity, as hydro- to the use of flow controlled, pressure limited mechani- xyl radical formation is increased beyond the immature cal devices specifically designed for neonatal resuscita- infant’s capacity to limit oxidative damage (67). Bio- tion and short-term ventilation. chemical markers of oxidative stress are still elevated Postoperative cardiac patients are typically venti- 28 days after birth in newborns resuscitated with 100% lated using high tidal volumes, low rates, and variable oxygen compared with those who received room air amounts of PEEP. This strategy minimizes effects on (68). The International Liaison Committee on Resusci- right heart afterload, but these higher tidal volumes tation stated air is as effective as 100% oxygen for the place immature lungs at high risk for VILI. Ades and resuscitation of most infants at birth (69). A United colleagues suggest transitioning postoperatively to Kingdom survey found no consistency in the use of NICU ventilation strategies as soon as hemodynamics oxygen during anesthesia in premature and newborn

Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd 541 Open-heart surgery for low birth weight infants G.D. Williams and R.S. Cohen infants (70). For LBW infants with CHD, the desired exceedingly difficult to maintain oxygen saturations in oxygen saturation levels to avoid pulmonary over-circu- the range recommended for premature infants, and clin- lation in ductal-dependent lesions and those obtainable ical outcome studies would be challenging. Our CPB in cyanotic heart lesions are similar to the saturations management of oxygen administration in premature recommended to minimize oxygen toxicity in premature infants does not differ from our practice in term infants. infants. In other cardiac lesions, maintaining this lower saturation goal should not be harmful (57). Near infra- Cardiovascular issues red spectroscopy (NIRS) monitoring of cerebral satura- tions may be helpful in guiding oxygen therapy. Systemic vascular resistance in the fetus is very low, Oxygen administration will decrease PVR in some with both ventricles contributing to systemic cardiac patients with pulmonary hypertension (PHT). output. At birth, the left ventricular output approxi- ROP is found primarily in preterm infants mately doubles. There are fewer contractile elements in <32 weeks gestational age and results from abnormal the immature heart, and the myocardium is less com- development of the retinal vasculature. In Phase I of the pliant. This stiffness renders the ventricle relatively disease, oxygen levels in the preterm infant’s retina are intolerant of volume and pressure loading. There is higher than normal in utero values, and this relative hy- limited ability to increase stroke volume, and cardiac peroxia impairs retinal vessels growth. This is followed output is heart rate sensitive. Parasympathetic tone in Phase II by hypoxia-driven uncontrolled abnormal predominates. Hypotension is often noted in the pre- vessel growth. Evidence is accumulating that one of the term neonate and is commonly because of myocardial strongest predictors of ROP, in addition to low gesta- dysfunction or altered peripheral vasoregulation. The tional age, is poor weight gain during the first weeks of neonatal myocardium is more sensitive to changes in life (71–73), and a common mechanism is proposed extracellular ionized calcium and relies more on glu- (74). During the third trimester, physiological ‘hypoxia’ cose metabolism (compared with fat) than that of of the retina stimulates vascular endothelial growth fac- older patients. Thus, premature and SGA infants may tor (VEGF) secretion. VEGF requires insulin-like be more at risk for myocardial dysfunction given their growth factor I (IGF-I) to initiate normal retinal vascu- decreased stores of calcium, inadequate glycogen larisation (75). IGF-I is essential for both pre- and post- stores, and impaired gluconeogenesis (35). Alpha and natal general growth as well as for growth of the retinal beta adrenergic receptors in premature infants are ade- vasculature, and serum levels are low in the premature quate in number and function. However, coupling of (76). Thus, the relative retinal hyperoxia of extra-uterine adrenergic receptors to adenylate cyclase is incomplete life (compounded by supplemental oxygen) and the defi- and may render milrinone less effective. Unfortunately, ciency of IGF-I combine to prevent normal retinal ves- inotropic agents have not been well studied in the pre- sel growth. Postnatal growth retardation is a major term population, and data regarding safety and effi- problem in very preterm infants, especially if they have cacy are lacking (82). Thus, choice of inotrope therapy CHD. Poor early weight gain and low serum IGF-1 lev- is mostly determined by institutional practice. Hydro- els both correlate with severity of ROP (73,77). This cortisone is an option for ‘pressor-resistant’ hypoten- may explain the observed association between ROP and sion (83) or capillary leak syndromes, which are CHD. A meta-analysis found that restriction of oxygen common in cardiac and premature infants. However, significantly reduced the incidence and severity of ROP steroid therapy has been associated with adverse neu- without increasing death rates (78), although increased rodevelopmental outcomes and bowel injury. mortality remains a concern (79). Many units will now PHT can complicate the perioperative period in tolerate SpO2 levels in the 70–90% range for premature LBW infants undergoing OHS. Acidemia, hypoxia, infants during their acute illness (80), while a higher CPB, and noxious stimuli can potentiate PVR. Certain

SpO2 range may be indicated during Phase II of ROP cardiac and noncardiac congenital anomalies are at (81). Therefore, for LBW infants with CHD, it seems increased risk for PHT. Patients with single ventricle reasonable to follow neonatal intensive care ROP guide- palliation do poorly when PVR is elevated. PHT is lines, unless this conflicts with the needs of managing more likely in infants with BPD and, when present, is the physiology of the underlying defect. associated with higher mortality after OHS (34). There are virtually no data to guide oxygen manage- ment of the LBW infant during CPB. Variables such as Gastrointestinal issues hemodilution, CPB circuit primed with banked adult red blood cells, hypothermia, cerebral blood flow, and Premature and SGA infants are at increased risk for variations in pump flow and circuit volumes make it necrotizing enterocolitis (NEC), a disease in which the

542 Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd G.D. Williams and R.S. Cohen Open-heart surgery for low birth weight infants gut mucosal barrier is damaged by pathogenic enteric (98), and there are multiple reasons for this (97). Prema- bacteria and may progress to bowel necrosis, sepsis, and ture infants have lower glomerular filtration rates death. Coexisting CHD further increases the incidence (GFR), higher renovascular resistance, and impaired of NEC (84). Risk factors include poor systemic perfu- concentrating and diluting abilities compared to term sion, particularly if the cardiac anomaly causes substan- infants. The immature tubular function leads to bicar- tial aortic run-off to the lungs (e.g., truncus arteriosus) bonate and sodium losses, thus requiring replacement to and prostaglandin infusion rates >0.05 mcgÆkg)1Æmin)1 maintain normal sodium balance and avoid acidemia. (7,84). NIRS monitoring of the anterior abdomen has Adequate GFR is maintained by postglomerular, effer- been reported to be a valid noninvasive measure of ent arteriolar vasoconstriction, which mainly is depen- splanchnic oxygen saturation in infants following OHS dent on a high level of angiotensin II and angiotensin- (85). Jaundice is more severe in LBW infants because of converting enzyme inhibitors should be used with cau- reduced red cell survival and liver immaturity. Preterm tion (99). Similarly, prostaglandin synthase inhibitors infants suffer brain damage at lower unconjugated bili- administered to close a patent ductus arteriosus may rubin levels than term babies because of a relative defi- blunt the vasodilation needed to maintain adequate per- ciency of the protective substance P-glycoprotein and fusion of the LBW kidney. Hypoxemia, hypothermia, the increased exposure to exacerbating factors such as and sepsis-associated vasoactive substances (e.g., endo- hypoxia, acidosis, hypercapnea, and hypothermia (86). thelin, thromboxane A2) reduce renal blood flow and Attention to the pre- and postoperative nutritional GFR and activate the renin-aldosterone-angiotensin requirements is vital for LBW infants undergoing OHS. system. Both before and after surgery, low cardiac out- The complications of intravenous hyperalimentation put syndrome was a major determinant of renal injury can be severe and include line infections, thrombosis, (95). Positive pressure ventilation decreases venous and liver injury. Therefore, enteric feeding is preferable return and increases renal sympathetic nervous activity with the understanding that these patients are at risk for and serum vasopressin levels (97). LBW infants are NEC and aspiration (increased incidence of gastrointes- often exposed to nephrotoxic agents such as antibiotics tinal reflux and vocal cord paresis). and intravenous contrast media. CPB impairs renal Neonatologists have been focusing attention lately function by activation of inflammation, nonpulsatile on what has been termed ‘extra-uterine growth retar- flow, hemolysis, and ischemia. Circulatory arrest and dation’: the common phenomenon of poor growth for duration of CPB are associated with acute kidney injury very low birth weight infants during their NICU stay in infants (95). Recent reports have suggested that the- (87–89). This failure to grow appropriately and main- ophylline may be protective against renal injury in neo- tain weight above the 10th percentile for postmenstrual nates with respiratory distress (100) and asphyxia (101). age has been associated with a more complicated This has not been studied in the context of CPB, but the NICU course, and a worse long-term outcome (90). analogy could be made. Thus, theophylline may be Providing more protein intake in the first week of life considered a potential means of prophylaxis against post- may at least in part ameliorate this; an extra gm per operative renal dysfunction in infants undergoing OHS. kg of protein daily during that time resulted in better growth and an 8.2-point improvement in Mental Neurological issues Development Index (91,92). Thus, there has been an increased focus on providing more protein and calories Compared to normal weight infants, LBW infants are starting immediately upon NICU admission up to more likely to incur brain damage and have long-term 3gÆkg)1 even on the first day of life (93,94). Because neurological disabilities; this is especially true for the premature infants with CHD frequently require vari- extremely LBW (<1000 g) child (102). Factors that ous intravenous infusions (e.g., alprostadil, dopamine), are implicated include genetic disorders, brain injury in there commonly is conflict between the desire to pro- utero, perinatal asphyxia, IVH (associated with prema- vide maximized parenteral nutrition and the need for turity), and postnatal insults such as hypoxia, hypoper- fluid restriction. fusion, and infection. OHS during infancy is also associated with early and late neurological morbidities. Neurological complications such as stroke and seizures Renal issues can present in the early postoperative period, but are In children and infants, acute kidney injury incidence less common than the longer-term issues such as cogni- after OHS varies from 2.7% to 24.6% (95,96), with sur- tive and intellectual impairment, attention and execu- vival rates ranging from 21% to 80% (97). Young age tive function deficits, visual-spatial and visual-motor and small size were risk factors for poor renal outcome skills deficiencies, speech and language delays, and

Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd 543 Open-heart surgery for low birth weight infants G.D. Williams and R.S. Cohen behavioral difficulties (103–105). Mechanisms of injury open posture. Body core temperature typically include patient-specific factors such as genetic predis- responds quickly to thermo-manipulation during CPB. position, gender, race, socioeconomic status, and in Babies are particularly prone to becoming hypothermic utero central nervous system development. Modifiable during transport and in the operating room prior to factors include not only intraoperative variables (CPB, draping for surgery when the child is fully exposed. deep hypothermic circulatory arrest, and hemodilution) Measures that help thermoregulation include transpor- but also such variables as hypoxemia, hypotension, tation in an incubator, preparing the OR to appropri- and low cardiac output (106). In addition, there are ate temperature and humidity, heating lights, plastic concerns about the possibility of anesthesia-induced wrap, warm, humidified inspired gases, and warmed neuronal cell loss in the developing brain (107–109). topical and intravenous fluids. The strongest predictors of a worse neurodevelopmen- tal outcome at 1 year of age were reported to be Hematology patient-specific factors including presence of a genetic LBW infants are often anemic (immature marrow, low syndrome, LBW, and presence of the apolipoprotein E erythropoietin concentration, reduced red blood cell life- genotype (110). A large, prospective cohort study dem- span, iron deficiency, frequent blood sampling). Their onstrated a 10% prevalence of stroke in infants with coagulation profile is different but adequate, with a ten- CHD undergoing OHS, half of which occurred preop- dency toward thrombosis – particularly if polycythemia, eratively. LBW was independently associated with low cardiac output, or dehydration are present (113). It stroke (111). is pointless to treat clotting study results for a premature Measures to protect the infant brain during OHS infant that may be in the ‘abnormal’ range for older continue to evolve (103). There has been concern that children; ‘normal’ results will not be achieved, blood the immature germinal matrix of premature infants products will be wasted, additional fluid load will be would greatly increase the risk of IVH following OHS. given, and hypercoagulation may be the result. Addi- However, the incidence of severe IVH in premature tionally, erythropoietin studies for the treatment of ane- infants (median gestational age = 33 weeks) undergoing mia have demonstrated that premature infants tolerate OHS was not different from a control group without quite well much lower hematocrits than was previously CHD (7). Most cases of severe IVH present in the first believed. Current transfusion recommendations for pre- 72 h after birth and occur in infants <30 weeks gesta- mature infants now are much lower, with hematocrits in tional age. Therefore, it might be reasonable to delay the 18–20% range deemed acceptable by many (114). OHS beyond the first few days of life if the infant is very premature. Pharmacology LBW infants may differ from term infants in dose response because of pharmacokinetic and pharmacody- Other issues namic effects. For example, there are differences because Infection of organ immaturity (liver, kidney, gut), body composi- Extremely premature newborns have a 5- to 10-fold tion (total body water, fat, protein binding, pH, biliru- higher incidence of microbial infection than term new- bin), and age-dependent metabolic pathways (e.g., borns (112). Infections decrease the chance of survival. theophylline) that affect drug metabolism and clearance. Unfortunately, premature infants are more likely to be exposed to risk factors for infection such as indwelling Invasive therapies vascular catheters, mechanical ventilation, intravenous Central intravascular catheters are often present. Posi- hyperalimentation, NEC, invasive procedures, and pro- tioning of the catheter tip should be checked because longed hospital stay. Patients with conotruncal cardiac adverse events are more common at certain sites (e.g., defects are more likely to have 22q11.2 deletion syn- renal thrombosis when umbilical artery catheter drome (DiGeorge anomaly) and be immunocomprom- located at origin of renal arteries). Long duration in ized. Blood transfusions and steroid therapy may also situ and urgency of placement usually raise our con- depress immunity. cern for catheter-associated infections.

Thermoregulation Anesthetic management LBW infants at risk for hypo- and hyperthermia and dehydration because they have a larger body surface Cardiopulmonary bypass area/weight ratio, very thin skin, decreased brown fat Table 4 lists details of surgery and CPB management are unable to shiver or sweat and assume a flaccid of the Stanford LBW patients. Owing to a lack of

544 Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd G.D. Williams and R.S. Cohen Open-heart surgery for low birth weight infants

Table 4 Details of cardiac diagnosis and cardiopulmonary bypass is different for LBW patients in the following (CPB) (n = 65) aspects: Surgery and CPB parameters Value* Hemodilution Diagnosis details Hemodilution is extreme, resulting in increased blood Tetralogy of Fallot, with/without pulmonary 12 (18%) product requirements to correct anemia and coagula- atresia tion function. Mean prime volume ranged from 1.8 to Transposition of great arteries 10 (15%) Truncus arteriosus 7 (11%) 6.5 times greater than the estimated blood volume of Total anomalous pulmonary venous return 6 (9%) Stanford study patients. Miniaturized CPB circuits are Hypoplastic left heart variant (Sano-modified 6 (9%) being developed, but are not currently available for Norwood I) clinical use. Successful use of a nonheme prime for a Ventricular septal defect and aortic arch 6 (9%) 1700-g premature infant has been reported (115), but obstruction hematocrit values during CPB fell to values that others Ventricular septal defect, with/without atrial 5 (8%) have associated with poorer neurological outcome septal defect Norwood aortic arch repair + Rastelli repair 3 (5%) (116). ventricular septal defect Pulmonary atresia with intact ventricular septum 2 (3%) Inflammatory response to CPB Tetralogy of Fallot, pulmonary atresia, 2 (3%) Tremendous activation of inflammation can result in aorto-pulmonary collaterals fluid exudates and substantial tissue edema. Often our Partial or complete atrioventricular septal defect 2 (3%) surgeons opt for delayed closure of the sternum, and Hypoplastic left heart variant (Classic Norwood I) 1 (2%) many LBW infants require insertion of a peritoneal Shone’s complex (repair of mitral valve, 1 (2%) aortic valve, aortic arch) dialysis catheter to avoid an abdominal compartment + ventricular septal defect 1 (2%) syndrome. In hope of mitigating the inflammatory Aorto-pulmonary window + aortic arch 1 (2%) response, we perform dilutional conventional ultrafil- reconstruction tration and usually remove 600–1000 ml of ultrafil- Single ventricle physiology 9 (14%) trate. Methylprednisolone (30 mgÆkg)1) and mannitol )1 CPB details (0.5 gÆkg ) are added to the CPB prime. The addition Duration of CPB (min, median ± IQR) 142 ± 67 of modified ultrafiltration may be useful if the duration Duration of aortic cross clamping 67 ± 51 of conventional ultrafiltration is curtailed by an (min, median ± IQR) extended period of circulatory arrest (117). Deep hypothermic circulatory arrest 30 ± 20/11 (min, mean ± SD/%) Oxygen toxicity Selective antegrade cerebral perfusion 16 (% of total) As mentioned earlier, the optimal oxygen values dur- pH stat acid-base management (% of total) 67 ing CPB are unknown. It would be technically chal- Aortic cannula sizes: 4/11/85 lenging during CPB to provide adequate systemic and

14G IV catheter/6 F/8 F (%) cerebral perfusion if SaO2 was maintained at 85% Venous cannula sizes: 19/67/14 because blood flow into and out of the CPB circuit is 10 F/12 F/‡14F (% of total) not constant and the oxygen carrying capacity is lower Circuit prime: packed red blood cells 213 ± 34 (hemodilution). (ml, mean ± SD) Circuit prime: fresh frozen plasma 60 ± 18 (ml, mean ± SD) Anticoagulation management Circuit prime: crystalloid (ml, mean ± SD) 187 ± 82 The risk of intracranial hemorrhage was discussed ear- Circuit prime: total volume (ml, mean ± SD) 393 ± 182 lier. Heparin dosing and cofactor activity differs in the Phentolamine administered (%) 35 premature. Higher doses per kg body weight may be Methylprednisolone administered (%) 100 required to achieve target-activated clotting time. We Minimum core temperature (C, mean ± SD) 22.3 ± 5.4 manage heparin therapy by monitoring heparin con- Minimum hematocrit on CPB (%, mean ± SD)32±5 Maximum hematocrit on CPB (%, mean ± SD)45±6centration (using heparin-protamine titration) and acti- vated clotting times. *Values expressed as percentage of total patients unless stated otherwise. Neuroprotection relevant data, CPB management of LBW infants is Strategies used for normal birth weight infants are mostly extrapolated from experience gained caring adopted because of the paucity of data about neuro- for term neonates. This is suboptimal because CPB protection during CPB for LBW infants. The prefer-

Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd 545 Open-heart surgery for low birth weight infants G.D. Williams and R.S. Cohen ence at Stanford is a hematocrit of 28–30% during diac lesion have similar cardiac pathophysiology and CPB (103). We employ pH stat acid-base management present the anesthesiologist with similar management if the induced hypothermia is severe enough (<26– challenges. However, there are some anesthetic issues 28C) to impair cerebral autoregulation (103). Elevated that are especially relevant to LBW infants. blood glucose concentrations are not treated intraoper- atively and usually remain below 150 mgÆdl)1. Safe Anesthetic agents limits for cerebral blood flow and cerebral perfusion The safety of general anesthesia for neonates and LBW pressure during CPB are unknown. This is particularly babies in particular is currently under intense debate relevant during antegrade cerebral perfusion. Cerebral because anesthetic-induced neurotoxicity in immature oxygenation monitoring by NIRS may be difficult animal models has been reported (108,109,118–122). because of the baby’s small forehead. Implicated drugs include N-methyl D-aspartate receptor antagonists (e.g., ketamine, nitrous oxide) and agents Patient size with GABA mimetic properties (e.g., diazepam, isoflura- The technical conduct of CPB is more challenging. ne, halothane, propofol). There are also data suggesting CPB cannulae are small and can easily become dis- that these same anesthetics may be neuroprotective placed or obstructed. LBW patients seem more prone against hypoxic-ischemic injury and that inadequate to electrolyte and pH abnormalities during CPB, per- analgesia during painful procedures may lead to haps because of the large circuit volumes. Administra- increased neuronal cell death in animals and long-term tion of banked blood, , and saline can behavioral changes in humans (119). Additionally, it is cause acidosis, hyperkalemia, and hypernatremia. Tro- well recognized that inappropriate choice of anesthetic methamine is a useful low-sodium buffer that is FDA agent or dose can lead to suboptimal hemodynamics approved for use during pediatric CPB. and cerebral hypoxia, thereby resulting in adverse neu- rological outcome. OHS in critically ill neonates and children has been associated with impaired developmen- Operating room preparation and patient transport tal outcomes (104,123–125), but the contribution of Standard protocol at Stanford includes: (i) scheduling anesthesia is unknown. The current Stanford approach nonurgent LBW cases during time of day when special- is anesthesia induction with ketamine and rocuronium, ized hospital resources are readily available; (ii) preoper- maintenance with midazolam, low-dose isoflurane, and ative collaboration between surgical, nursing, perfusion, high-dose fentanyl (Table 5). A TIVA technique is man- and anesthesia staff to finalize intraoperative plan; (iii) dated when an ICU ventilator is employed. priming the CPB circuit prior to transport of patient to operating room; (iv) use of neonatal ventilator; (v) if Monitoring necessary, ability to deliver nitric oxide, nitrogen (to Optimally positioned umbilical artery and vein catheters provide FiO2 values in range of 0.16–0.21), or carbon are very useful for hemodynamic monitoring, drug dioxide to optimize pulmonary blood flow (59); (vi) administration, volume resuscitation, and blood sam- operating room at suitable temperature and humidity pling, but are not always in situ. Peripherally inserted and (vii) drug dilutions appropriate to patient weight. central catheters are less practical because of their nar- Considerations for patient transport include: (i) row lumen. We do not insert subclavian or internal jug- appropriate equipment and medications for possible ular venous catheters because of concerns about trauma patient decompensation; (ii) use of gas blender to main- and thrombosis. Instead, two right atrial catheters are tain pretransport FiO2; (iii) measures to minimize heat surgically inserted after sternotomy; one to monitor and water loss from patient; (iv) full monitoring, includ- atrial pressure and the other for drug administration ing ETCO2; (v) portable suction apparatus if endotra- and volume resuscitation. When wrist arterial access is cheal secretions are a concern; (vi) vigilance during hand compromised, we prefer axillary artery (22G, 25-mm ventilation to ensure Qp : Qs balance is maintained and intravenous catheter) rather than femoral artery cannu- (vii) confirm access to baby can be easily achieved when lation because of concerns about infection (proximity to incubator is disconnected from power source. perineum) and leg ischemia, although there are insuffi- cient data to support this bias. Epicardial echocardiog- raphy is our standard for LBW infants although we do Anesthetic technique use intracardiac echocardiography catheters under a Appropriate anesthetic management for OHS requires research protocol for patients >2 kg. Point of care lab- an appreciation of the patient’s cardiac pathophysiol- oratory tests facilitates ongoing assessment of the ogy. LBW and term infants with the same type of car- patient’s status. LBW infants are particularly prone to

546 Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd G.D. Williams and R.S. Cohen Open-heart surgery for low birth weight infants

Table 5 Anesthetic management (n = 65) ment. We attempt to maintain the preoperative mechanical ventilation and FiO goals, but not at the Variable Valuea 2 expense of hemodynamic stability or cerebral oxygena-

Monitoring tion. Accurate tidal volume delivery and ETCO2 moni- Umbilical vein catheter (%) 32 toring at the ETT are required. These are best Peripheral intravenous central catheter 21 achieved with a sophisticated neonatal ventilator. An (PICC) (%) alternative for physicians who prefer an anesthetic Femoral venous catheter (%) 12 Internal jugular vein catheter (%) 2 machine ventilator would be additional respiratory Transthoracic right or common atrial catheter (%) 89 monitoring with neonate-appropriate devices such as Transthoracic left atrial catheter (%) 46 the NICO monitor (Philips Respironics/Novametrix, Umbilical artery catheter (%) 37 Andover, MA, USA). Surfactant administration post- Radial or ulnar or axillary artery catheter (%) 30 CPB has not been necessary to date. Femoral artery catheter (%) 33 Echocardiography: 89/11 Hemodynamic support epicardial/transesophagealb (%) Milrinone, epinephrine, and dopamine infusions are Ventilation typically used at Stanford. Ionized calcium values are Cuffed ETT (%) 47 maintained at the upper end of the age-appropriate ETT sizes (mm) 2.5/3.0/3.5 (%) 0/28/71 range by infusing calcium chloride. Inhaled nitric oxide, FiO2 < 0.21 at induction (%) 9 Intensive care ventilator: 25/2 typically £20 p.p.m. (126), is administered if PHT is conventional/oscillator (%) suspected. Temporary epicardial pacing is employed for arrhythmias when indicated, but also when a sinus Anesthetic technique rhythm rate is relatively slow. When rewarming after Anesthesia induction included ketamine (%) 79 Anesthesia maintenance included ketamine (%) 9 CPB, target core temperature is limited to 35–36Cif Anesthesia maintenance included isoflurane (%) 75 there is concern about junctional ectopic tachycardia. Isoflurane administered during CPB (%) 97 Midazalom administered (%) 77 Hemostasis Fentanyl administered (%) 100 The consequences of immature hemostatic function, Muscle relaxant: rocuronium/pancuronium (%) 83/17 cyanotic heart disease, liver dysfunction, hemodilution, Inotropes: dopamine/epinephrine/ 100/57/81/100 inflammation, hypothermia, and surgical bleeding com- milrinone/calcium (%) Furosemide administered (%) 63 bine to greatly perturb coagulation. We do not use an- Temporary pacemaker used (%) 35 tifibrinolytic agents for LBW infants; the risks and benefits are unclear for this population. Fresh (<48 h) Blood products administered intraoperatively whole blood is not available at our institution. Packed red blood cells (units, median ± IQR) 3 (1) Apheresed platelets (units, median ± IQR) 1 (0.3) Activated Factor VII is only administered as rescue Fresh frozen plasma (units, median ± IQR) 1 (0.4) therapy. At Stanford, fresh frozen plasma is added to Cryoprecipitate (units, median ± IQR) 2 (1) the CPB circuit during dilutional conventional ultrafil- Total blood products (units, median ± IQR) 7 (1.3) tration. Post-CPB, 1 unit (50 ml) of volume-reduced apheresed platelets and 2 units of cryoprecipitate are CPB, cardiopulmonary bypass; ETT, endotracheal tube. aValues expressed as percentage of total patients unless stated transfused to correct a predictable deficiency of plate- otherwise. lets and fibrinogen (127). Further products are adminis- bStudy probe used, as per research protocol. tered as necessary. All red cell units are washed to limit the hyperkalemia and acidosis associated with banked hypo- and hyperglycemia during OHS, and frequent blood. Calcium chloride is infused (10–40 mgÆkg)1Æh)1) blood glucose monitoring is required. to mitigate citrate-induced hypocalcemia. All fluids are delivered via syringe pump to avoid inadvertent volume Pulmonary overload. Total volume of blood products transfused If the ETT has been in situ for a > 5 days or if secre- exceeds the patient’s estimated blood volume. tions are copious, we are usually inclined to change the ETT. We may upsize the ETT or replace it with a Other cuffed ETT if there is a substantial air-leak. This is Meticulous care is advised when padding and position- done in anticipation of a CPB-induced lung dysfunc- ing the patient to minimize heat loss and skin trauma. tion. After positioning the patient for surgery, a chest The urinary catheter can be problematic; it is narrow X-ray is obtained to verify satisfactory ETT place- and often obstructs or urine leaks around it.

Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd 547 Open-heart surgery for low birth weight infants G.D. Williams and R.S. Cohen

nates (22,128). The most common morbidity identified Adverse intraoperative events in a recent study was infections of the blood stream. In the Stanford series, adverse events that affected Infections and chronic lung disease were associated anesthetic management occurred in 28% of surgeries with increased length of stay (22). (Table 6). Common problems were arrhythmia, pulmo- nary hemorrhage, decreased lung compliance after Mortality CPB, and excessive blood loss. Several patients desatu- rated during induction, suggesting marginal cardio-pul- LBW is a known risk factor for mortality following monary reserve. In 48% of patients, the surgeon opted OHS with early mortality rates ranging from 10% to to not close the sternum and placed a silastic patch. 42% (7,13–22,128–131). Interestingly, increased mortal- Intraoperative extracorporeal membrane oxygenation ity and morbidity after OHS is recently reported for (ECMO) support was never required. Similar reports neonates delivered at 37–38 weeks gestation when com- for comparison were not found. pared to a control group of neonates born at 39–40 weeks gestation (128). Seventeen percent of the Stanford LBW patients died during postoperative hospitalization Outcome (Table 7). Mortality for two-ventricle repair (12%) was Postoperative course within the reported range for LBW infants (20,22,35). As indicated in Table 7, LBW patients typically have a Mortality has been associated with high RACHS-1 and prolonged recovery after OHS compared to term neo- Aristotle preoperative risk assessment scores

Table 6 Intraoperative adverse a Event Weight (kg) In-hospital mortality events that complicated anes- Hemodynamic thetic management Hypotension with chest closure 0.98 Died POD 1 from junctional arrhythmia and hypotension Temporary ventricular dysfunction 2.4 Alive (air in coronary) Ventricular fibrillation during sternotomy 2 Died POD 37 from sepsis and multi-organ failure Slow junctional arrhythmia 2.45 Alive Junctional ectopic tachycardia 1.7 Alive Junctional ectopic tachycardia 2.5 Died POD 67 from sepsis and multi-organ failure Supraventricular tachycardia 1.1 Alive

Respiratory Brief hypoxia with induction 1.8 Alive Hypoxia, bradycardia with induction; 2.2 Alive Pre-CPB hemorrhage in left upper lobe of lung Difficult intubation: Failed direct and fiberoptic 2.25 Alive attempts, success using rigid bronchoscope Significant airleak around ETT post-CPB; PHT 2.2 Alive Blood in ETT causing occlusion 2.11 Alive Blood in ETT causing occlusion 2.3 Alive Blood in ETT causing occlusion 2.3 Alive Hypoxia from inadequate pulmonary blood flow 2.42 Died POD 7 from (Norwood stage I). Sano shunt revised. bleeding following chest tube insertion

Other PHT; liver hemorrhage 1.9 Died day of surgery from liver hemorrhage Air conditioning problem after patient induced. 2.04 Alive Surgery delayed until issue resolved. Bleeding caused return to CPB 2.1 Alive

POD, postoperative day; CPB, cardiopulmonary bypass; ETT, endotracheal tube; PHT, pulmonary hypertension. aWeight on day of surgery.

548 Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd G.D. Williams and R.S. Cohen Open-heart surgery for low birth weight infants

Table 7 Patient in-hospital outcome (n = 65) 0.6 kg. Others have failed to find an association between very LBW (<1.5 kg) and mortality Variable Value (7,18,19,22,131). It is likely that few infants in the low- Hospital survivors est gestational age and lowest weights at birth had Total in-hospital survivors 54 (83%) severe forms of congenital cardiac disease, as many of Postoperative mechanical ventilation 6 (8) those infants do not survive to repair, or are not (days, median ± IQR) offered repair owing to extreme prematurity (22). SGA Duration in cardiac intensive care 10 (12) (days, median ± IQR) infants undergoing OHS have been reported to have Postoperative duration in hospital 20 (15) increased risk for mortality and morbidity when com- (days, median ± IQR) pared to controls matched for gestational age (22), but Discharged to medical facility (% of total) 44 several studies have found no association (131) or a Nonsurvivors suggestion of increased survival (17,22). Perhaps gesta- Total in-hospital nonsurvivors 11 (17%) tional age is a more important risk factor for poor Duration of ventilation (days, median ± IQR) 21 (28) outcome than weight (22). There is contradictory evi- Duration in hospitala (days, median ± IQR) 22 (32) dence on the influence of associated anomalies on mor- tality (7,17,20,22,130,131). The difference in reported Survivor if discharged from the hospital where cardiac surgery was performed. outcomes could be related to the selection criteria for aAll nonsurvivors died during stay in Cardiovascular Intensive Care surgery. Postoperative factors that have been associ- Unit. ated with mortality include cardiopulmonary resuscita- tion, ECMO, infection, and reoperation in the same admission (22,133,136). (13,132,133) and certain cardiac diagnoses (133). A meta-analysis demonstrated that first-stage reconstruc- tion for hypoplastic left heart syndrome (HLHS) was a The future predictor of mortality in LBW infants (134), and mortal- As these complex infants with multiple risk factors for ity rates of 38% to 67% are reported (22,131,135,136). adverse outcomes begin to have improved survival, the Hospital survival for Stanford infants who underwent morbidity burden will be higher with a commensurate this procedure (43%) was much lower than the institu- increase in length of stay and resource utilization. Peri- tion’s reported survival (93%) for the overall HLHS operative care for LBW neonates with complex CHD population (137). LBW was also associated with frequently requires unplanned diagnostic and therapeu- increased mortality for infants undergoing the hybrid tic interventions as well as multiple subspecialty con- procedure for HLHS (138). sultations. This infrastructure may be most efficiently In the Stanford LBW series, patient weight was not supported in high-volume centers in countries with predictive of outcome, and the mortality rate for very sophisticated healthcare systems. The perioperative and extremely LBW babies (11%) was no worse than management described may not easily apply to popula- that for LBW infants (18%), suggesting it is reason- tions and/or healthcare systems elsewhere (134,139). able to perform OHS on children weighing as little as

References 1 Rosenberg A. The IUGR newborn. Semin tal cardiac disease. Eur J Pediatr 1990; 149: operation for complex D-transposition. Ann Perinatol 2008; 32: 219–224. 752–757. Thorac Surg 2008; 85: 1698–1702. 2 Tanner K, Sabrine N, Wren C. Cardiovas- 6 Ferencz C, Boughman JA, Neill CA et al. 10 Prandstetter C, Hofer A, Lechner E et al. cular malformations among preterm infants. Congenital cardiovascular malformations: Early and mid-term outcome of the arterial Pediatrics 2005; 116: e833–e838. questions on inheritance. J Am Coll Cardiol switch operation in 114 consecutive patients: 3 Rosenthal GL, Wilson PD, Pemutt T et al. 1989; 14: 756–763. a single center experience. Clin Res Cardiol Birth weight and cardiovascular malforma- 7 Dees E, Lin H, Cotton RB et al. Outcome 2007; 96: 723–729. tions: a population based study. Am J Epi- of preterm infants with congenital heart dis- 11 Tibballs J, Kawahira Y, Carter BG et al. demiol 1991; 133: 1273–1281. ease. J Pediatr 2000; 137: 653–659. Outcomes of surgical treatment of infants 4 Levy RJ, Rosenthal A, Fyler DC et al. 8 Chang AC, Hanley FH, Lock JE et al. with hypoplastic left heart syndrome: an Birthweight of infants with congenital car- Management and outcome of low birth institutional experience 1983–2004. J Paedi- diac disease. Am J Dis Child 1978; 132: weight neonates with congenital heart dis- atr Child Health 2007; 43: 746–751. 249–254. ease. J Pediatr 1994; 124: 461–466. 12 McGuirk SP, Sticklev J, Griselli M et al. 5 Kramer HH, Trampisch HJ, Rammos S 9 Gottlieb D, Schwartz ML, Bischoff K et al. Risk assessment and early outcome follow- et al. Birthweight of children with congeni- Predictors of outcome of arterial switch ing the Norwood procedure for hypoplastic

Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd 549 Open-heart surgery for low birth weight infants G.D. Williams and R.S. Cohen

left heart. Eur J Cardiothorac Surg 2006; 29: 26 Bernstein IM, Horbar JD, Badger GJ et al. Care. Baltimore, MD: Lippincott Williams 675–681. Morbidity and mortality among very-low- & Wilkins, 1998: 107–125. 13 Curzon CL, Milford-Beland S, Li JS et al. birth-weight neonates with intrauterine 40 Pessonen EJ, Korpela R, Peltola K et al. Cardiac surgery in infants with low-birth growth restriction: the Vermont Oxford Regional generation of free oxygen radicals weight is associated with increased mortal- Network. Am J Obstet Gynecol 2000; 182: during cardiopulmonary bypass in children. J ity: analysis of the Society of Thoracic Sur- 198–206. Thorac Cardiovasc Surg 1995; 110: 768–773. geons Congenital Heart Database. J Thorac 27 Aucott SW, Donohue PK, Northington FJ. 41 Newth JL, Hammer J. Pulmonary issues. In: Cardiovasc Surg 2008; 135: 546–551. Increased morbidity in severe early intra- Chang AC, Hanley FL, Wernovsky G, Wes- 14 Beyens T, Biarent D, Bouton JM et al. uterine growth restriction. J Perinatol 2004; sel DL, eds. Pediatric Cardiac Intensive Cardiac surgery with extracorporeal circula- 24: 435–440. Care. Baltimore, MD: Lippincott Williams tion in 23 infants weighing 2500 g or less: 28 Tommiska V, Heinonen K, Lehtonen L & Wilkins, 1998: 351–367. short and intermediate term outcome. Eur et al. No improvement in outcome of 42 Rabinovitch M, Grady S, David I et al. J Cardiothorac Surg 1998; 14: 165–172. nationwide extremely low birthweight infant Compression of intra-pulmonary bronchi by 15 Dimmick S, Walker K, Badawi N et al. populations between 1996–7 and 1999–2000. abnormally branching pulmonary arteries Neonatal Intensive Care Units Study Pediatrics 2007; 119: 29–36. associated with absent pulmonary valves. (NICUS) of the New South Wales Preg- 29 Coalson JJ. Pathology of new bronchopul- Am J Cardiol 1982; 50: 804–813. nancy and Newborn Services Network. Out- monary dysplasia. Semin Neonatol 2003; 8: 43 Lister G, Talner NS. Management of respi- comes following surgery for congenital heart 73–81. ratory failure of cardiac origin. In: Gregory disease in low-birthweight infants. J Paedi- 30 Huiskamp G, Lum S, Stocks J et al. Associ- G, ed. Clinics in Critical Care Medicine: atr Child Health 2007; 43: 370–375. ation of prematurity, lung disease and body Respiratory Failure in the Child. New York, 16 Bove T, Francois K, De Groote K et al. size with lung volume and ventilation inho- NY: Churchill Livingstone, 1981: 67–88. Outcome analysis of major cardiac opera- mogeneity in unsedated neonates: a multi- 44 Schulze-Neick I, Ho SY, Bush A et al. tions in low weight neonates. Ann Thorac centre study. Thorax 2009; 64: 240–245. Severe airflow limitation after the unifocal- Surg 2004; 78: 181–187. 31 Horbar JD, Badger GJ, Carpenter JH et al. ization procedure: clinical and morphologi- 17 Oppido G, Napoleone CP, Formigari R et Trends in morbidity and mortality for very cal correlates. Circulation 2000; 102: III142– al. Outcome of cardiac surgery in low birth low birth weight infants, 1991–1999. Pediat- III147. weight and premature infants. Eur J Cardio- rics 2002; 110: 143–151. 45 Gorenflo M, Vogel M, Obladen M. Pulmo- thorac Surg 2004; 26: 44–53. 32 National Institute of Child Health and nary vascular changes in bronchopulmonary 18 Pawade A, Waterson K, Laussen P et al. Human Development. Report of the dysplasia: a clinicopathologic correlation in Cardiopumonary bypass in neonates weigh- Consensus Development Conference on the short and long-term survivors. Pediatr ing less than 2.5 kg: analysis of the risk fac- Effect of Corticosteroids for Fetal Matura- Pathol 1991; 11: 851–866. tors for early and late mortality. J Card tion on Perinatal Outcomes. Washington, 46 Bland RD, Albertine KH, Carlton DP et al. Surg 1993; 8: 1–8. DC: National Institutes of Health, 1994: Chronic lung injury in preterm lambs: 19 Reddy VM, Hanley FL. Cardiac surgery in 95–3784. NIH Publication No. 1994. abnormalities of the pulmonary circulation infants with very low birth weight. Semin 33 Abman SH, Accurso FJ, Bowman CM. and lung fluid balance. Pediatr Res 2000; Pediatr Surg 2000; 9: 91–95. Unsuspected cardiopulmonary abnormalities 48: 64–74. 20 Reddy MV, McElhinney DB, Sagrado T complicating bronchopulmonary dysplasia. 47 Parker TA, Abman SH. The pulmonary et al. Results of 102 cases of complete repair Arch Dis Child 1984; 59: 966–970. circulation in bronchopulmonary dysplasia. of congenital heart defects in patients weigh- 34 McMahon CJ, Penny DJ, Nelson DP et al. Semin Neonatol 2003; 8: 51–61. ing 700 to 2500 grams. J Thorac Cardiovasc Preterm infants with congenital heart dis- 48 Bush A, Busst CM, Knight WB et al. Surg 1999; 117: 324–331. ease and bronchopulmonary dysplasia: post- Changes in pulmonary circulation in severe 21 Rossi AF, Seiden HS, Sadeghi AM et al. operative course and outcome after cardiac bronchopulmonary dysplasia. Arch Dis The outcome of cardiac operations in surgery. Pediatrics 2005; 116: 423–430. Child 1990; 65: 739–745. infants weighing two kilograms or less. J 35 Ades A, Johnson BA, Berger S. Manage- 49 Doull IJ, Mok Q, Tasker RC. Tracheobron- Thorac Cardiovasc Surg 1998; 116: 28–35. ment of low birth weight infants with con- chomalacia in preterm infants with chronic 22 Ades AM, Dominguez TE, Nicolson SC et genital heart disease. Clin Perinatol 2005; lung disease. Arch Dis Child Fetal Neonatal al. Morbidity and mortality after surgery 32: 999–1015. Ed 1997; 76: F203–F205. for congenital cardiac disease in the infant 36 Merrill JD, Ballard RA, Cnaan A et al. 50 Sharek PJ, Baker R, Litman F et al. Evalu- born with low weight. Cardiol Young 2010; Dysfunction of pulmonary surfactant in ation and development of potentially better 20: 8–17. chronically ventilated premature infants. Pe- practices to prevent chronic lung disease 23 Kecskes Z, Cartwright DW. Poor outcome diatr Res 2004; 56: 918–926. and reduce lung injury in neonates. Pediat- of very low birthweight babies with serious 37 Paul DA, Greenspan JS, Davis DA et al. rics 2003; 111(4 Pt 2): e426–e431. congenital heart disease. Arch Dis Child The role of cardiopulmonary bypass and 51 Kamlin COF, Davis PG. Long versus short Fetal Neonatal Ed 2002; 87: F31–F33. surfactant decompensation after surgery for inspiratory times in neonates receiving 24 McIntire DD, Bloom SL, Casey BM et al. congenital heart disease. J Thorac Cardio- mechanical ventilation. Cochrane Database Birth weight in relation to morbidity and vasc Surg 1999; 117: 1025–1026. Syst Rev 2003; 4: CD004503.pub2. DOI: mortality among newborn infants. N Engl J 38 Rinker C, Jansen S, Wilnhammer C et al. 10.1002/14651858.CD004503.pub2. Med 1999; 340: 1234–1238. Cardiopulmonary bypass reduces pulmonary 52 van Kaam AH, Rimensberger PC. Lung- 25 Regev RH, Lusky A, Dolfin T et al. Excess surfactant activity in infants. J Thorac protective ventilation strategies in neonatol- mortality and morbidity among small-for- Cardiovasc Surg 1999; 118: 237–244. ogy: what do we know—what do we need gestational-age premature infants: a popula- 39 Bohn D. Cardiopulmonary interactions. In: to know? Crit Care Med 2007; 35: 925–931. tion based study. J Pediatr 2003; 143: 186– Chang AC, Hanley FL, Wernovsky G, 53 Ambalavanan N, Carlo WA. Ventilatory 191. Wessel DL, eds. Pediatric Cardiac Intensive strategies in the prevention and management

550 Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd G.D. Williams and R.S. Cohen Open-heart surgery for low birth weight infants

of bronchopulmonary dysplasia. Semin Per- 67 O’Donovan D, Fernandes C. Mitochondrial tion in extremely preterm infants. N Engl J inatol 2006; 30: 192–199. glutathione and oxidative stress: implica- Med 2010; 362: 1959–1969. 54 Bjo¨rklund LJ, Ingimarsson J, Curstedt T tions for pulmonary oxygen toxicity in pre- 80 Tin W, Milligan DW, Pennefather P et al. et al. Manual ventilation with a few large mature infants. Mol Genet Metab 2000; 71: Pulse oximetry, severe retinopathy, and out- breaths at birth compromises the therapeu- 352–358. come at one year in babies of less than tic effect of subsequent surfactant replace- 68 Niermeyer S, Vento M. Is 100% oxygen 28 weeks gestation. Arch Dis Child Fetal ment in immature lambs. Pediatr Res 1997; necessary for the resuscitation of newborn Neonatal Ed 2001; 84: F106–F110. 42: 348–355. infants? J Matern Fetal Neonatal Med 2004; 81 Chen ML, Guo L, Smith LE et al. High or 55 Ingimarsson J, Bjo¨rklund LJ, Curstedt T et 15: 75–84. low oxygen saturation and severe retinopa- al. Incomplete protection by prophylactic 69 ILCOR. The International Liaison Commit- thy of prematurity: a meta-analysis. Pediat- surfactant against the adverse effects of tee on Resuscitation (ILCOR) consensus on rics 2010; 125: e1483–e1492. large lung inflations at birth in immature science with treatment recommendations for 82 Evans JR, Short BL, Van Meurs K et al. lambs. Intensive Care Med 2004; 30: 1446– pediatric and neonatal patients: neonatal Cardiovascular support in preterm infants. 1453. resuscitation. Pediatrics 2006; 117: 978–988. Clin Therap 2006; 28: 1366–1384. 56 Hillman NH, Moss TJ, Kallapur SG et al. 70 Short JA, Van der Walt J. Oxygen in neo- 83 Suominen PK, Dickerson HA, Moffett BS Brief, large tidal volume ventilation initiates natal and infant anesthesia – current prac- et al. Hemodynamic effects of rescue proto- lung injury and a systemic response in fetal tice in the UK. Pediatr Anesth 2008; 18: col hydrocortisone in neonates with low car- sheep. Am J Respir Crit Care Med 2007; 378–387. diac output syndrome after cardiac surgery. 176: 575–581. 71 Lofqvist C, Andersson E, Sigurdsson J et al. Pediatr Crit Care Med 2005; 6: 655–659. 57 Turki M, Young MP, Wagers SS et al. Peak Longitudinal postnatal weight and insulin- 84 McElhinney DB, Hedrick H, Bush D et al. pressures during manual ventilation. Respir like growth factor I measurements in the Necrotizing enterocolitis in infants with con- Care 2005; 50: 340–344. prediction of retinopathy of prematurity. genital heart disease: risk factors and out- 58 Thomas W, Speer CP. Nonventilatory strat- Arch Ophthalmol 2006; 124: 1711–1718. comes. Pediatrics 2000; 106: 1080–1087. egies for prevention and treatment of bron- 72 Filho JB, Bonomo PP, Maia M et al. 85 Kaufman J, Almodovar MC, Zuk J et al. chopulmonary dysplasia – what is the Weight gain measured at 6 weeks after birth Correlation of abdominal site near-infrared evidence? Neonatology 2008; 94: 150–159. as a predictor for severe retinopathy of pre- spectroscopy with gastric tonometry in 59 Ramamoorthy C, Tabbutt S, Kurth CD et al. maturity: study with 317 very low birth infants following surgery for congenital Effects of inspired hypoxic and hypercapnic weight preterm babies. Graefes Arch Clin heart disease. Pediatr Crit Care Med 2008; gas mixtures on cerebral oxygen saturation in Exp Ophthalmol 2009; 247: 831–836. 9: 62–68. neonates with univentricular heart defects. 73 Hellstrom A, Hard AL, Engstrom E et al. 86 AAP Subcommittee on Neonatal Hyperbi- Anesthesiology 2002; 96: 283–288. Early weight gain predicts retinopathy in lirubinemia. Neonatal jaundice and kernicte- 60 Mariani G, Cifuentes J, Carlo WA. Ran- preterm infants: new, simple, efficient rus. Pediatrics 2001; 108: 763–765. domized trial of permissive hypercapnia in approach to screening. Pediatrics 2009; 123: 87 Ehrenkranz RA, Younes N, Lemons JA preterm infants. Pediatrics 1999; 104: 1082– e638–e645. et al. Longitudinal growth of hospitalized 1088. 74 Hellstro¨m A, Ley D, Hansen-Pupp I et al. very low birth weight infants. Pediatrics 61 Pfister RH, Goldsmith JP. Quality improve- New insights into the development of reti- 1999; 104(2 Pt 1): 280–289. ment in respiratory care: decreasing bron- nopathy of prematurity – importance of 88 Radmacher PG, Looney SW, Rafail ST chopulmonary dysplasia. Clin Perinatol early weight gain. Acta Paediatr 2010; 99: et al. Prediction of extrauterine growth 2010; 37: 273–293. 502–508. retardation (EUGR) in VLBW infants. J 62 Vannucci RC, Towfighi J, Brucklacher RM 75 Chan-Ling T, Gock B, Stone J. The effect of Perinatol 2003; 23: 392–395. et al. Effect of extreme hypercapnia on hyp- oxygen on vasoformative cell division. Evi- 89 Clark RH, Thomas P, Peabody J. Extra- oxic–ischemic brain damage in the immature dence that ‘physiological hypoxia’ is the stim- uterine growth restriction remains a serious rat. Pediatr Res 2001; 49: 799–803. ulus for normal retinal vasculogenesis. Invest problem in prematurely born neonates. 63 Fabres J, Carlo WA, Phillips V et al. Both Ophthalmol Vis Sci 1995; 36: 1201–1214. Pediatrics 2003; 111(5 Pt 1): 986–990. extremes of arterial carbon dioxide pressure 76 Hellstrom A, Carlsson B, Niklasson A et al. 90 Ehrenkranz RA, Dusick AM, Vohr BR et al. and the magnitude of fluctuations in arterial IGF-I is critical for normal vascularization Growth in the neonatal intensive care unit carbon dioxide pressure are associated with of the human retina. J Clin Endocrinol influences neurodevelopmental and growth severe intraventricular hemorrhage in pre- Metab 2002; 87: 3413–3416. outcomes of extremely low birth weight term infants. Pediatrics 2007; 119: 299–305. 77 Hellstrom A, Engstrom E, Hard AL et al. infants. Pediatrics 2006; 117: 1253–1261. 64 Miller JD, Carlo WA. Safety and effective- Postnatal serum insulin-like growth factor I 91 Cooke R, Embleton N, Rigo J et al. High ness of permissive hypercapnia in the pre- deficiency is associated with retinopathy of protein pre-term infant formula: effect on term infant. Curr Opin Pediatr 2007; 19: prematurity and other complications of pre- nutrient balance, metabolic status and 142–144. mature birth. Pediatrics 2003; 112: 1016– growth. Pediatr Res 2006; 59: 265–270. 65 Cohen RS. Acidosis and alkalosis. In: 1020. 92 Stephens BE, Walden RV, Gargus RA et al. Stevenson DK, Benitz WE, Sunshine P, 78 Askie LM, Henderson-Smart DJ, Ko H. First-week protein and energy intakes are Hintz SR, Druzin ML, eds. Fetal and Neo- Restricted versus liberal oxygen exposure associated with 18-month developmental natal Brain Injury, 4th edn. New York: for preventing morbidity and mortality in outcomes in extremely low birth weight Cambridge University Press, 2009: 402–408. preterm or low birth weight infants. Cochra- infants. Pediatrics 2009; 123: 1337–1343. 66 Van der Walt J. Oxygen – elixir of life or ne Database Syst Rev 2009; 1: CD001077. 93 Hay WW Jr. Strategies for feeding the pre- Trojan horse? Part 2: oxygen and neonatal DOI: 10.1002/14651858.CD001077.pub2. term infant. Neonatology 2008; 94: 245–254. anesthesia. Pediatr Anesth 2006; 16: 1205– 79 SUPPORT Study Group of the Eunice Ken- 94 Uhing MR, Das UG. Optimizing growth in 1212. nedy Shriver NICHD Neonatal Research the preterm infant. Clin Perinatol 2009; 36: Network. Target ranges of oxygen satura- 165–176.

Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd 551 Open-heart surgery for low birth weight infants G.D. Williams and R.S. Cohen

95 Picca S, Principato F, Mazzera E et al. poptosis: status of the evidence. Anesth 123 Bellinger DC, Rappaport LA, Wypij D Risks of acute renal failure after cardiopul- Analg 2008; 106: 1659–1663. et al. Patterns of developmental dysfunction monary bypass surgery in children: a retro- 109 Loepke AW, McGowan FX Jr, Soriano SG. after surgery during infancy to correct trans- spective 10-year case-control study. Nephrol CON: the toxic effects of anesthetics in the position of the great arteries. J Dev Behav Dial Transplant 1995; 10: 630–636. developing brain: the clinical perspective. Pediatr 1997; 18: 75–83. 96 Bailey D., Phan V., Litalien C. et al. Risk Anesth Analg 2008; 106: 1664–1669. 124 Bellinger DC, Wypij D, du Duplessis AJ et factors of acute renal failure in critically ill 110 Gaynor JW, Wernovsky G, Jarvik GP et al. al. Neurodevelopmental status at eight years children: a prospective descriptive epidemio- Patient characteristics are important deter- in children with dextro-transposition of the logical study. Pediatr Crit Care Med 2007; minants of neurodevelopmental outcome at great arteries: the Boston Circulatory Arrest 8: 29–35. one year of age after neonatal and infant Trial. J Thorac Cardiovasc Surg 2003; 126: 97 Picca S, Ricci Z, Picardo S. Acute kidney cardiac surgery. J Thorac Cardiovasc Surg 1385–1396. injury in an infant after cardiopulmonary 2007; 133: 1344–1353. 125 Wright M, Nolan T. Impact of cyanotic bypass. Semin Nephrol 2008; 28: 470–476. 111 Chen J, Zimmerman RA, Jarvik GP et al. heart disease on school performance. Arch 98 Chan KL, Ip P, Chiu CS et al. Peritoneal Perioperative stroke in infants undergoing Dis Child 1994; 71: 64–70. dialysis after surgery for congenital heart open heart operations for congenital heart 126 Macrae DJ, Field D, Mercier JC et al. disease in infants and young children. Ann disease. Ann Thorac Surg 2009; 88: 823–829. Inhaled nitric oxide therapy in neonates and Thorac Surg 2003; 76: 1443–1449. 112 Clapp W. Developmental regulation of the children: reaching a European consensus. 99 Guignard JP. Nephrotoxicity of vasoactive immune system. Semin Perinatol 2006; 30: Intensive Care Med 2004; 30: 372–380. drugs in the fetus and the newborn infant. 69–72. 127 Williams GD, Bratton SL, Riley EC et al. Therapie 1988; 43: 407–411. 113 Kettner SC, Pollak A, Zimpfer M et al. Coagulation tests during cardiopulmonary 100 Cattarelli D, Spandrio M, Gasparoni A Heparinase-modified thrombelastography in bypass correlate with blood loss in children et al. A randomised, double blind, placebo term and preterm neonates. Anesth Analg undergoing cardiac surgery. J Cardiothorac controlled trial of the effect of theophylline 2004; 98: 1650–1652. Vasc Anesth 1999; 13: 398–404. in prevention of vasomotor nephropathy in 114 Valieva OA, Strandjord TP, Mayock DE 128 Costello JM, Polito A, Brown DW et al. very preterm neonates with respiratory dis- et al. Effects of transfusions in extremely Birth before 39 weeks’ gestation is associ- tress syndrome. Arch Dis Child Fetal Neona- low birth weight infants: a retrospective ated with worse outcomes in neonates with tal Ed 2006; 91: F80–F84. study. J Pediatr 2009; 155: 331–337. heart disease. Pediatrics 2010; 126: 101 Bhat MA, Shah ZA, Makhdoomi MS et al. 115 Huebler M, Redlin M, Boettcher W et al. e277–e284. Theophylline for renal function in term neo- Transfusion-free arterial switch operation in 129 Clancy RR, McGaurn SA, Wernovsky G nates with perinatal asphyxia: a randomized, a 1.7 kg premature neonate using a new et al. Preoperative risk of death prediction placebo-controlled trial. J Pediatr 2006; 149: miniature cardiopulmonary bypass system. J model in heart surgery with deep hypother- 180–184. Card Surg 2008; 23: 358–360. mic circulatory arrest in the neonate. J Tho- 102 Hack M, Taylor HG, Drotar D et al. 116 Jonas RA, Wypij D, Roth SJ et al. The rac Cardiovasc Surg 2000; 119: 347–357. Chronic conditions, functional limitations, influence of hemodilution on outcome after 130 Numa A, Butt W, Mee RBB. Outcome of and special health care needs of school-aged hypothermic cardiopulmonary bypass: infants with birthweight 2000 g or less who children born with extremely low-birth- results of a randomized trial in infants. J undergo major cardiac surgery. J Paediatr weight in the 1990s. JAMA 2005; 294: Thorac Cardiovasc Surg 2003; 126: Child Health 1992; 28: 318–320. 318–325. 1765–1774. 131 Lechner E, Wiesinger-Eidenberger G, Weiss- 103 Williams GD, Ramamoorthy C. Brain mon- 117 Williams GD, Ramamoorthy C, Chu L ensteiner M et al. Open-heart surgery in itoring and protection during pediatric et al. Modified and conventional ultrafiltra- premature and low-birth-weight infants – a cardiac surgery. Semin Cardiothorac tion during pediatric cardiac surgery: clinical single-centre experience. Eur J Cardiothorac Vasc Anesth 2007; 11: 23–33. outcomes compared. J Thorac Cardiovasc Surg 2009; 36: 986–991. 104 Bellinger DC, Wypij D, Kuban KC et al. Surg 2006; 132: 1291–1298. 132 Miyamoto T, Sinzobahamvya N, Photiadis J Developmental and neurological status of 118 Durrmeyer X, Vutskits L, Anand KJS et al. et al. Survival after surgery with cardiopul- children at 4 years of age after heart surgery Use of analgesic and sedative drugs in the monary bypass in low weight patients. Asian with hypothermic circulatory arrest or low- NICU: integrating clinical trials and labora- Cardiovasc Thorac Ann 2008; 16: 115–119. flow cardiopulmonary bypass. Circulation tory data. Pediatr Res 2010; 67: 117–127. 133 Netz BC, Hoffmeier A, Krasemann T et al. 1999; 100: 526–532. 119 McCann ME, Soriano SG. Is anesthesia Low weight in congenital heart surgery: is it 105 Mahle WT, Wernovsky G. Neurodevelop- bad for the newborn brain? Anesthesiol Clin the right way? Thorac Cardiovasc Surg 2005; mental outcomes in hypoplastic left heart 2009; 27: 269–284. 53: 330–333. syndrome. Semin Thorac Cardiovasc Surg 120 Sanders RD, Ma D, Brooks P et al. Balanc- 134 Abrishamchian R, Kanhai D, Zwets E et al. Pediatr Cardiac Surg Ann 2004; 7: 39–47. ing paediatric anaesthesia: preclinical Low birth weight or diagnosis, which is 106 Ballweg JA, Wernovsky G, Gaynor JW. insights into analgesia, hypnosis, neuropro- higher risk?-a meta-analysis of observational Neurodevelopmental outcomes following tection, and neurotoxicity. Br J Anaesth studies. Eur J Cardiothorac Surg 2006; 30: congenital heart surgery. Pediatr Cardiol 2008; 101: 597–609. 700–705. 2007; 28: 126–133. 121 Loepke AW, Soriano SG. An assessment of 135 Pizarro C, Davis DA, Galantowicz ME 107 Loepke AW. Developmental neurotoxicity the effects of general anesthetics on develop- et al. Stage I palliation for hypoplastic left of sedatives and anesthetics: a concern for ing brain structure and neurocognitive func- heart syndrome in low birth weight neo- neonatal and pediatric critical care medicine. tion. Anesth Analg 2008; 106: 1681–1707. nates: can we justify it? Eur J Cardiothorac Pediatr Crit Care Med 2010; 11: 217–226. 122 Wilder RT, Flick RP, Sprung J et al. Early Surg 2002; 21: 716–720. 108 Jevtovic-Todorovic V, Olney JW. PRO: exposure to anesthesia and learning disabili- 136 Weinstein S, Gaynor JW, Bridges ND et al. anesthesia-induced developmental neuroa- ties in a population-based birth cohort. Early survival of infants weighing 2.5 kilo- Anesthesiology 2009; 110: 796–804. grams or less undergoing first-stage recon-

552 Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd G.D. Williams and R.S. Cohen Open-heart surgery for low birth weight infants

struction for hypoplastic left heart syn- Norwood procedure. Circulation 2006; 114: palliation? Eur J Cardiothorac Surg 2008; drome. Circulation 1999; 100(Suppl II): II- 594–599. 33: 613–618. 167–II-170. 138 Pizarro C, Derby CD, Baffa JM et al. 139 Dorfman AT, Marino BS, Wernovsky G et 137 Reinhartz O, Reddy VM, Petrossian E et al. Improving the outcome of high-risk neo- al. Critical heart disease in the neonate: pre- Homograft valved right ventricle to pulmo- nates with hypoplastic left heart syndrome: sentation and outcome at a tertiary care cen- nary artery conduit as a modification of the hybrid procedure or conventional surgical ter. Pediatr Crit Care Med 2008; 9: 193–202.

Pediatric Anesthesia 21 (2011) 538–553 ª 2011 Blackwell Publishing Ltd 553 Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Total intravenous anesthesia (TIVA) in pediatric cardiac anesthesia Grace L. S. Wong & Neil S. Morton

Department of Anesthesia, Royal Hospital for Sick Children, Yorkhill, Glasgow, UK

Keywords Summary total intravenous anesthesia; pediatrics; Although inhalational anesthesia with moderate- to high-dose opioid anal- cardiac surgery; propofol gesia has been the mainstay of pediatric cardiac anesthesia, the availability of new short-acting drugs, new concepts in pharmacokinetic modeling and Correspondence Grace L. S. Wong, computer technology, and advances in surgery and perfusion have made Flat F, 9/Fl, Block 6, Beacon Heights, 6 total intravenous anesthesia (TIVA) an attractive option. In this article, we Lung Ping Road, Kowloon Tong, Hong Kong review some of the TIVA techniques used in pediatric cardiac anesthesia. Email: [email protected]

Section Editor: Chandra Ramamoorthy

Accepted 18 February 2011 doi:10.1111/j.1460-9592.2011.03565.x

anesthetic agents. These effects may be additive, for Introduction example, cerebral metabolic depression, in addition Total intravenous anesthesia (TIVA) (1,2) is useful for to that caused by hypothermia, may be produced certain pediatric cardiac procedures and has become with thiopentone or propofol or isoflurane titrated to more feasible with improved infusion pump design, EEG suppression during CPB (3,10). With thiopen- more appropriate pediatric software, and more clinical tone, this may result in reduced cerebral damage experience. There is a better understanding of the (3,11). However, to achieve EEG suppression, a large pharmacodynamics (PD) and pharmacokinetics (PK) dose of thiopentone is usually required, which can of intravenous anesthetic agents in children undergoing produce hemodynamic instability, prolonged anes- cardiopulmonary bypass (CPB) and potential beneficial thetic effects, delayed tracheal extubation, and effects of these agents. increased sedation during the first few postoperative days (11). Pharmacodynamic effects of anesthetic agents during cardiac surgery Cardiovascular effects of ketamine, propofol, and opioids CPB may be associated with multiorgan damage, resulting in both immediate- and long-term effects Ketamine (3,4). This damage may be related in part to inade- Ketamine is frequently chosen for anesthetic induction quate regional tissue oxygen delivery and also to the in patients with cyanotic conditions, as it increases sys- ‘stress response’ during and after CPB (3). Reduction temic vascular resistance and cardiac output but with- in the stress response to cardiac surgery in neonates out worsening right-to-left shunting (12). A recent and children is associated with improved postopera- prospective, randomized study by Tugrul et al. (13) tive outcome (3,5–9). Damage caused by inadequate demonstrated that ketamine anesthesia provided more oxygen delivery and tissue hypoxia may be reduced stable precardiopulmonary bypass conditions than by metabolic depression using both hypothermia and isoflurane anesthesia. Arterial oxygen tension, oxygen

560 Pediatric Anesthesia 21 (2011) 560–566 ª 2011 Blackwell Publishing Ltd G.L.S Wong and N.S. Morton TIVA in pediatric cardiac anesthesia saturation, and mean arterial pressure were better Preconditioning maintained poststernotomy in those patients receiving ketamine (13). Ketamine does not significantly increase Previous exposure of the brain and myocardium to the pulmonary vascular resistance in children with pul- minor insults, chemicals, or pharmacological agents can monary hypertension (14). ‘precondition’ or increase the tolerance of neural and cardiac cells to lethal ischemic injury. Preconditioning Propofol with either volatile or intravenous agents may be Propofol has a short half-life and context-sensitive advantageous. Some volatile agents, most notably iso- half-time (CSHT), and when given by infusion with flurane, when given before an ischemic insult have been opioids, propofol can significantly attenuate the found to induce ischemic tolerance in the brain and adverse hemodynamic and metabolic effects of spinal cord (23–26) and to prevent ischemia/reperfusion bypass and surgery while producing rapid smooth (I/R) myocardial injury (27). There are several proposed recovery and transition to postoperative care (1). mechanisms for protection (e.g., attenuating calcium Propofol reduces whole body oxygen uptake during overload, anti-inflammatory and antioxidant effects, hypothermic CPB (28C) in adults (3,15) and reduces pre- and postconditioning-like protection) (22). Anes- cerebral perfusion pressure, oxidative stress, cerebral thetic preconditioning with volatile anesthetics blood flow, velocity, and microembolic delivery. improves the recovery of contractile function and Propofol has also been shown to suppress seizures reduced calcium overload after I/R (28). Clarke et al. (3,16,17) and may act as an oxygen free radical (29) showed recently that a critical factor mediating rep- scavenger (3,18). Propofol during CPB caused signifi- erfusion injury of the heart is the mitochondrial perme- cant increases in mixed venous oxygen saturation, ability transition pore (MPTP). The opening of the significant reductions in systemic oxygen uptake, MPTP causes mitochondrial swelling with release of reduction in glucose and cortisol concentrations with proapoptotic proteins and uncoupling of mitochondrial no difference in lactate concentrations, mean arterial oxidative phosphorylation. The resulting ATP depriva- pressure during CPB, and inotrope requirement after tion causes the disruption of ionic homeostasis and con- CPB, or in recovery times (3). Catecholamine and tractile function and ultimately sarcolemma rupture cortisol are released during stress, which produces and necrosis. Inhibition of MPTP opening during reper- alterations in microvascular permeability and multi- fusion protects hearts from reperfusion injury. organ damage (3,19,20). These hormonal changes Propofol has been shown to protect the heart also result in hyperglycemia, which is known to wor- against cardiac insults in a variety of experimental sen neurological outcome after cerebral ischemia (3). models (30–32). These effects are attributed to its abil- During rewarming and normothermia, propofol can ity to act as a free radical scavenger (33), enhancing prevent hyperglycemia, attenuate the stress response, tissue antioxidant capacity (34), and through inhibition and produce rapid recovery (3). Propofol can be of plasma membrane calcium channels (35,36). Its anti- used in patients with or at risk of long QT oxidant properties are responsible for its inhibitory syndrome without increasing the risk of torsades de action of MPTP opening in the Langendorff-perfused pointes (TdP) (21). rat heart (30) and its antiapoptotic properties (36). The Propofol produces arterial and venous dilation, both clinical benefits of propofol’s antioxidant capacity in systemic and in pulmonary circulation. These effects during CPB are more evident when using a high will be more prominent with the concomitant use of maintenance dose (plasma levels of approximate other vasodilators like alpha blockers, phosphodiester- 4.2 ngÆml)1) by Xia et al. (37). Propofol protects the ase III inhibitors, and angiotensin-converting enzyme myocardium against I/R injury, owing to its antioxi- inhibitors. These circulatory effects can be modulated dant effect and inhibition of the MPTP. by optimizing the patient’s intravascular volume, surgi- cal stimulation, propofol dose increments, and the rate Cardiopulmonary bypass and pharmacokinetics of dose changes. of anesthetic agents Opioids During CPB, a variety of factors act to alter drug Opioids do not cause myocardial depression and pro- disposition, metabolism, and elimination to a clinically tect the heart by a preconditioning-like mechanism, significant extent. At the onset of CPB, the bypass and therefore, concurrent administration of opioids circuit priming fluid is mixed with the patient’s blood. with either propofol or volatile agents produces an The volume of the prime and reservoir varies with age additive effect (22). from around 350 ml in the neonate to 1500 ml in the

Pediatric Anesthesia 21 (2011) 560–566 ª 2011 Blackwell Publishing Ltd 561 TIVA in pediatric cardiac anesthesia G.L.S Wong and N.S. Morton adult. This prime may be crystalloid or crystalloid and third compartments would be 4600, 1340, and mixed with albumin or blood. This mixture between 8200 ml, respectively (43). Developments are underway patient’s blood and bypass circuit priming fluid will to allow effect-site targeting in children (2). lead to a reduction in the patient’s packed red cell vol- ume (PCV) to approximately 25%; plasma volume will Opioids be increased by 40–50%. All these changes may alter drug protein binding and distribution (38). This results All opioids show a decrease in total drug concentration in reduction in total drug concentration, although not on commencing CPB. The degree of this decrease is always free drug. The potential changes in acid/base greater with fentanyl where a significant proportion of balance during CPB will result in changes in ionized the drug adheres to the surface of the CPB circuit (38). and unionized drug concentrations, which affects pro- The decrease is least with opioids that have a high vol- tein binding. The ratio of the CPB circuit volume to the ume of distribution when the addition of the prime vol- patients’ blood volume is greater for smaller patients ume is less important and in those that can equilibrate than that for adults; consequently, this hemodilution rapidly to minimize the dilution effect (38). For alfenta- effect and its effects on protein binding, volume of dis- nil, fentanyl, and sulfentanil (i.e. drugs that undergo tribution, and drug clearance may be more pronounced hepatic biotransformation and elimination), CPB is in children as opposed to adults (39). associated with marked changes in PK properties. After The mean arterial pressure is determined by the CPB, the elimination half-life of fentanyl is prolonged, pump speed and the systemic vascular resistance which plasma clearance is decreased, and volume of distribu- can be altered by the use of vasodilators and vasocon- tion is increased (39,40). For alfentanil, CPB appears to strictors. Regional blood flow distribution and thus the prolong the elimination half-life and increase the vol- drug distribution and metabolism can be varied. Hypo- ume of distribution, whereas clearance is unchanged thermia reduces hepatic and renal enzyme function, (39,44). Free alfentanil concentrations remain relatively which affects drug metabolism. Many drugs bind to stable throughout CPB and pharmacologically active components of the CPB circuit (e.g., fentanyl) (39,40). concentration remains unchanged (38). Remifentanil There are important developmental influences on body kinetics appears to be minimally affected by CPB with composition (lipid, body water), body proportions, no change in the volume of distribution at steady state, blood–brain barrier maturation, regional blood flow the volume of the central compartment, or the elimina- distribution, hepatic and renal functions which must tion half-life (39). Target effect site concentration of be considered. remifentanil commonly used in TIVA varies. A target of 2–3 lgÆl)1 is adequate for laryngoscopy, 6–8 lgÆl)1 for laparotomy, and 10–12 lgÆl)1 for ablating the stress Propofol response associated with cardiac surgery (45). Conflicting results have been obtained for propofol. CSHT is defined as the time taken for blood plasma The total concentration of propofol may decrease on concentration of a drug to decline by 50% after an commencing CPB with increase in the free fraction, or infusion designed to maintain a state has stopped. the total concentration may remain unchanged. A pro- CSHT is important when TIVA is used. Fentanyl has longed elimination half-life has been demonstrated in a short CSHT when given by infusion for a short time, one study, but the redistribution half-life was short; but this dramatically increases as the duration of the concentration decreased rapidly after stopping the drug infusion increases. Alfentanil’s CSHT becomes con- and patients made a rapid recovery. In general, the free stant after 90-min infusion (Table 1). Remifentanil in active concentrations of these drugs remain unchanged contrast is context-insensitive with offset time indepen- but their action may be prolonged (38,41,42). dent of the duration of infusion because of its unique In a prospective trial of the accuracy of a pediatric elimination by esterase cleavage. target-controlled infusion (TCI) of propofol which employed the Paedfusor pharmacokinetic dataset (43), Table 1 Context-sensitive half-times (CSHT) of opioids in children children did not show a significant change in PD effect [adopted from Mani and Morton (2).]

(level of consciousness), because, even in small chil- Infusion duration (min) 10 100 200 300 600 dren, the volume of bypass circuit and reservoir is small in comparison with the volume of the central Remifentanil 3–6 3–6 3–6 3–6 3–6 and other compartments. For example, in a child Alfentanil 10 45 55 58 60 Fentanyl 12 30 100 200 weighing 10 kg, the volume of prime would be 270 ml, Sulfentanil 20 25 35 60 whereas the volumes of the theoretical central, second,

562 Pediatric Anesthesia 21 (2011) 560–566 ª 2011 Blackwell Publishing Ltd G.L.S Wong and N.S. Morton TIVA in pediatric cardiac anesthesia

build up long-chain acyl-carnitine intermediates. Prop- Use of target-controlled infusion of propofol in ofol infusion syndrome can manifest as unexplained children severe metabolic, usually lactic acidosis, rhabdomyoly- TIVA in children has been reviewed in detail recently sis, cardiac failure, renal failure, and usually of high (2). TIVA can be delivered either by using a manual mortality. In contrast to adults, higher target plasma infusion scheme or by using a TCI. TCI uses a real- and effect site concentration is needed for children to time PK model to calculate the bolus dose and infusion induce anesthesia and takes longer to reach the peak rates to achieve a user-defined target blood or effect effect (2,59) and slower recovery from propofol infu- site concentration. This is achieved by an infusion sion (2). This creates potential problems such as lipid pump controlled by a microprocessor, which incorpo- overload and propofol infusion syndrome, especially rates PK models with age-appropriate parameters. after a long cardiac surgery. A healthy child requires Comparative studies between TCI and manual infusion 2–3 gÆkg)1Æday)1 of lipid per day, equivalent to 4 mgÆkg)1Æ have shown better hemodynamic stability, lower induc- h)1 of 1% propofol in 10% soya oil. Lipid overload can tion dose, and faster recovery with TCI (46–49). be overcome by using 2% propofol solution, or by reduc- ing soya oil to 5% from 10%, or by using propofol- sparing measures such as premedication, regional block- Pediatric pharmacokinetic models ade, and/or concurrent use of systemic opioids; using There are several PK models for pediatric TCI devices, TCI, monitoring depth of anesthesia, and tapering the among them, with Paedfusor and Kataria datasets the infusion toward end of case, if possible, could allow more most often used (50). TCI with propofol is limited to rapid recovery and may reduce such complications in the age group of 3 years or more for most models (51). children (2). Recently, the Paedfusor has been modified to allow use Propofol induces a drop in systemic vascular resis- down to 1 year of age and a lower weight limit of 5 kg. tance, which may worsen the hemodynamics in The Paedfusor is a prototype TCI system developed patients with aortic stenosis, tetralogy of fallot, hyper- from a model concept developed by Schuttler whereby trophic obstructive cardiomyopathy, balanced circula- the clearance is adjusted for age. In children <12 years tion, or right-to-left shunt. of age, clearance increases as age decreases (2). In the Also, there are considerable gaps in PK models for Paedfusor, the central compartment volume and clear- some drugs for ill children and for young children, ance have a nonlinear correlation with weight and the infants, and neonates, so caution is needed when size of the central compartment is quite larger compared applying such programs to these populations. Future with the Marsh pediatric (52) and Schuttler models. models will incorporate more sophisticated PK/PD Thus, the user enters weight and age and the pump algorithms. Hence, when using the TCI technique at automatically enters the correct microconstants into the present, the anesthesiologist must still use knowledge three-compartment PK algorithm. The accuracy of the and experience to titrate the intravenous agents to Paedfusor system has been prospectively evaluated in 29 effect to avoid awareness, pain, and adverse effects (2). children aged from 1 to 15 scheduled for cardiac proce- dures (43,46,51); general anesthesia was provided using Future of TIVA/TCI propofol administered by the Paedfusor system. Accu- racy of the system was evaluated by obtaining up to The wide variability of PK parameters in the pediatric nine arterial samples for the measurement of propofol population compared with adults (46,60) suggests that concentration both during anesthesia and in the recov- the ability of a mathematical model to predict propo- ery period. The predictive indices of median perfor- fol cerebral effect is limited. This variability may be mance error and median absolute performance error of attenuated by using an open-loop TCI with clinical the Paedfusor system were found to be much better feedback input from EEG analysis, may play this role. than those found with the adult ‘Diprifusor’ system. Despite the lack of pediatric algorithm in the bispectral index (BIS) monitor, several studies have demonstrated a very good correlation between BIS values and mea- Potential problems with propofol sured or estimated propofol concentrations in children Propofol infusion syndrome is a metabolic disease usu- (59,61,62). Jeleazcov also reported that BIS can be ally associated with high dose (53–56) or long duration used to monitor anesthetic effect produced by propofol (55,57,58) of propofol and low carbohydrate intake. in children above 1 year (63). With the impairment of mitochondrial fatty acid Currently, closed-loop TCI devices remain the pur- oxidation, ATP production will be reduced and thus view of research laboratories and have not reached the

Pediatric Anesthesia 21 (2011) 560–566 ª 2011 Blackwell Publishing Ltd 563 TIVA in pediatric cardiac anesthesia G.L.S Wong and N.S. Morton open market for use in humans. One major advance application. Alternatively, we can perform inhalational reported recently was the application of mass spectro- induction with sevoflurane prior to iv catheter inser- scopy to measure propofol concentrations in the tion. To reduce injection pain of 2% propofol, we exhaled breath of adult and pediatric patients preinject lignocaine 0.2 mgÆkg)1 and alfentanil 10– (64).There is a very good correlation between the 20 lgÆkg)1, wait for around 1 min, and then start a exhaled and blood concentrations of propofol. blood-targeted infusion of 2% propofol using an Alar- The importance of interindividual variability that is PK pump programmed with the Paedfusor dataset. characterizes the children population may be investi- An initial set blood target concentration of 2 lgÆml)1 gated using covariates as weight, age, and height to delivers a loading bolus dose of approximately describe metabolic processes during physiological 1mgÆkg)1. We assess the patient’s response and titrate development in pediatric PK/PD modeling (46). up the target concentration by increments of 1 lgÆml)1 Possible new formulations of propofol include emul- to achieve the desired induction conditions. Pancuroni- sions that contain medium-chain rather than long-chain um is used for muscle relaxation as it counteracts any triglycerides (to possibly avoid propofol infusion syn- tendency of the induction agents to reduce heart rate. drome), formulations with greater concentrations of We taper our TCI target during vascular access. propofol (2%), and water solution emulsions (micro- Alfentanil will be titrated manually prior to sterno- emulsions and phosphorylated propofol pro-drugs) (65). tomy to a total of 50 lgÆkg)1 and then maintained Refinement of fast-tracking cardiac anesthesia tech- with an infusion of 1–5 lgÆkg)1Æmin)1; alfentanil is niques may be useful in the future. Experience with maintained during surgery and bypass with additional TIVA techniques in cardiac anesthesia is increasing. increments of 10 lgÆkg)1 prior to surgical stimulation, The next step is in making TCI equipment and pediatric sternotomy, chest closure, insertion etc. The software more widely available. propofol target is maintained around 3 lgÆml)1 during bypass. As propofol is a vasodilator, systemic vascular resistance is reduced and caution must be exercised Clinical example of TIVA for pediatric cardiac during concomitant use of other vasodilators such as anesthesia in the Royal Hospital for Sick milrinone. Morphine and midazolam are typically Children, Glasgow given 30–60 min prior to discharge to the pediatric After preoperative assessment and patient preparation, intensive care unit to assist transitioning to ongoing we insert an intravenous cannula after EMLA cream sedation and analgesia.

References 1 Anderson BJ, Hodkinson B. Are there still the metabolic and endocrine stress responses 12 Miller R. Anaesthesia. In: Greeley W, limitations for the use of target-controlled of newborn infants undergoing operation? Kern F, eds. Anesthesia for Pediatric Car- infusion in children? Curr Opin Anaesthesiol Br Med J 1988; 296: 668–672. diac Surgery, 4th edn. New York: Churchill 2010; 23: 356–362. 7 Anand KJ, Hansen DD, PR H. Hormonal- Livingstone, 1994: pp. 1811–1850. 2 Mani V, Morton NS. Overview of total metabolic stress responses in neonates 13 Tugrul M, Camci E, Pembeci K. Ketamine intravenous anesthesia in children. Pediatr undergoing cardiac surgery. Anaesthesiology infusion versus isoflurane for the mainte- Anesth 2010; 20: 211–222. 1990; 73: 661–670. nance of anesthesia in the prebypass 3 Laycock GJA, Mitchell IM, Paton RD et al. 8 Anand KJ, Aynsley-Green A. Measuring period in children with tetralogy of Fallot. EEG burst suppression with propofol dur- the severity of surgical stress in newborn J Cardiothorac Vasc Anesth 2000; 14: 557– ing cardiopulmonary bypass in children: a infants. J Pediatr Surg 1988; 23: 297–305. 561. study of the haemodynamic, metabolic and 9 Anand KJ, Hickey PR. Halothane-mor- 14 Williams GD, Philip BM, Chu LF et al. endocrine effects. Br J Anaesth 1992; 69: phine compared with high-dose sulfentanil Ketamine does not increase the pulmonary 356–362. for anesthesia and postoperative analgesia vascular resistance in children with pulmo- 4 Tinker JH. Cardiopulmonary bypass: in neonatal cardiac surgery. N Engl J Med nary hypertension. Anesth Analg 2007; 105: current concepts and controversies. In: Go- 1992; 326: 1–9. 1578–1584. view AV, ed. Central Nervous System Com- 10 Woodcock TE, Murkin JM, Farrar JK et al. 15 Laycook GJA, Alston RP. Propofol during plications After Cardiopulmonary Bypass. Pharmacological EEG supression during car- hypothermic cardiopulmonary bypass: Philadelphia: W.B. Sauders, 1989: pp. 41– diopulmonary bypass: cerebral hemodynamic effects on hemodynamics and metabolism. 68. and metabolic effects of thiopental or isoflu- J Cardiothorac Anesth 1990; 4: 105. 5 Anand KJS, Sippell WG, Aynsley-Green A. rane during hypothermia and normorther- 16 Lowson S, Gent JP, Goodchild CS. Anti- Randomised trial of fentanyl anesthesia in mia. Anaesthesiology 1987; 67: 218–224. convulsant properties of propofol and thio- preterm babies undergoing surgery: effects on 11 Nussmeier NA, Arlund C, Alston RP. EEG pentone: comparison using tow tests in the stress response. Lancet 1987; 1: 243–247. suppression induced by propofol during laboratoratory mice. Br J Anaesth 1990; 64: 6 Anand KJS, Sippell WG, Schofield NM hypothermic cardiopulmonary bypass. 59–63. et al. Does halothane anesthesia decrease J Cardiothorac Anesth 1990; 4: 106.

564 Pediatric Anesthesia 21 (2011) 560–566 ª 2011 Blackwell Publishing Ltd G.L.S Wong and N.S. Morton TIVA in pediatric cardiac anesthesia

17 Mackiezie S, Kapadia F, Grant IS. Propofol transition pore opening by ischemic precon- 43 Absalom A, Amutike D, Lai A et al. Accu- infusion for control of status epilepticus. ditioning is probably mediated by reduction racy of the ‘‘Paedfusor’’ in children under- Anesthesia 1990; 45: 1043–1045. of oxidative stress rather than mitochondrial going cardiac surgery or catheterization. 18 Murphy PG, Myers D, Webster NR et al. protein phosphorylation. Circ Res 2008; Br J Anaesth 2003; 91: 507–513. THe antioxidant potential of propofol. Br 102: 1082–1090. 44 Hug CC, Burm AGL, de Lange S. Alfenta- J Anaesth 1991; 67: 209P. 30 Javadov SA, Lim KHH, Kerr PM et al. nil pharmacokinetics in cardiac surgical 19 Tinker JH. Cardiopulmonary bypass: cur- Protection of hearts from reperfusion injury patients. Anesth Analg 1994; 78: 231–239. rent concepts and controversies. In: Kirklin by propofol is associated with inhibition of 45 Sammartino M, Garra R, Sbaraglia F et al. JK, ed. The Postperfusion Syndrome: the mitochondrial permeability transition. Remifentanil in children. Pediatr Anesth Inflammation and the Damaging Effects of Cardiovasc Res 2000; 45: 360–369. 2010; 20: 246–255. Cardiopulmonary Bypass. Philadelphia: 31 Kokita N, Hara A. Propofol attenuates 46 Constant I, Rigouzzo A. Which model for W.B. Sauders, 1989: pp. 131–146. hydrogen peroxide-induced mechanical and propofol TCI in children. Pediatr Anesth 20 Tinker JH. Cardiopulmonary bypass: cur- metabolic derangements in the isolated rat 2010; 20: 233–239. rent Concepts and Controversies. In: Reves heart. Anaesthesiology 1996; 84: 117–127. 47 Hentgen E, Houfani M, Billard V et al. JG, Croughwell N, Jacobs JR et al., eds. 32 Kokita N, Hara A, Abiko Y et al. Propofol Propofol-Sufentanil anesthesia for thyroid Anesthesia During Cardiopulmonary improves functional and metabolic recovery surgery: optimal concentrations for hemo- Bypass: Does It Matter? Philadelphia: W.B. in ischemic reperfused isolated rat hearts. dynamic and electrocephalogram stability, Sauders, 1989: pp. 69–97. Anesth Analg 1998; 86: 252–258. and recovery. Anesth Analg 2002; 95: 597– 21 Whyte SD, Booker PD, Buckley DG. The 33 Stratford N, Murphy P. Antioxidant activity 605. effects of propofol and sevoflurane on the of propofol in blood from anaesthetized 48 Passot S, Servin F, Allary R et al. Target- OT interval and transmural dispersion of patients. Eur J Anaesthesiol 1998; 15: 158– controlled versus manually-controlled infu- repolarization in children. Anesth Analg 160. sion of propofol for direct laryngoscopy and 2005; 100: 71–77. 34 Xia Z, Godin DV, Ansley DM. Propofrol bronchoscopy. Anesth Analg 2002; 94: 1212– 22 Suleiman M-S, Zacharowski K, Angelini enhances ischemic tolerance of middle-aged 1216.

GD. Inflammatory response and cardio- rat hearts: effects on 15-F2t-isoprostane 49 Servin FS. TCI compared with manually protection during open-heart surgery: the formation and tissue antioxidant capacity. controlled infusion of propofol: a multicen- importance of anaesthetics. Br J Pharmacol Cardiovasc Res 2003; 59: 113–121. tre study. Anesthesia 1998; 53: 82–86. 2008; 153: 21–33. 35 Buljubasic N, Marijic J, Berczi V et al. 50 Varveris DA, Morton NS. Target controlled 23 Sang H, Cao L, Qiu P et al. Isoflurane pro- Differential effects of etomidate, propofol, infusion of propofol for induction and duces delayed preconditioning against spinal and midazolam on calcium and potassium maintenance of anaesthesia using the cord ischemic injury via release of free radi- channel currents in canine myocardial cells. paedfusor: an open pilot study. Paediatr cals in rabbits. Anaesthesiology 2006; 105: Anaesthesiology 1996; 85: 1092–1099. Anaesth 2002; 12: 589–593. 953–960. 36 Roy N, Friehs I, Cowan DB et al. Dopamine 51 Absalom A, Kenny G. ‘‘Paedfusor’’ phar- 24 Wang L, Traystman RJ, Murphy SJ. Inha- induces postischemic cardiomyocyte apopto- macokinetic data set. Br J Anaesth 2005; 95: lational anesthetics as preconditioning sis in vivo:an efect ameliorated by propofol. 110–113. agents in ischemic brain. Curr Opin Pharma- Ann Thorac Surg 2006; 82: 2192–2199. 52 Marsh B, White M, Morton NS et al. Phar- col 2008; 8: 104–110. 37 Xia Z, Huang Z, Ansley DM. Large-dose macokinetic model driven infusion of propo- 25 Xiong L, Zheng Y, Wu M et al. Precondi- propofol during cardiopulmonary bypass fol in children. Br J Anaesth 1991; 67: 41–48. tioning with isoflurane produces dose-depen- decreases biochemical markers of myocar- 53 Burow BK, Johnson ME, Packer DL. dent neuroprotection via activation of dial injury in coronary surgery patients: a Metabolic acidosis associated with propofol adenosine triphosphate-regulated potassium Comparison with isoflurane. Anesth Analg in the absence of other causative factors. cahnnels after focal cerebral ischemia in 2006; 103: 527–532. Anaesthesiology 2004; 101: 239–241. rats. Anesth Analg 2003; 96: 233–237. 38 Gedney JA, Ghosh S. Pharmacokinetics of 54 Liolios A, Guerit J-M, Scholtes J-L et al. 26 Zhang HP, Yuan LB, Zhao RN et al. Iso- analgesics, sedatives and anaesthetic agents Propofol infusion syndrome associated with flurane preconditioning induces neuropro- during cardiopulmonary bypass. Br J Ana- short term large dose infusion during surgi- tection by attenuating Ubiquitin-conjugated esth 1995; 75: 344–351. cal anesthesia in an adult. Anesth Analg protein aggregation in a mouse model of 39 Davis PJ, Wilson AS, Siewers RD et al. The 2005; 100: 1804–1806. transient global cerebral ischemia. Anesth effects of cardiopulmonary bypass on remif- 55 Rozet I, Lam AM. Propofol infusion Analg 2010; 111: 506–514. entanil kinetics in children undergoing atrial syndrome or probably overinterpretation 27 Julier K, da Silva R, Garcia C et al. Precon- septal defect repair. Anesth Analg 1999; 89: syndrome? Anaesthesiology 2008; 108: 330. ditioning by sevoflurane decreases biochemi- 904–908. 56 Salengros J-C, Velghe-Lenelle CE, Bollens cal markers for myocardial and renal 40 Bovill JG, Sebel PS. Pharmacokinetics of R et al. Lactic acidosis during propofol- dysfunction in coronary artery bypass graft high-dose fentanyl: a study in patients remifentanil anesthesia in an adult. Anaes- surgery:a double-blinded, placebo-con- undergoing cardiac surgery. Br J Anaesth thesiology 2004; 101: 241–243. trolled, multicenter study. Anaesthesiology 1980; 52: 795–801. 57 Cremer OL, Moons KGM, Bouman EAC 2003; 98: 1315–1327. 41 Massey NJA, Sherry KM, Oldroyd S et al. et al. Long term propofol infusion and car- 28 An J, Bosnjak ZJ, Jiang MT. Myocardial Pharmacokinetics of an infusion of propofol diac failure in adult head-injured patients. protection by isoflurane preconditioning pre- during cardiac surgery. Br J Anaesth 1990; Lancet 2001; 357: 117–118. serves Ca2+ cycling proteins independent of 65: 475–479. 58 Merz TM, Regli B, Rothen HU. Propofol

sarcolemmal and mitochondrial KATP chan- 42 Russell GN, Wright EL, Fox MA et al. infusion syndrome-A fatal case at a low nels. Anesth Analg 2007; 105: 1207–1213. Propofol-fentanyl anaesthesia for coronary infusion dose. Anesth Analg 2006; 103: 1050. 29 Clarke SJ, Khaliulin I, Das M et al. artery surgery and cardiopulmonary bypass. 59 Jeleazcov C, Ihmsen H, Schmidt J et al. Inhibition of mitochondrial permeability Anaesthesia 1988; 44: 205–208. Pharmacodynamic modelling of the bispec-

Pediatric Anesthesia 21 (2011) 560–566 ª 2011 Blackwell Publishing Ltd 565 TIVA in pediatric cardiac anesthesia G.L.S Wong and N.S. Morton

tral index response to propofol-based anaes- tion among children. Pediatrics 2005; 115: producing hypnosis in children and adults: thesia during general surgery in children. 1666–1674. comparison using the bispectral index. Acta Br J Anaesth 2008; 100: 509–516. 62 Rigouzzo A, Girault L, Louvet N et al. Anaesthesiol Scand 2006; 50: 882–887. 60 Hu C, Horstman DJ, Shafer SL. Variability THe relationship between bispectral index 64 Hornuss C, Praun S, Villinger J et al. Real- of target-controlled infusion is less than the and propofol during target-controlled infu- time monitoring of propofol in expired air in variability afterbolus injection. Anaesthesiol- sion anesthesia: a comparative study humans undergoing total intravenous anes- ogy 2005; 102: 639–645. between children and young adults. Anesth thesia. Anaesthesiology 2007; 106: 665–674. 61 Powers KS, Nazarian EB, Tapyrik SA Analg 2008; 106: 1109–1116. 65 Lerman J. TIVA, TCI, and pediatrics: et al. Bispectral index as a guide for titra- 63 Munoz HR, Cortinez LI, Ibacache ME where are we and where are we going? tion of propofol during procedural seda- et al. Effect site concentrations of propofol Pediatric Anesth 2010; 20: 273–278.

566 Pediatric Anesthesia 21 (2011) 560–566 ª 2011 Blackwell Publishing Ltd Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Role of transesophageal echocardiography in the management of pediatric patients with congenital heart disease Komal Kamra1, Isobel Russell2 & Wanda C. Miller-Hance3

1 Lucile Packard Children’s Hospital, Stanford University, Palo Alto, CA, USA 2 University of California San Francisco, San Francisco, CA, USA 3 Baylor College of Medicine, Texas Children’s Hospital, Houston, TX, USA

Keywords Summary congenital heart disease; transesophageal echocardiography Transesophageal echocardiography (TEE) has become a critical diagnostic and perioperative management tool for patients with congenital heart dis- Correspondence ease (CHD) undergoing cardiac and noncardiac surgical procedures. This Komal Kamra, review highlights the role of TEE in routine management of pediatric car- Department of Anesthesia, Stanford diac patient population with focus on indications, views, applications and University School of Medicine, 300 Pasteur technological advances. Drive, Room H3580, Stanford, CA 94305- 5640, USA Email: [email protected]

Section Editor: Chandra Ramamoorthy

Accepted 7 March 2011 doi:10.1111/j.1460-9592.2011.03570.x

stances where other imaging modalities such as trans- Introduction thoracic echocardiography are not diagnostic (23). Over the past two decades, transesophageal echocardio- This review highlights the use of TEE in the pediat- graphy (TEE) has solidified its role as a critical diag- ric age group, focusing on indications, applications, nostic and perioperative management tool for patients complications, and contraindications of this imaging with congenital heart disease (CHD). TEE has proven modality. A brief overview of TEE cross-sections and to be invaluable in confirming preoperative diagnoses, anatomical information derived from these is provided formulating surgical plans, evaluating immediate opera- as relevant to the examination in CHD. Relevant tive results, identifying patients with residual defects, information obtainable from intraoperative TEE in the and guiding surgical revisions (1–8). Technological pre and postbypass settings for selected congenital car- advancements, particularly the use of small probe sizes, diac pathologies is described. Lastly, recent technologi- have significantly improved patient safety and success cal advances related to TEE in this patient population of cardiac surgery in infants and children (9–20). are noted. Transesophageal echocardiography has been shown to play an important role during cardiac catheteriza- Indications and applications of TEE in pediatric tion procedures such as device placement, percutaneous patients with CHD valve replacement (21,22), and intracardiac ablations, as well as during noncardiac surgery in patients with Surgical applications CHD. Imaging through the transesophageal approach Surgical indications for TEE include both cardiac and has been demonstrated to be of benefit in circum- noncardiac procedures.

Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd 479 Role of TEE in pediatric CHD patients K. Kamra et al.

Indications and Applications of TEE

Cardiac Surgical Catheterization Other Uses

Interventional Cardiac Non-Cardiac

Catheter Ablation

Cardiopulmonary No Bypass Bypass

Bypass Pre-Bypass Initiation Post-Bypass

Cardiac surgery Intraoperative use is currently the most common indi- cation of TEE in the pediatric age group. Numerous reports since the technology became available have documented the benefits of this approach, and the experience to date accounts for the incorporation of TEE into the standard of care of patients undergoing surgery for CHD at many centers worldwide.

Precardiopulmonary bypass Figure 2 Coronary sinus. Retroflexion of the probe at the level of Transesophageal echocardiography is a beneficial diag- the mid-esophageal four-chamber view displays the coronary sinus as it courses longitudinally along the atrioventricular groove to drain nostic tool during the prebypass phase of surgery for into the right atrium. An enlarged coronary sinus suggests the pres- CHD. TEE can corroborate preoperative diagnoses ence of a persistent left superior vena cava. RA, right atrium; RV, and define the anatomical abnormalities prior to the right ventricle; CS, coronary sinus; LV, left ventricle. surgical intervention. In case of unsatisfactory trans- thoracic images or incomplete diagnostic data, TEE shunting, and ventricular function, thereby guiding can help acquire the relevant missing information. fluid management, anesthetic drug selection, and the Anesthesiologists can utilize TEE to assess ventricular use of inotropic/vasocative agents. Additional applica- preload, hemodynamic status, intracardiac/great artery

Figure 3 Left pulmonary vein, spectral Doppler. View obtained by Figure 1 Mid-esophageal four-chamber view. All four cardiac starting at the mid-esophageal four-chamber view, slightly with- chambers can be visualized in this view. Applications of this cross- drawing the probe and rotating to the left. Pulsed-wave Doppler section include the evaluation of chamber size and morphology, interrogation of left lower pulmonary vein as it enters the left atrioventricular valve function, intracardiac shunting across the atrial atrium. The spectral Doppler interrogation is guided by color flow and ventricular levels and ventricular function (global and segmen- mapping (red signal represents venous flow), as it drains normally tal). RA, right atrium; TV, tricuspid valve; RV, right ventricle; LA, left into the left atrium. This examination is useful in the evaluation of atrium; MV, mitral valve; LV, left ventricle; IAS, interatrial septum; pulmonary vein obstruction, diastolic dysfunction, mitral regurgita- IVC, inter ventricular septum. tion, and atrial rhythm disturbances.

480 Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd K. Kamra et al. Role of TEE in pediatric CHD patients

Figure 4 Right upper pulmonary veins, color Doppler. View Figure 6 Mid-esophageal aortic valve short axis view. Cross-sec- obtained by starting at the mid-esophageal four-chamber view, tion obtained with the probe in neutral position at the mid esopha- slightly withdrawing the probe and rotating to the right. Color flow gus and transducer rotation to approximately 30. This view Doppler interrogation of the right upper pulmonary veins as these provides anatomical detail of the aortic valve that includes valve enter the left atrium (arrow). Further probe withdrawal and rotation morphology, number of cusps, and commissures. Color Doppler to the right is usually able to display all right-sided pulmonary veins. may be applied to assess for the aliased flow associated with valve RA, right atrium; LA, left atrium; RV, right ventricle; IVS, interven- stenosis or to evaluate a regurgitant jet across the valvar orifice. tricular septum; PV, pulmonary vein. AV, aortic valve; RA, right atrium; LA, left atrium; PV, pulmonary valve; RVOT, right ventricular outflow tract; IAS, interatrial septum.

Figure 5 Mid-esophageal ascending aortic short axis view. View obtained after slight probe withdrawal from the mid-esophageal four-chamber view and anteflexion with probe rotation to between Figure 7 Mid-esophageal bicaval view. View obtained at approxi- 20 and 30. In addition to displaying the anatomical details of the mately 90 in the mid esophagus displays the right atrium, right ascending aorta, main and proximal branched pulmonary arteries, atrial appendage, a portion of the left atrium, the interatrial septum, this view is particularly useful in the assessment of pulmonary out- and the systemic veins as they join the right atrium (left panel). The flow tract obstruction. The ability to align the spectral Doppler cur- Eustachian valve can be visualized in this image at the mouth of sor with the direction of flow allows for optimal measurement the inferior vena cava. Color Doppler interrogation of the interatrial of peak gradients in this setting. RPA, right pulmonary artery; PA, septum in this view complements the evaluation of atrial level com- pulmonary artery. munications (right panel). This view is also useful in optimizing cath- eter position during central venous line placement. RA, right atrium; tions of intraoperative prebypass TEE include confir- LA, left atrium; IAS, interatrial septum; IVC, inferior vena cava; mation of appropriate position of central venous cath- SVC, superior vena cava. eters placed percutaneously (24,25).

Bypass initiation ventricular venting device. During beating-heart sur- Prior to initiation of cardiopulmonary bypass, TEE gery, TEE is helpful in assessing the presence of left- can confirm placement of venous cannulae and left sided intracardiac air, allowing for recognition of

Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd 481 Role of TEE in pediatric CHD patients K. Kamra et al.

Figure 10 Mid-esophageal aortic valve long axis view. This view is obtained at the level of the mid esophagus by rotating the scanning plane to approximately 120–140. The view displays the left atrium, mitral valve, left ventricle, left ventricular outflow tract, aortic valve, and proximal ascending aorta. Portions of the interventricular sep- tum and right ventricle are usually seen. Pathologies that may be Figure 8 Mid-esophageal right ventricular inflow–outflow view. characterized by this view include mitral and aortic valve disease, Cross-section obtained at 60 displays both atria, the inlet and out- left ventricular outflow tract obstruction (discrete and tunnel-type), let portions of the right ventricle; the aortic valve in short axis, the supravalvar aortic disease, and ventricular septal defects. In pulmonary valve and proximal main pulmonary artery. This view is patients with hypertrophic cardiomyopathies, this view provides for essential in evaluating ventricular septal defects and defining their characterization of septal thickness, anterior systolic motion of the specific type. Perimembranous defects are seen in close proximity mitral valve, mitral regurgitation, and the determination of outflow to the tricuspid valve (membranous region). In contrast, outlet tract gradients. LA, left atrium; MV, mitral valve; AV, aortic valve; (conal) ventricular septal defects are anatomically near the pulmo- LVOT, left ventricular outflow tract; LV, left ventricle; RV, right nary valve. This view provides detailed information regarding the ventricle; IVS, interventricular septum. nature and severity of obstruction in lesions that involve the right ventricular outflow tract. In the figure, mild prominence of the conal septum is seen. RA, right atrium; LA, left atrium; TV, tricuspid valve; AV, aortic valve; RVOT, right ventricular outflow tract; RPA, right pulmonary artery.

Figure 11 Transgastric mid short axis view. This cross-section is obtained by advancing the probe to the distal esophagus. Anteflexion is required to maintain adequate probe contact. This view provides for a quick assessment of ventricular filling, as well as global and Figure 9 Mid-esophageal two-chamber view. This cross-section segmental function. This mid papillary ventricular cross-section is displays the left atrium, left atrial appendage, anterior and posterior particularly useful in determining the etiology of hypotension upon mitral valve leaflets, and left ventricle. The anterior and inferior separation from cardiopulmonary bypass. walls of the left ventricle are well seen in this view and can be examined in the evaluation of segmental wall motion abnormalities. LA, left atrium; PMVL, posterior mitral valve leaflet; MV, mitral potential systemic air embolization. TEE can be of valve; LV, left ventricle; AMVL, anterior mitral valve leaflet; LAA, benefit in eliminating the possibility of left ventricular left atrial appendage. dilation because of inadequate venting.

482 Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd K. Kamra et al. Role of TEE in pediatric CHD patients

Figure 12 Deep transgastric long axis view of the left ventricular Figure 14 Sinus venosus atrial septal defect. The mid-esophageal outflow tract. The main applications of this view relate to the assess- bicaval view displays the deficiency in the superior aspect of the ment of ventriculo-arterial connections, left ventricular outflow tract, interatrial septum characteristic of sinus venosus atrial septal aortic valve, and ventricular septum. In the figure, the image display defect (left panel). The defect is located near the entrance of the has been inverted to follow a normal anatomical orientation (superior superior vena cava and overrides the crest of the atrial septum. structures at the top of the screen and inferior ones at the bottom). Color flow imaging displays left to right shunting across the defect This view provides an excellent alignment of the interrogating spec- (right pane). This view is essential in the comprehensive evaluation tral Doppler cursor to the left ventricular outflow tract allowing for of the atrial septum as this particular defect may not be evident in optimal estimation of pressure gradients. LV, left ventricle; RV, right the mid-esophageal four-chamber view. RA, right atrium; LA, left ventricle; AV, aortic valve; LVOT, left ventricular outflow tract. atrium; ASD, atrial septal defect.

Figure 13 Deep transgastric long axis view of the right ventricular Figure 15 Perimembranous ventricular septal defect. In this mid- outflow tract. This view is obtained after advancing the probe into esophageal four-chamber view, a defect is shown in the perimem- the fundus of the stomach. If the left ventricular outflow tract is dis- branous region. The ventricular septal defect is in close proximity played, this cross-section can be obtained by releasing the flexion to the aortic valve. The septal leaflet of tricuspid valve is seen her- on the imaging probe allowing for the normally more anterior right niating through the ventricular communication. This can limit the ventricular outflow tract to be seen. In the figure, the image display magnitude of the shunt in some cases. RA, right atrium; LA, has been inverted to follow a normal anatomical orientation (superior left atrium; LV, left ventricle; RV, right ventricle; VSD, ventricular structures at the top of the screen and inferior ones at the bottom). septal defect. Doppler-derived measurements of right ventricular outflow gradients are facilitated by this view. The evaluation of pulmonary or right ven- surgical repair and in the detection of residual pathol- tricular conduit regurgitation may also be feasible. LV, left ventricle; ogy such as outflow tract obstruction, shunting, PA, pulmonary artery; RVOT, right ventricular outflow tract. valvular regurgitation, and stenosis. Such conditions can lead to difficulty in separation from bypass. IN Postcardiopulmonary bypass addition, TEE is used to facilitate cardiac de-airing Transesophageal echocardiography has been shown to and closely monitor ventricular function after the be of major benefit in assessing the adequacy of the bypass period. This technology is also useful in the

Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd 483 Role of TEE in pediatric CHD patients K. Kamra et al.

Figure 18 Subaortic membrane. In this long axis view of the left ventricular outflow track, a discrete membranous ridge is seen attached to the ventricular septum below the aortic valve. In addi- Figure 16 Complete atrioventricular canal defect (Atrioventricular tion to assessing the gradient across the outflow tract, the exami- septal defect). The figure displays the anatomical defects associ- nation should address aortic valve competence, as aortic ated with a complete canal defect. The defect in the inferior aspect regurgitation may be an associated finding. AV, aortic valve; RV, of the interatrial septum is seen (primum atrial septal defect) just right ventricle. above the level of the common atrioventricular valve. A large inlet- type ventricular septal defect is also visualized. A, right atrium; LA, left atrium; LV, left ventricle; RV, right ventricle; VSD, ventricular septal defect; ASD, atrial septal defect.

Figure 19 Tetralogy of Fallot. Mid-esophageal long axis view in a patient with tetralogy of Fallot demonstrating the hallmark of this defect and anterosuperior deviation of the infundibular septum (left panel). This accounts for malalignment of the conal and trabecular Figure 17 Supravalvar aortic stenosis. Mid-esophageal aortic valve portions of the ventricular septum and leads to crowding of the long axis view depicting narrowing just above the level of the sino- right ventricular outflow tract and associated obstruction. The aortic tubular junction in supravalvar aortic stenosis. This pathology char- root is seen to override the ventricular septal defect. Color flow acterized by a decreased in the caliber of the aorta above the valve interrogation (right panel) is able to determine the magnitude and level may be seen in patients with William syndrome. Evaluation by direction of shunting across the ventricular septal defect. LV, left color Doppler typically demonstrates turbulent flow at the level of ventricle; RV, right ventricle; IVS, interventricular septum; VSD, ven- obstruction. AV, aortic valve. tricular septal defect, assessment of hemodynamic status, thereby guiding fluid management and inotropic therapy (26–30). Postbypass TEE provides valuable information to the Noncardiac surgery anesthesia and critical care teams in formulating post- During noncardiac surgery, patients with CHD, operative management plans. The surgical placement acquired heart disease, heart failure, and pulmonary of transthoracic lines may also be facilitated by the hypertension may present significant challenges to the use of TEE. anesthesiologist. TEE serves as a useful tool in the

484 Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd K. Kamra et al. Role of TEE in pediatric CHD patients

Figure 20 Tetralogy of Fallot. Mid-esophageal right ventricular Figure 22 Ebstein’s anomaly of the tricuspid valve. This mid- inflow-outflow view depicting severe stenosis of the infundibular esophageal four-chamber view demonstrates apical displacement opening and the large ventricular septal defect (left panel). Color of the septal and posterior tricuspid valve leaflets in Ebstein’s Doppler imaging demonstrates turbulent, aliased flow across the anomaly. Relative to the hinge point of the mitral valve at the crux right ventricular outflow tract. This narrowing occurs between the of the heart there is severe displacement of the tricuspid valve anteriorly displaced conal septum and septoparietal trabeculation. causing atrialization of the right ventricle and a reduction in the size RA, right atrium; LA, left atrium; RV, right ventricle; PV, pulmonary of this chamber. LV, left ventricle; RV, right ventricle; RA, right valve; VSD, ventricular septal defect; RVOT, right ventricular out- atrium. flow tract.

Figure 23 Hypoplastic left heart syndrome. Mid-esophageal four- Figure 21 Corrected transposition of the great arteries, postdouble chamber view displaying a hypoplastic left ventricle and mitral switch operation. Mid-esophageal four-chamber view with right- valve. The right ventricle in this anomaly functions as the systemic ward transducer rotation to evaluate the intra-atrial baffle (arrow) chamber. LV, left ventricle; RV, right ventricle. following a double switch operation in corrected transposition. This surgical procedure involves re-routing of the systemic and pulmo- detection of wall motion abnormalities, thereby deter- nary venous return and an arterial switch operation to restore the mining regional function and serving as a monitor for normal ventriculoarterial relation. TEE is able to exclude obstruction myocardial ischemia. Complications of operative pro- across the pulmonary and systemic venous pathways and assess cedures such as pericardial effusions can be detected the integrity of the intra-atrial baffle. LV, left ventricle. by TEE and appropriate drainage if necessary insti- tuted (31–41). clinical management of these patients. As in the case Transesophageal echocardiography during cardiac of cardiac surgery, TEE can assist the anesthesiologist catheterization to effectively optimize ventricular preload, afterload and contractility, and guide inotropic support and Interventional procedures fluid management. In addition to the evaluation of The use of TEE has been documented during cardiac global ventricular function, TEE is useful in the catheterization procedures that involve transcatheter

Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd 485 Role of TEE in pediatric CHD patients K. Kamra et al. device closures of septal defects in the determination of Extra corporeal membrane oxygenation exact position of the defect, size and relationship to sur- In patients undergoing extra corporeal membrane oxy- rounding structures such as atrioventricular valves, genation (ECMO), the thorax may be covered by surgi- superior vena cava right pulmonary veins and aortic cal dressings or the presence of an open sternum may valve (21,22). Imaging via the transesophageal approach preclude transthoracic imaging. In such cases, TEE may has been also shown to assist in positioning of wires, be used in the evaluation of ventricular function, can- catheters, and devices (Clip S1), as well as in assessing nula position and suitability for weaning of support. In residual shunting and encroachment of devices upon the event of left atrial distension, TEE can determine the adjacent structures. Additional benefits include the eval- need for left atrial decompression and facilitate cannula uation of ventricular function and pericardial effusion/ positioning directly into the left atrium or stent place- other procedure-related complications. ment across the interatrial septum if necessary (46). In patients with hypertrophic obstructive cardiomy- opathy, TEE can precisely determine the site of injec- Diagnostic applications tion during alcohol ablation. TEE can be used during Diagnostic applications of TEE in patients with CHD balloon valvuloplasties or laser perforation of valvar include assessment of venous pathways in patients with atresia for monitoring, demonstrating immediate transposition after baffle procedures, aortic root assess- results of the procedures and to exclude any postproce- ment in patients with repaired tetralogy of Fallot and dural complications. other conditions associated with aortic root enlargement, and assessment of ventricular and valvar function in adult patients after Fontan surgery. In the case of endo- In patients with supraventricular arrhythmias, TEE carditis, TEE can precisely locate vegetation(s) or abscess can be used to guide transeptal puncture and sheath and the resulting sequelae in affected areas. TEE can placement into the left atrium for electrophysiologic help determine the size, position and mobility of intracar- mapping and ablation of left-sided pathways. This diac tumors and monitor the effect of this on surround- imaging approach can potentially prevent cardiac per- ing structures and physiological changes related to this. foration. TEE facilitates positioning and manipulation In the postoperative or critical care setting, TEE of catheters used for in nor- may be able to provide diagnostic information that is mal and abnormal hearts, and serves as an adjunct to not otherwise obtainable by the transthoracic echocar- fluoroscopy (42–44). diography because of poor windows, dressings, or limi- tations related to patient size or body habitus. This is Other applications of transesophageal echocardio- of particular benefit in adult patients with CHD. In graphy such cases, TEE can be of benefit in the evaluation of cardiac tamponade, thrombus, ventricular function, Ventricular assist devices volume status, and responsiveness to fluid (47–49). After insertion of ventricular assist devices (VAD), the presence of an intracardiac communication can lead to potential right to left shunting causing hypoxemia. Complications related to transesophageal Valvar insufficiency and stenosis can compromise ade- echocardiography quate device output and cause ventricular dysfunction Extensive experience has demonstrated that TEE can based on pressure and volume overload. As these be safely performed in most pediatric patients with lesions can be addressed at the time of device insertion, minimal complications. Complications related to probe TEE can play an important role in detecting these insertion include failure to intubate the esophagus and structural and functional abnormalities. Other useful vagal stimulation resulting in bradycardia. Although applications of TEE include ensuring appropriate hemodynamic and respiratory complications occur cannulae position and facilitating cardiac de-airing. rarely, it important to recognize that airway compro- TEE can assist in fine-tuning the settings for mechani- mise by compression of airway structures may be seen cal circulatory support and assessment of complica- in smaller patients. In neonates and small infants, the tions after VAD insertion such as pericardial imaging probe can also result in hemodynamic altera- tamponade and endocarditis (45). In the setting of tions. Compression of the descending aorta as well as biventricular dysfunction and placement of single hypotension in neonates with total anomalous pulmon- support device, TEE is of significant benefit in deter- ary venous return due to compression of posterior mining if a second device (biventricular assist devices venous confluence by the probe has been reported. or biVAD) is necessary.

486 Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd K. Kamra et al. Role of TEE in pediatric CHD patients

Probe compression of anomalous vascular structures vascular Anesthesiologists for performance of a com- can also occur, as in the case of a retroesophageal sub- prehensive intraoperative multiplane examination (55). clavian artery, leading to alterations in the pressure This consists of a series of 20 cross-sectional views of monitoring tracing of the corresponding arm during the heart and great vessels. Although these guidelines probe manipulation. assume normal cardiac arrangement (cardiac mass in Other problems such as dental trauma, jaw subluxa- the left chest or levocardia, concordant atrioventricular tion, mucosal erosion, and bleeding have been reported and ventricular connections), they serve as a reference to occur rarely. Prolonged probe use can result in lip point for the pediatric TEE examination. These views and tongue swelling. Unusual but more serious compli- can be modified as necessary to obtain the pertinent cations related to probe manipulation include deep information in most patients with structural heart dis- mucosal lacerations or perforation of the esophagus or ease, both children and adults. stomach. A comprehensive TEE examination in children Although the use of TEE in the young age groups is should include a systematic evaluation of all the considered to be safe and complications occur rarely, cardiac chambers, valves, and vascular structures. The the benefit-risk ratio should be examined in every case study should be systematic and organized so as not to and physicians should be vigilant of the risks associ- miss any relevant information (55). In cases where the ated with this imaging modality and monitor patients study needs to be limited or may need to be terminated accordingly (27,50–54). prematurely because of ventilatory/hemodynamic instability or electrocautery interference, the examina- tion should be focused and targeted to the relevant Contraindications to transesophageal echo- pathology if that can be anticipated. cardiography in the pediatric patient A transthoracic echocardiographic examination in Contraindications to TEE in children as described by patients with CHD follows a segmental approach. This the Task Force of the Pediatric Council of the Ameri- analysis involves defining the thoracoabdominal situs, can Society of Echocardiography are noted in Table 1 cardiac position in the thorax, atrial situs, atrioventric- (27). In the intraoperative setting during cases where ular, and ventriculo-arterial connections. It also esophageal instrumentation is not feasible, consider- includes assessment of ventricular topology and rela- ation may be given to epicardial imaging. Considerable tionship between the great arteries. Most of these ana- experience over many years prior to the availability of tomical details can be confirmed by TEE. The suitable TEE probes in children provided evidence that examination typically includes two-dimensional (2D) this approach is safe and is of benefit in the intraoper- imaging, color flow interrogation, and spectral Dopp- ative assessment of CHD. ler assessment. In some cases, additional modalities such as M-mode echocardiography may provide addi- tional information. Contrast echocardiography using The transesophageal examination in the patient agitated saline or albumin injection into a peripheral with CHD or central vein may be particularly helpful in the deter- Guidelines have been established by the American mination of small intracardiac shunts, residual septal Society of Echocardiography and Society of Cardio- defects, and other vascular variants/pathology that may impact the perioperative management. Table 1 Contraindications to transesophageal echocardiography Anatomic information derived from selected TEE Absolute contraindications Relative contraindications cross-sections is presented in Table 2 as a reference to evaluate normal and abnormal cardiovascular struc- Unrepaired History prior esophageal surgery tures. Details of the pre- and postsurgical evaluation tracheoesophageal fistula of specific congenital lesions are described in Table 3 Esophageal obstruction Esophageal varices or diverticulum (55–61). or stricture Perforated hollow viscus Gastric or esophageal bleeding Poor airway control Vascular ring, aortic arch anomaly Recent advances in transesophageal echocardio- with or without airway compromise graphy Severe respiratory depression Oropharyngeal pathology Uncooperative, unsedated Severe coagulopathy Two important advances in the field echocardiography patient Cervical spine injury or anomaly with applications to pediatric and CHD include the Reprinted from Ref. (27), Copyright (January 2005), with permission development of the micromultiplane TEE probe and from Elsevier. three-dimensional (3D) technology.

Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd 487 Role of TEE in pediatric CHD patients K. Kamra et al.

Table 2 Anatomical information obtainable from selected tee views

Reference view Multiplane angle () Probe manipulation Structures visualized

ME 4 Ch 0–30 Neutral–slight retroflex • Left and right atria (Figure 1, Clip S2) • Left and right ventricles • Atrioventricular valves • Segments of atrial and ventricular septum • Three septal and three lateral left ventricular segments: basal (septal and lateral), mid (septal and lateral), and apical (septal and lateral) walls Start at ME 4 Ch 0–30 Retroflex • Coronary sinus (Figure 2, Clip S3) Start at ME 4 Ch 0–30 Anteflex/slightly withdraw probe • Left pulmonary veins (Figure 3) and rotate to the left Start at ME 4 Ch 0–30 Anteflex/slightly withdraw probe • Right pulmonary veins (Figure 4, Clip S4) and rotate to the right • Superior vena cava • Right atrial appendage ME Asc Ao short 30 Withdraw from ME 4 Ch and • Aorta axis (Figure 5) slightly anteflex • Main pulmonary artery • Origin of branched pulmonary arteries ME Ao valve short 30 Advance probe in neutral position • Aortic valve axis (Figure 6, Clip S5) from upper esophagus to display • Coronary ostia aortic valve in short axis. Withdraw • Pulmonary valve probe to assess coronary origins • Left and right atria and proximal course. • Interatrial septum ME RV inflow–outflow 60–80 Neutral • Right and left atrium (Figure 8, Clip S6) • Tricuspid valve • Right ventricular inflow • Right ventricular outflow tract • Pulmonary valve • Proximal main pulmonary artery ME bicaval 90 Neutral • Left and right atria (Figure 7, Clip S7) • Interatrial septum • Superior and inferior vena cava ME 2 Ch 80–100 Neutral • Left atrial appendage (Figure 9, Clip S8) • Coronary sinus • Mitral valve • Left ventricle • Three anterior and three posterior left ventricular segments: basal (anterior, inferior), mid (anterior, inferior), and apical (anterior, inferior) ME Ao valve long axis 120–140 Neutral • Left atrium (Figure 10, Clip S9) • Mitral valve • Left ventricle • Two anteroseptal and two posterior left ventricular segments: basal (anteroseptal and posterior), and mid (anteroseptal and posterior) walls • Left ventricular outflow tract • Aortic valve and ascending aorta TG mid short axis 0–30 Anteflex • Left and right ventricles (short axis) (Figure 11, Clip S10) • Interventricular septum • All six left ventricular segments at the mid level: mid (anteroseptal, anterior, lateral, septal, inferior, posterior)

488 Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd K. Kamra et al. Role of TEE in pediatric CHD patients

Table 2 Continued

Reference view Multiplane angle () Probe manipulation Structures visualized

Deep TG long axis (left ventricular 0 Anteflex • Left atrium outflow tract view) (Figure 12, Clip S11) • Mitral valve • Left and right ventricles • Interventricular septum • Left ventricular outflow tract • Aortic valve • Proximal ascending aorta Deep TG long axis (right 0 Anteflex and rotate probe to the right • Right ventricular outflow tract ventricular outflow tract view) • Pulmonary valve Start at deep TG long axis (left ventricular • Proximal main pulmonary artery outflow view) (Figure 13, Clip S12) • Superior vena cava Start at deep TG long axis 90–120 Anteflex and rotate of probe to the left • Left ventricular outflow tract (complementary cross-section to • Aortic valve deep TG long axis view at 0) • Outlet ventricular septum • Pulmonary veins • Atrioventricular valves

Asc, ascending; Ao, aorta; Ch, chamber; ME, mid-esophageal; TG, transgastric.

Table 3 Intraoperative evaluation of congenital heart defects

Congenital anomaly Presurgical echocardiographic evaluation Postsurgical echocardiographic evaluation

Atrial septal defect • Size • Residual shunting (Figure 14) • Position • Valve competence • Magnitude and direction of shunt • Interrogation of pulmonary and systemic veins to • Associated anomalies (e.g. PAPVR, cleft mitral exclude obstruction valve, left SVC) • Ventricular systolic function (applies to all defects) • Mitral valve competence • Monitoring of ventricular preload, air (applies to all • Right atrial and right ventricular size defects) • Baseline ventricular systolic function (applies to all defects) Ventricular septal • Size, position and number of defects • Evaluation or residual shunt(s) (2D, Doppler, and con- defect (Figures 8 • Magnitude and direction of shunt trast and 15, Clip S13) • AV valve competence echocardiography) • Associated anomalies (e.g. tricuspid and aortic • AV and semilunar valve competence valve involvement) • Estimation of systolic pulmonary artery pressure • Chamber enlargement (TR/PR jets) • Estimation of systolic pulmonary artery pressure (TR/PR jets) Atrioventricular • Morphology and distribution of atrioventricular valve • Adequacy of the surgical repair (residual shunts, atrio septal defect tissue across ventricles ventricular valve regurgitation or stenosis) (Figure 16, Clip S14) • Size and location of septal defects, and shunting • Estimation of systolic pulmonary artery pressure (magnitude/direction) (TR/PR jets) • Left AV valve cleft • Left ventricular outflow tract (exclude obstruction) • Regurgitation/stenosis across common atrioventricular valve • Patency of outflow tracts • Ventricular dimensions (balanced versus unbalanced) Obstructive lesions • Level (valvar, subvalvar, supravalvar) • Residual obstruction involving the LVOT • Cause and severity of the obstruction • Valvar regurgitation (aortic and mitral) (Figures 17 and 18, • Morphology of LVOT/aortic valve/ascending aorta and • Iatrogenic pathology resulting from the repair Clips S15 and S16) measurements of involved/related structures (e.g., ventricular septal defect and mitral • Left ventricular size and wall thickness regurgitation) • Aortic valve competence • Associated defects

Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd 489 Role of TEE in pediatric CHD patients K. Kamra et al.

Table 3 Continued

Congenital anomaly Presurgical echocardiographic evaluation Postsurgical echocardiographic evaluation

Obstructive lesions • Evaluation of right ventricular outflow • Residual obstruction involving the RVOT tract for level and severity obstruction • Valvar regurgitation (pulmonary and tricuspid) (Figures 19 and 20, • Size of main and branch pulmonary arteries • Estimation of right ventricular/pulmonary artery Clip S17) • Evaluation of pulmonary and tricuspid valves for systolic pressure morphology, stenosis/regurgitation, and size of valve annuli • Right ventricular size and wall thickness • Associated defects Transposition of the • Confirmation of pathology by assessing segmental • Residual defects (e.g., shunts, in atrial baffle great arteries anatomy operations evaluation of baffle leak and obstruction (Figure 21) • Morphology, size and function of semilunar valves of venous pathways) • AV valve competence • Evaluation of both semilunar valves for • Outflow tract patency regurgitation and stenosis • Origin and course of coronary arteries • Outflow tract patency • Presence of associated cardiac anomalies • Biventricular function (e.g. septal defects) • Left ventricular segmental wall motion abnormalities Truncus arteriosus • Truncal valve morphology, stenosis, regurgitation • Truncal valvar regurgitation/stenosis • Anatomy of the main and proximal • Right ventricle to pulmonary artery reconstruction branched pulmonary arteries for stenosis or regurgitation • Presence, position, size, and shunting • Residual defects across intracardiac defects • Estimate of RV pressure • Atrioventricular valve competence • Coronary arteries Ebstein’s anomaly • Morphology and displacement of • Tricuspid valve function (regurgitation/stenosis) (Figure 22, Clip S18) tricuspid valve leaflets • Residual intracardiac shunting • Severity of tricuspid regurgitation • Residual RVOT obstruction • Right atrial size • Right ventricular systolic function • Right ventricular size and function • RVOT/pulmonary valve patency • Associated AV valve pathology • Atrial level shunts Single ventricle • Segmental anatomy Post-Norwood Procedure: (Figure 23, Clip S19) • Ventricular morphology • Adequacy of interatrial communication • Relationships between AV valves and ventricles. • AV and semilunar valve competence • Morphology and function of AV valves • Patency of systemic outflow • Presence and adequacy of interatrial communication • Residual pathology if concomitant defect(s) addressed • Outflow tract patency Post-Glenn Operation: • Associated defects • Patency of cavopulmonary connection • Antegrade flow across pulmonary valve (if pulsatile Glenn) Post-Fontan/total cavopulmonary connection: • Inferior vena cava to pulmonary artery pathway to exclude obstruction • Assess for leaks if lateral tunnel Fontan completion • Flow across fenestration • AV valve competence Anomalous • Site of drainage and presence of obstruction across • Exclude pulmonary venous obstruction pulmonary venous pulmonary venous pathways • Residual intracardiac defect return • Atrial septum • Estimation of systolic right ventricular/pulmonary • Associated cardiac defect(s) artery pressure • Estimation of systolic right ventricular/pulmonary artery pressure

2D, two-dimensional; AV, atrioventricular; LVOT, left ventricular outflow tract; PAPVR, partial anomalous pulmonary venous return; PR, pul- monary regurgitation; RVOT, right ventricular outflow tract; SVC, superior vena cava; TR, tricuspid regurgitation.

490 Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd K. Kamra et al. Role of TEE in pediatric CHD patients

adult patients with CHD already document the value of live 3D TEE imaging over conventional 2D scanning.

Supporting information Additional Supporting Information may be found in the online version of this article: Clip S1 Transcatheter device closure of secundum atrial septal defect. Clip S2 Mitral stenosis. Clip S3 Coronary sinus. Figure 24 Micromultiplane transesophageal echocardiographic pro- Clip S4 Right pulmonary veins, color flow Doppler. be. The dimensions of the imaging probe are shown. Reprinted Clip S5 Mid-esophageal aortic valve short axis view. from Ref. (9), Copyright (June 2010), with permission from Clip S6 Mid-esophageal right ventricular inflow-out- Elsevier. flow view. Clip S7 Mid-esophageal bicaval view. Clip S8 Mid-esophageal two-chamber view. Clip S9 Mid-esophageal aortic valve long axis view. Micromultiplane TEE probe Clip S10 Transgastric mid short axis view. A miniaturized probe, referred to as a micromultiplane Clip S11 Deep transgastric long axis view of the left TEE device, has been recently developed to potentially ventricular outflow tract. increase the margin of safety of this imaging approach Clip S12 Deep transgastric long axis view of the in neonates and the smallest infants (Figure 24). Initial right ventricular outflow tract. experience have demonstrated that this probe can Clip S13 Tetralogy of Fallot. obtain good quality images in children weighing Clip S14 Complete atrioventricular canal defect <3.5 kg without causing hemodynamic or ventilatory (Atrioventricular septal defect). compromise (9,11,13,62,63). Clip S15 Subaortic membrane. Clip S16 Supravalvar aortic stenosis. Clip S17 Tetralogy of Fallot. Three-dimensional transesophageal technology Clip S18 Ebstein’s anomaly of the tricuspid valve. For patients over 30 kg in weight, the currently avail- Clip S19 Hypoplastic left heart syndrome. able 3D TEE matrix array probe can be used to Data S1 Full legend for video clips. enhance the morphologic, structural, and functional Please note: Wiley-Blackwell are not responsible for assessment. Particularly applications that may benefit the content or functionality of any supporting materi- from 3D imaging include the evaluation of valve pathol- als supplied by the authors. Any queries (other than ogy and percutaneous transcatheter interventions aimed missing material) should be directed to the correspond- at occlusion of septal defects (64–83). Recent data in ing author for the article.

References 1 Bettex DA, Schmidlin D, Bernath MA et al. 4 Kaushal SK, Dagar KS, Singh A et al. during repair of congenital cardiac defects: Intraoperative transesophageal Intraoperative echocardiography as a rou- identification of residual problems necessi- echocardiography in pediatric congenital tine adjunct in assessing repair of congenital tating reoperation. J Am Soc Echocardiogr cardiac surgery: a two-center observational heart defects: experience with 300 cases. 1993; 6: 356–365. study. Anesth Analg 2003; 97: 1275–1282. Ann Card Anaesth 1998; 1: 36–45. 8 Benheim A, Karr SS, Sell JE et al. Routine 2 Randolph GR, Hagler DJ, Connolly HM 5 Loick HM, Scheld HH, Van Aken H. use of transesophageal echocardiography et al. Intraoperative transesophageal echo- Impact of perioperative transesophageal and color flow imaging in the evaluation cardiography during surgery for congenital echocardiography on cardiac surgery. and treatment of children with congenital heart defects. J Thorac Cardiovasc Surg Thorac Cardiovasc Surg 1997; 45: 321–325. heart disease. Echocardiography 1993; 10: 2002; 124: 1176–1182. 6 O’Leary PW, Hagler DJ, Seward JB et al. 583–593. 3 Mochizuki Y, Patel AK, Banerjee A et al. Biplane intraoperative transesophageal 9 Zyblewski SC, Shirali GS, Forbus GA et al. Intraoperative transesophageal echocardio- echocardiography in congenital heart dis- Initial experience with a miniaturized graphy: correlation of echocardiographic ease. Mayo Clin Proc 1995; 70: 317–326. multiplane transesophageal probe in small findings and surgical pathology. Cardiol Rev 7 Stevenson JG, Sorensen GK, Gartman DM infants undergoing cardiac operations. Ann 1999; 7: 270–276. et al. Transesophageal echocardiography Thorac Surg 2010; 89: 1990–1994.

Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd 491 Role of TEE in pediatric CHD patients K. Kamra et al.

10 Bichell DP. Invited commentary. Ann esophageal echocardiography. An updated 36 Bilotta F, Tempe DK, Giovannini F et al. Thorac Surg 2010; 89: 1237–1238. report by the American Society of Anesthe- Perioperative transoesophageal echocardio- 11 Scohy TV, Matte G, van Neer PL et al. A siologists and the Society of Cardiovascular graphy in noncardiac surgery. Ann Card new transesophageal probe for newborns. Anesthesiologists Task Force on Trans- Anaesth 2006; 9: 108–113. Ultrasound Med Biol 2009; 35: 1686–1689. esophageal Echocardiography. Anesthesiol- 37 Patteril M, Swaminathan M. Pro: intrao- 12 Scohy TV, Gommers D, Schepp MN et al. ogy 2010; 112: 1084–1096. perative transesophageal echocardiography Image quality using a micromultiplane 24 Andropoulos DB. Transesophageal echocar- is of utility in patients at high risk of transesophageal echocardiographic probe in diography as a guide to central venous cath- adverse cardiac events undergoing noncardi- older children during cardiac surgery. Eur eter placement in pediatric patients ac surgery. J Cardiothorac Vasc Anesth J Anaesthesiol 2009; 26: 445–447. undergoing cardiac surgery. J Cardiothorac 2004; 18: 107–109. 13 Scohy TV, Gommers D, Jan ten Harkel Vasc Anesth 1999; 13: 320–321. 38 Hofer CK, Zollinger A, Rak M et al. Thera- AD et al. Intraoperative evaluation of mi- 25 Andropoulos DB, Stayer SA, Bent ST et al. peutic impact of intra-operative transo- cromultiplane transesophageal echocardio- A controlled study of transesophageal esophageal echocardiography during graphic probe in surgery for congenital echocardiography to guide central venous noncardiac surgery. Anaesthesia 2004; 59:3– heart disease. Eur J Echocardiogr 2007; 8: catheter placement in congenital heart 9. 241–246. surgery patients. Anesth Analg 1999; 89: 39 Collard CD. Con: intraoperative trans- 14 Ma XJ, Huang GY, Liang XC et al. 65–70. esophageal echocardiography is not of util- Transoesophageal echocardiography in 26 Khalid O, Koenig P. The use of echocardio- ity in patients at high risk of adverse monitoring, guiding, and evaluating surgical graphy in congenital heart surgery and cardiac events undergoing noncardiac sur- repair of congenital cardiac malformations intervention. Expert Rev Cardiovasc Ther gery. J Cardiothorac Vasc Anesth 2004; 18: in children. Cardiol Young 2007; 17: 301– 2006; 4: 263–271. 110–111. 306. 27 Ayres NA, Miller-Hance W, Fyfe DA et al. 40 Maslow A, Bert A, Schwartz C et al. Trans- 15 Balmer C, Barron D, Wright JG et al. Indications and guidelines for performance esophageal echocardiography in the noncar- Experience with intraoperative ultrasound in of transesophageal echocardiography in the diac surgical patient. Int Anesthesiol Clin paediatric cardiac surgery. Cardiol Young patient with pediatric acquired or congenital 2002; 40: 73–132. 2006; 16: 455–462. heart disease: report from the task force of 41 Skiles JA, Griffin BP. Transesophageal 16 Mart CR, Fehr DM, Myers JL et al. Intra- the Pediatric Council of the American Soci- echocardiographic (TEE) evaluation of ven- operative transesophageal echocardiography ety of Echocardiography. J Am Soc Echo- tricular function. Cardiol Clin 2000; 18: in a 1.4-kg infant with complex congenital cardiogr 2005; 18: 91–98. 681–697, vii. heart disease. Pediatr Cardiol 2003; 24: 28 Agricola E, Oppizzi M, Melisurgo G et al. 42 Clark J, Bockoven JR, Lane J et al. Use of 84–85. Transesophageal echocardiography: a com- three-dimensional catheter guidance and 17 Aronson LA. Transnasal placement of plementary view of the heart. Expert Rev trans-esophageal echocardiography to elimi- biplane transesophageal echocardiography Cardiovasc Ther 2004; 2: 61–75. nate fluoroscopy in catheter ablation of left- probe intraoperatively in an adolescent with 29 Yumoto M, Katsuya H. Transesophageal sided accessory pathways. Pacing Clin Elec- congenital heart disease. Anesth Analg 2003; echocardiography for cardiac surgery in trophysiol 2008; 31: 283–289. 97: 1617–1619. children. J Cardiothorac Vasc Anesth 2002; 43 Tucker KJ, Curtis AB, Murphy J et al. 18 Rice MJ, McDonald RW, Li X et al. New 16: 587–591. Transesophageal echocardiographic guid- technology and methodologies for intraoper- 30 Kavanaugh-McHugh A, Tobias JD, Doyle ance of transseptal left heart catheterization ative, perioperative, and intraprocedural T et al. Transesophageal echocardiography during radiofrequency ablation of left-sided monitoring of surgical and catheter inter- in pediatric congenital heart disease. Cardiol accessory pathways in humans. Pacing Clin ventions for congenital heart disease. Echo- Rev 2000; 8: 288–306. Electrophysiol 1996; 19: 272–281. cardiography 2002; 19: 725–734. 31 Subramaniam B, Park KW. Impact of TEE 44 Hahn K, Gal R, Sarnoski J et al. Trans- 19 Bruce CJ, O’Leary P, Hagler DJ et al. Min- in noncardiac surgery. Int Anesthesiol Clin esophageal echocardiographically guided iaturized transesophageal echocardiography 2008; 46: 121–136. atrial transseptal catheterization in patients in newborn infants. J Am Soc Echocardiogr 32 Mahmood F, Christie A, Matyal R. Trans- with normal-sized atria: incidence of compli- 2002; 15: 791–797. esophageal echocardiography and noncardi- cations. Clin Cardiol 1995; 18: 217–220. 20 Shiota T, Lewandowski R, Piel JE et al. Mi- ac surgery. Semin Cardiothorac Vasc Anesth 45 Chumnanvej S, Wood MJ, MacGillivray TE cromultiplane transesophageal echocardio- 2008; 12: 265–289. et al. Perioperative echocardiographic exam- graphic probe for intraoperative study of 33 Catena E, Mele D. Role of intraoperative ination for ventricular assist device implan- congenital heart disease repair in neonates, transesophageal echocardiography in tation. Anesth Analg 2007; 105: 583–601. infants, children, and adults. Am J Cardiol patients undergoing noncardiac surgery. 46 Marcus B, Atkinson JB, Wong PC et al. 1999; 83: 292–295, A7. J Cardiovasc Med (Hagerstown) 2008; 9: Successful use of transesophageal echocardi- 21 Rigby ML. Transoesophageal echocardiog- 993–1003. ography during extracorporeal membrane raphy during interventional cardiac cathe- 34 Schulmeyer MC, Santelices E, Vega R et al. oxygenation in infants after cardiac opera- terisation in congenital heart disease. Heart Impact of intraoperative transesophageal tions. J Thorac Cardiovasc Surg 1995; 109: 2001; 86(Suppl 2): II23–II29. echocardiography during noncardiac sur- 846–848. 22 Van Der Velde ME, Perry SB. Transesoph- gery. J Cardiothorac Vasc Anesth 2006; 20: 47 Noritomi DT, Vieira ML, Mohovic T et al. ageal echocardiography during intervention- 768–771. Echocardiography for hemodynamic evalua- al catheterization in congenital heart 35 Memtsoudis SG, Rosenberger P, Loffler M tion in the intensive care unit. Shock 2010; disease. Echocardiography 1997; 14: 513– et al. The usefulness of transesophageal 34(Suppl 1): 59–62. 528. echocardiography during intraoperative 48 Guarracino F, Baldassarri R. Transesopha- 23 Thys DM, Brooker RF, Cahalan MK et al.. cardiac arrest in noncardiac surgery. Anesth geal echocardiography in the OR and ICU. Practice guidelines for perioperative trans- Analg 2006; 102: 1653–1657. Minerva Anestesiol 2009; 75: 518–529.

492 Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd K. Kamra et al. Role of TEE in pediatric CHD patients

49 Salem R, Vallee F, Rusca M et al. Hemo- heart disease in the adult. Cardiol Clin 1993; before and after transcatheter closure. dynamic monitoring by echocardiography in 11: 505–520. Circulation 2009; 120: e92–e93. the ICU: the role of the new echo 62 Wolfe LT, Myers JL. Invited commentary. 74 Lerakis S, Babaliaros V, Junahadhwalla Z techniques. Curr Opin Crit Care 2008; 14: Ann Thorac Surg 2010; 89: 1994. et al. Three-dimensional transesophageal 561–568. 63 Cannesson M, Henaine R, Metton O et al. echocardiographic guidance during retrieval 50 Andropoulos DB, Stayer SA, Bent ST et al. Images in cardiovascular medicine. Intra- of an embolized percutaneous atrial septal The effects of transesophageal echocardiog- operative transesophageal echocardiography defect closure device. Echocardiography raphy on hemodynamic variables in small using a miniaturized transducer in a neonate 2009; 26: 970–972. infants undergoing cardiac surgery. undergoing Norwood procedure for hypo- 75 Enar S, Singh P, Douglas C et al. Live/real J Cardiothorac Vasc Anesth 2000; 14: plastic left heart syndrome. Circulation time three-dimensional transthoracic echo- 133–135. 2008; 117: 702–703. cardiographic assessment of transposition of 51 Cote G, Denault A. Transesophageal echo- 64 Vegas A, Meineri M. Core review: three- the great arteries in the adult. Echo- cardiography-related complications. Can J dimensional transesophageal echocardio- cardiography 2009; 26: 1095–1104. Anaesth 2008; 55: 622–647. graphy is a major advance for intraoperative 76 Balzer J, Kelm M, Kuhl HP. Real-time 52 Kostolny M, Schreiber C, Henze R et al. clinical management of patients undergoing three-dimensional transoesophageal echo- Temporary pulmonary vein stenosis during cardiac surgery: a core review. Anesth Analg cardiography for guidance of non-coro- intraoperative transesophageal echocardiog- 2010; 110: 1548–1573. nary interventions in the catheter raphy in total cavopulmonary connection. 65 van Noord PT, Scohy TV, McGhie J et al. laboratory. Eur J Echocardiogr 2009; 10: Pediatr Cardiol 2006; 27: 134–136. Three-dimensional transesophageal echo- 341–349. 53 Stevenson JG. Incidence of complications in cardiography in Ebstein’s anomaly. Interact 77 Baker GH, Shirali G, Ringewald JM et al. pediatric transesophageal echocardiography: Cardiovasc Thorac Surg 2010; 10: 836–837. Usefulness of live three-dimensional trans- experience in 1650 cases. J Am Soc Echocar- 66 Rinaldi JP, Latcu DG, Saoudi N. Real-time esophageal echocardiography in a congenital diogr 1999; 12: 527–532. 3-dimensional transesophageal echocardio- heart disease center. Am J Cardiol 2009; 54 Bensky AS, O’Brien JJ, Hammon JW. graphy for the diagnosis of a thrombus 103: 1025–1028. Transesophageal echo probe compression of straddling the patent foramen ovale. JAm 78 Aggarwal G, Schlosshan D, Arronis C et al. an aberrant right subclavian artery. JAm Coll Cardiol 2010; 55: e7. Images in cardiovascular medicine. Real- Soc Echocardiogr 1995; 8: 964–966. 67 Price MJ, Smith MR, Rubenson DS. Utility time 3-dimensional transesophageal echocar- 55 Shanewise JS, Cheung AT, Aronson S et al. of on-line three-dimensional transesophageal diography in the evaluation of a patient ASE/SCA guidelines for performing a com- echocardiography during percutaneous atrial with concomitant double-orifice mitral prehensive intraoperative multiplane trans- septal defect closure. Catheter Cardiovasc valve, bicuspid aortic valve, and coarctation esophageal echocardiography examination: Interv 2010; 75: 570–577. of the aorta. Circulation 2009; 120: e277– recommendations of the American Society 68 Kwak J, Andrawes M, Garvin S et al. 3D e279. of Echocardiography Council for Intraoper- transesophageal echocardiography: a review 79 Salgo IS. 3D echocardiographic visualiza- ative Echocardiography and the Society of of recent literature 2007–2009. Curr Opin tion for intracardiac beating heart surgery Cardiovascular Anesthesiologists Task Anaesthesiol 2010; 23: 80–88. and intervention. Semin Thorac Cardiovasc Force for Certification in Perioperative 69 Bockeria L, Plakhova V. eComment: re: Surg 2007; 19: 325–329. Transesophageal Echocardiography. Anesth three-dimensional transesophageal 80 Mercer-Rosa L, Seliem MA, Fedec A et al. Analg 1999; 89: 870–884. echocardiography in Ebstein’s anomaly. Illustration of the additional value of real- 56 Sundar S, DiNardo JA. Transesophageal Interact Cardiovasc Thorac Surg 2010; 10: time 3-dimensional echocardiography to echocardiography in pediatric surgery. Int 837–838. conventional transthoracic and transesopha- Anesthesiol Clin 2008; 46: 137–155. 70 Azran MS, Romig CB, Locke A et al. Echo geal 2-dimensional echocardiography in 57 Smallhorn JF. Intraoperative transesopha- rounds: application of real-time 3-dimen- imaging muscular ventricular septal defects: geal echocardiography in congenital heart sional transesophageal echocardiography in does this have any impact on individual disease. Echocardiography 2002; 19: 709– the percutaneous closure of a mitral para- patient treatment? J Am Soc Echocardiogr 723. valvular leak. Anesth Analg 2010; 110: 2006; 19: 1511–1519. 58 Miller-Hance WC, Silverman NH. Trans- 1581–1583. 81 Baweja G, Nanda NC, Kirklin JK. Defini- esophageal echocardiography (TEE) in con- 71 Taniguchi M, Akagi T, Watanabe N et al. tive diagnosis of cor triatriatum with genital heart disease with focus on the Application of real-time three-dimensional common atrium by three-dimensional adult. Cardiol Clin 2000; 18: 861–892. transesophageal echocardiography using a transesophageal echocardiography in 59 Russell IM, Silverman NH, Miller-Hance W matrix array probe for transcatheter closure an adult. Echocardiography 2004; 21: et al. Intraoperative transesophageal echo- of atrial septal defect. J Am Soc Echocardi- 303–306. cardiography for infants and children under- ogr 2009; 22: 1114–1120. 82 Miller AP, Nanda NC, Aaluri S et al. going congenital heart surgery: the role of 72 Piatkowski R, Budaj-Fidecka A, Scislo P Three-dimensional transesophageal echo- the anesthesiologist. J Am Soc Echocardiogr et al. Transesophageal real time three- cardiographic demonstration of anatomical 1999; 12: 1009–1014. dimensional echocardiography in assess- defects in AV septal defect patients present- 60 Muhiudeen Russell IA, Miller-Hance WC, ment of partial atrioventricular septal ing for reoperation. Echocardiography 2003; Silverman NH. Intraoperative transesopha- defect. Echocardiography 2009; 26: 1092– 20: 105–109. geal echocardiography for pediatric patients 1094. 83 Ahmed S, Nekkanti R, Nanda NC et al. with congenital heart disease. Anesth Analg 73 Marek T, Zelizko M, Kautzner J. Images in Three-dimensional transesophageal echo- 1998; 87: 1058–1076. cardiovascular medicine. Real-time 3-dimen- cardiographic demonstration of intraatrial 61 Marelli AJ, Child JS, Perloff JK. Trans- sional transesophageal echocardiography baffle obstruction. Echocardiography 2003; esophageal echocardiography in congenital imaging: adult patent ductus arteriosus 20: 683–686.

Pediatric Anesthesia 21 (2011) 479–493 ª 2011 Blackwell Publishing Ltd 493 Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Anesthesia during surgical repair for congenital heart disease and the developing brain: neurotoxic or neuroprotective? Lisa Wise-Faberowski1 & Andreas Loepke2

1 Department of Anesthesiology, Stanford University, Palo Alto, CA, USA 2 University of Cincinnati, Cincinnati, OH, USA

Keywords Introduction anesthesia; brain; congenital heart disease; neurotoxic; neuroprotective Every year, approximately 10 000 children require anesthesia for congenital heart surgery during their first year of life (1,2). Advances in surgical tech- Correspondence niques, cardiopulmonary bypass, anesthesia, and intensive care manage- Lisa Wise-Faberowski, ment have dramatically improved long-term survival in these children and Assistant Professor of Anesthesiology, have therefore focused attention on quality of life and neurocognitive out- Stanford University Medical Center, Lucile comes (3–6). As the numbers of survivors increase and surgical procedures Packard Children’s Hospital, Palo Alto, CA 94305, USA become more complex, there is a growing concern about the risk of abnor- Email: [email protected] mal neurologic outcome. The neurodevelopmental outcome of these chil- dren is quite variable: many are totally normal, while 20–50% of these Section Editor: Chandra Ramamoorthy children have neurologic impairment and developmental disabilities (7,8). The etiology of neurologic impairment in these children has been studied Accepted 18 February 2011 and is multifactorial: brain immaturity (9–11), timing of surgery in cyanotic infants (12), prolonged low brain regional oxygen saturation (13), prolonged doi:10.1111/j.1460-9592.2011.03586.x

deep hypothermic circulatory arrest (14), and chromo- ease. Each of these factors will be discussed indepen- somal anomalies (15). However, the effect of anesthetic dently. exposure and neurologic outcome in these children is 1. Chronic hypoxia: congenital heart disease, requir- just beginning to be addressed (16,17). The potential ing surgical repair in infancy, is one usually detrimental effect of prolonged anesthetic exposure on causing hypoxia, such as transposition of the great the developing brain has recently become a public con- vessels, truncus arteriosus, total anomalous pulmo- cern. Despite animal evidence supporting this concern nary venous return, or hypoplastic left heart syn- (18–20), the clinical evidence in humans is limited to drome. retrospective evaluations (21–23). More importantly, 2. Cardiopulmonary bypass (ischemia/reperfusion the effect of anesthetics on neurodevelopmental out- injury): cardiopulmonary bypass is usually low- come in infants after repair for congenital heart disease flow cardiopulmonary bypass; however, circulatory needs to be further evaluated. The question to be arrest is still used for some procedures at some answered, ‘Does the type of anesthetic used during institutions. More recent is the use of antegrade surgical repair of congenital heart disease affect neuro- cerebral perfusion in some institutions. Regardless logic outcome?’ is complex. Thus, the aim of this article of the method of CPB, normal pulsatile blood flow is to discuss the impact of chronic hypoxia, cardio- to the brain is disrupted pulmonary bypass, and anesthetics on the developing 3. Prolonged anesthetic exposure: surgical repair for brain. complex cyanotic heart lesions such as hypoplastic There are three factors to consider when assessing left heart syndrome, transposition of the great neurologic outcome in infants undergoing a prolonged arteries, etc. requires a 4–6 h general anesthetic at anesthetic for surgical repair of congenital heart dis- most institutions.

554 Pediatric Anesthesia 21 (2011) 554–559 ª 2011 Blackwell Publishing Ltd L. Wise-Faberowski and A. Loepke Anesthesia and congenital heart disease

comes studies is currently being addressed (36). Chronic hypoxia During circulatory arrest and possibly during low- The developing hypoxic brain may or may not be dif- flow cardiopulmonary bypass, cells in certain brain ferent than the developing normoxic brain. For exam- regions are known to die, whereas, other cells in the ple, using a perinatal rat model mimicking cyanotic same region or different regions do not. This phe- heart disease, in which rat pups were maintained in a nomenon, referred to as selective vulnerability, occurs hypoxic environment (10% oxygen) for 1 and 4 weeks, in mature and immature brain (37–39). In neonates, there was no difference in neuronal cell death in the neurons and, more importantly, oligodendrocytes in chronic hypoxia group as compared to those animals the neocortex and hippocampus are selectively vulner- maintained in a normoxic environment for the same able to death after deep hypothermic circulatory duration (24). Thus, a protective adaptive response arrest. These events are felt to occur early, within may exist in the neonatal brain when exposed to hours of reperfusion and continue for several days chronic hypoxia (25,26). postoperatively. The process of reperfusion during This adaptive response to chronic hypoxia has been cardiopulmonary bypass, is that in which areas of the attributed to the N-methyl-D-aspartate (NMDA) 2B brain, formally deprived of oxygen, are now abun- receptor subunit composition (27,28). The NR2B pre- dant in oxygen. dominant subunit composition of the NMDA receptor Overabundance of oxygen in the acutely hypoxic early in life allows the fetus to tolerate hypoxic condi- infant is a concern, and the use of hyperoxia, 100% tions in utero by avoiding excessive calcium influx via oxygen, in the resuscitation of acutely asphyxiated NMDA receptors (29–31) and thereby allowing the infants is now being questioned (40–43). Evidence in expression of brain-derived neurotrophic factor, which neonatal human and animal models of acute hypoxia prevents apoptotic cell death (32,33). The NMDA has determined that exposure to hyperoxia during and receptor subunit composition switches to NMDA 2A after a hypoxic period can generate excessive neuro- with increasing developmental age, in response to toxic compounds. Markers of increased oxidative stress increased PaO2 after birth (30,31). Anesthetic agents, have been found in neonates as long as 28 days after particularly, volatile agents act on the NMDA 2B resuscitation with 100% oxygen for acute hypoxia receptor subunit composition (34). But how anesthetic (44). Thus, the use of hyperoxia during cardiopulmo- agents affect the adaptive response to chronic hypoxia nary bypass, a period of increased oxidative stress and is not known. inflammatory mediators (45), in normoxic neonates Another adaptation to chronic hypoxia is cortical may not be as beneficial as previously thought. Fur- neurogenesis (35). It has been shown that 30% of cor- thermore, there are no studies that have evaluated the tical neurons are lost after chronic hypoxia and results effects of cardiopulmonary bypass, with associated in a reduction in brain volume. However, with the ischemia followed by hyperoxia, in a chronically hyp- establishment of normoxia, in previously hypoxic oxic human or animal model in the presence or brain, there is a 40% increase in cortical neurons but absence of anesthesia. no increase in astrocyte or oligodendrocyte number. In normoxic brain, made acutely hypoxic, the re-estab- Prolonged anesthetic exposure lishment of normoxia increases the number of oligo- dendrocytes and astrocytes, but not neurons. Thus, By nature of the surgical procedures being performed with the establishment of normoxia in the developing in these infants, anesthesia is required and the duration chronic hypoxic brain, there is a loss of supporting of exposure is several hours. A high-dose narcotic structures, which may promote white matter injury. technique, as that used in the Boston Circulatory Furthermore, as to be discussed later, narcotic agents Arrest Trials, was a contribution of Anand and Hickey can influence neurogenesis. (46). Their clinical investigation in children undergoing anesthesia for surgical repair of congenital heart dis- ease supported an improvement in overall outcome Cardiopulmonary bypass with the use of high-dose narcotics. High-dose narcot- Cardiopulmonary bypass has been a great example of ics reduced the stress response curve as determined by a global ischemic event followed by hyperoxic reper- serum catecholamine and stress markers. The effect of fusion with resultant white matter injury. More stress, catecholamine binding to serotonergic and nico- recently, the use of antegrade cerebral perfusion has tinic receptors, and the amelioration of stress, binding gained favor in some institutions. The impact of ante- of opioids to opioid receptors, can be influenced by grade cerebral perfusion in long-term neurologic out- development.

Pediatric Anesthesia 21 (2011) 554–559 ª 2011 Blackwell Publishing Ltd 555 Anesthesia and congenital heart disease L. Wise-Faberowski and A. Loepke

In general, binding of catecholamines to serotonergic for later inhibition via potassium-chloride transporter, and nicotinic receptors decreases dramatically as the KCC2 (61). Thus, loss of the GABAA a2 composition developing human brain transitions between fetal, or interaction with the GABAA ligand-gated channel infant, and mature brain. Mature serotonergic receptor by GABAA-specific agent such as sevoflurane or mi- binding is about 25% of fetal binding and mature nic- dazolam early in development may promote neuronal otinic binding is about 50% of fetal nicotinic binding cell death. (47), so the ability to mount a stress response decreases Exposure to volatile anesthetics may cause morpho- with increasing developmental age. Binding to opioid logical and functional responses in neural tissue that receptors remains relatively unchanged throughout are dependent on the developmental stage of the tissue brain development, but the opioid receptor subunit (62–65). This is likely attributable to developmental types do change (48,49). differences in GABA and NMDA receptor subunit Opioid receptors, specifically mu and kappa, are composition. As a result, conclusions drawn from adult present early in gestation with delta receptors present- brain regarding responses to anesthetics (66–68) may ing later in gestation (50,51). The pattern of opioid be very different from responses of developing neural receptor expression is cell type specific, and the action tissue in neonates and infants. of a specific opioid receptor may be different based on For example, Jevtovic-Todorovic et al. (18) exposed the cell type: astrocytes, oligodendrocytes, and normal PND 7 rat pups to combinations of isoflurane, neurons. Fetal expression of the mu receptor is midazolam, and N2O. Isoflurane caused neuronal significantly decreased in the mature brain. The mu apoptosis at the three concentrations studied; 1.5% receptor is involved in regulation of neuronal and glial isoflurane had the greatest effect on neuronal viability. proliferation. Exposure of PND 7 pups to the combination of 1.5%

Opioid action on mu receptors in the developing isoflurane, midazolam, and N2O caused neuronal brain inhibits neuronal and glial cell growth and divi- apoptosis, necrosis, and learning and memory deficits sion (52,53). During development, opioid action on measured months after anesthetic exposure. The effect kappa receptors promotes cell proliferation and offers in pups of other developmental stages was not evalu- neuroprotection. Thus, there is a developmental bal- ated. ance between mu and kappa receptors. Because of the Although this work has been criticized for lack influence of all three receptor subtypes (l, j, and d)on of physiologic control during anesthetic exposure, it astrocytes and oligodendrocytes, myelination is inhib- is consistent with other findings. NMDA antagonists ited with the use of narcotics early in development (MK-801) initiate apoptosis in immature rat brain, (54). Concerns about preexisting white matter injury in the extent of which is dependent on both duration of infants prior to cardiopulmonary bypass, which may exposure (4–6 h) and age (18,19). The effect on cell be exacerbated by chronic hypoxia and cardiopulmo- death increases as a function of postnatal age with nary bypass, (55–57) may further question the use of maximal effect at PND7. However, it has been shown narcotics before, during, and after these corrective sur- in preliminary work, from the same researchers, gical procedures. in animal models, that hypothermia, as used in As discussed previously with NMDA and opioid cardiopulmonary bypass, can prevent the neurotoxic receptors, GABAA receptor subunit composition effects of anesthetics (69). changes during development as well (58,59). The tran- Contrary to the above work, volatile anesthetics sition in NMDA and GABAA receptor subunit com- have demonstrated neuroprotective effects in adult rat position occurs around postnatal seven in a rat pup. A models of global and focal ischemia (70,71). Desflu- similar transition in the GABAA receptor subunit rane, a volatile anesthetic with properties similar to composition occurs in humans as well (60). GABAA isoflurane in terms of cerebral autoregulation, confers receptors are pentameric structures containing predom- neocortical and hippocampal neuroprotection in a pig- inantly a1ora2 subunits in combination with b2orb3 let model of deep hypothermic circulatory arrest or and a c2 subunit. The GABAA a and b subunits are low-flow cardiopulmonary bypass as opposed to a nar- crucial for the action of volatile anesthetics. The cotic-based technique. This was confirmed with acute

GABAA receptor subunit composition transitions from histologic examination and long-term neurologic evalu- being predominantly a2, early in development, to pre- ation. Neither the particular relevance of desflurane as dominantly a1 later in development. The GABAA a2 compared to other volatile anesthetics nor has the allows for early excitation via sodium–potassium-chlo- effect of chronic hypoxia as opposed to normoxia been ride transporter, NKCC1, which promotes synaptic evaluated using a cardiopulmonary bypass model of growth and development, whereas GABAA a1 allows developing brain.

556 Pediatric Anesthesia 21 (2011) 554–559 ª 2011 Blackwell Publishing Ltd L. Wise-Faberowski and A. Loepke Anesthesia and congenital heart disease

In summary, there are limited in vivo, human or ani- monary bypass, in the developing hypoxic brain may mal, investigations comparing the neuroprotective or need to be considered. Protective mechanisms devel- neurotoxic effects of anesthetic agents in the develop- oped in utero to tolerate chronic hypoxic conditions ing cyanotic brain undergoing cardiopulmonary bypass exist in the developing brain. Disruption of these in the absence or presence of surgical stimulus, let mechanisms by anesthetic agents and high concentra- alone surgical repair for congenital heart disease. Con- tions of oxygen and the subsequent effect on neurode- clusions drawn about anesthetic neurotoxicity in the velopment is not known. Thus, at present, no developing normoxic brain could be potentially differ- conclusions can be drawn as to which is the best anes- ent than those drawn in the developing hypoxic brain. thetic agent or agents to use in these children. Furthermore, the effects of hyperoxia, as in cardiopul-

References 1 Lloyd-Jones D. Heart disease and stroke brain injury before and after neonatal an evaluation in organotypic hippocampal statistics – a 2010 update: a report from the cardiac surgery with high-flow bypass slice cultures. Anesth Analg 2005; 101: 651– American Heart Association. Circulation and cerebral oxygen monitoring. 657. 2010; 121: e46–e121. J Thorac Cardiovasc Surg 2010; 139: 20 Loepke AW, Istaphanous GK, McAuliffe JJ 2 Welke KF, Shen I, Ungerleider RM. Cur- 543–556. et al. The effects of neonatal isoflurane rent assessment of mortality rates in congen- 12 Petit C, Rome JJ, Wernovsky G et al. exposure in mice on brain cell viability, ital cardiac surgery. Ann Thorac Surg 2006; Preoperative brain injury in transposition of adult behavior, learning and memory. 82: 164–170. the great arteries is associated with Anesth Analg 2009; 108: 90–104. 3 Hovels-Gurich HH, Bauer SB, Schnitker R oxygenation and time to surgery, not 21 Kalkman C, Sun LS, Kakavouli A et al. A et al. Long-term outcome of speech and lan- balloon . Circulation 2009; retrospective cohort study of the association guage in children after corrective surgery for 119: 709–716. of anesthesia and hernia repair surgery with cyanotic or acyanotic cardiac defects in 13 Kussman BD, Wypij D, Laussen P et al. behavioral and developmental disorders in infancy. Eur J Paediatr Neurol 2008; 12: Relationship of intraoperative cerebral young children. J Neurosurg Anesthesiol 378–386. oxygen saturation to neurodevelopmental 2009; 21: 283–285. 4 Bellinger DC, Jonas RA, Rappaport LA et al. outcome and brain magnetic resonance 22 Wilder RT, Flick RP, Sprung J et al. Rela- Developmental and neurologic status of chil- imaging at 1 year of age in infants under- tionship between exposure to anesthesia and dren after heart surgery with hypothermic cir- going biventricular repair. Circulation 2010; subsequent learning disabilities in children. culatory arrest or low-flow cardiopulmonary 122: 245–254. Anesthesiology 2009; 110: 796–804. bypass. N Engl J Med 1995; 332: 549–555. 14 Wypij D, newburger JW, Rappaport LA 23 DiMaggio C, Sun LS, Kakavouli A et al. A 5 du Plessis AJ. Neurologic complications of et al. The effect of duration of deep retrospective cohort study of the association cardiac disease in the newborn. Clin Perina- hypothermic circulatory arrest in infant of anesthesia and hernia repair surgery with tol 1997; 24: 807–826. heart surgery on late neurodevelopment: the behavioral and developmental disorders in 6 Ballweg JA, Wernovsky G, Gaynor JW. Boston circulatory arrest trial. J Thorac young children. J Anesthesiol 2009; 21: 283– Neurodevelopmental outcomes following Cardiovasc Surg 2003; 126: 1397–1403. 285. congenital heart surgery. Pediatr Cardiol 15 Atallah J, Joffe AR, Robertson CMT et al. 24 Bitar FF, el Sabban M, Bitar H et al. Lack 2007; 28: 126–133. Two year general and neurodevelopmental of apoptosis in the hypoxic brain of a rat 7 Majnemer A, Limperopoulas C, Shevell MI outcome after neonatal complex surgery in model mimicking cyanotic heart disease. et al. Long-term neuromotor outcome at patients with deletion 22q11.2: a compara- Brain Inj 2002; 16: 891–900. school entry of infants with congenital heart tive study. J Thorac Cardiovasc Surg 2007; 25 Chavez JC, Agani F, Pichiule P et al. disease requiring open heart surgery. J Pedi- 134: 772–779. Expression of hypoxia-inducible factor- atr 2006; 148: 72–77. 16 Andropoulos D, Stayer S, Dickerson H 1alpha in the brain of rats during chronic 8 Massaro AN, El-dib M, Glass P et al. et al. Strategies focused on cerebral hypoxia. J Appl Physiol 2000; 89: 1937– Factors associated with adverse neuro- oxygenation improve neurodevelopment 1942. developmental outcomes in infants with after neonatal cardiac surgery. Anesthesiol- 26 Bruer U, Weih MK, Isav NK. Induction of congenital heart disease. Brain Dev 2008; ogy 2010; 113: A199. tolerance in rat cortical neurons: hypoxic 300: 437–446. 17 Andropoulos D, Hunter J, Wilde E et al. preconditioning. FEBS Lett 1997; 414: 117– 9 Limperopoulos C, Tworetsky W, McElhin- Effect of isoflurane and midazolam on 121. ney DB et al. Brain volume and metabolism quantitative MRI brain maturation in car- 27 Bickler PE, Fahlman CS, Taylor DM. Oxy- in fetuses with congenital heart disease: diac surgical neonates. Anesthesiology 2010; gen sensitivity of NMDA receptors: rela- evaluation with magnetic resonance imaging 113: A201. tionship to NR2 subunit composition and and spectroscopy. Circulation 2010; 121: 18 Jevtovic-Todorovic V, Hartman RE, Izumi hypoxia tolerance of neonatal neurons. Neu- 26–33. Y et al. Early exposure to common anes- roscience 2003; 118: 25–35. 10 Miller SP, McQuillen P, Hamrick S et al. thetics causes widespread neurodegeneration 28 Bickler PE, Hansen BM. Hypoxia-tolerant Abnormal brain development in newborns in the developing rat brain and persistent neonatal CA1 neurons: relationship to sur- with congenital heart disease. N Engl J Med learning deficits. J Neurosci 2003; 23: 876– vival to evoked glutamate release and gluta- 2007; 357: 1928–1938. 882. mate receptor mediated calcium changes in 11 Andropoulos D, Hunter JV, Nelson DP 19 Wise-Faberowski L, Zhang H, Ing R et al. hippocampal slices. Dev Brain Res 1998; et al. Brain immaturity is associated with Isoflurane-induced neuronal degeneration: 106: 57–69.

Pediatric Anesthesia 21 (2011) 554–559 ª 2011 Blackwell Publishing Ltd 557 Anesthesia and congenital heart disease L. Wise-Faberowski and A. Loepke

29 Xia Y, Fung ML, O’Reilly JP et al. 100% oxygen exacerbates neurological dys- long-term cognitive deficits. Neuroscience Increased neuronal excitability after long- function following nine minutes of normo- 2009; 158: 1326–1337. term oxygen deprivation is mediated mainly thermic cardiac arrest in dogs. Resuscitation 54 Sanchez ES, Bigbee JW, Fobbs W et al. by sodium channels. Brain Res Mol Brain 1994; 27: 159–170. Opioid addiction and pregnancy: perinatal Res 2000; 76: 211–219. 43 Pomper JK, Graulich J, Kovacs R et al. exposure to buprenorphine affects myelina- 30 Hestrin S. Developmental regulation of High oxygen tension leads to acute cell tion in the developing brain. Glia 2008; 56: NMDA receptor mediated at a central syn- death in organotypic hippocampal slice cul- 1017–1027. apse. Nature 1992; 357: 686–689. tures. Dev Brain Res 2001; 126: 109–116. 55 Andropoulos DB, Hunter JV, Nelson DP 31 Haberny KA, Paule MG, Scallet AC et al. 44 Kaindl AM, Sifringer M, Zabel C et al. et al. Brain immaturity is associated Ontogeny of the N-methyl-d-aspartate Acute and long-term proteome changes with brain injury before and after (NMDA) receptor system and susceptibility induced by oxidative stress in the developing neonatal cardiac surgery with high-flow to neurodegeneration. Toxicol Sci 2002; 68: brain. Cell Death Differ 2006; 13: 1097– bypass and cerebral oxygenation monitor- 9–17. 1109. ing. J Thorac Cardiovasc Surg 2010; 139: 32 Jiang X, Tian F, Mearow K et al. The ex- 45 Nollert G, Nagashima M, Bucerius J et al. 543–556. citoprotective effect of N-methyl-D-aspar- Oxygenation strategy and neurologic dam- 56 Galli KK, Zimmerman RA, Jarvik GP tate receptors is mediated by a brain-derived age after deep hypothermic circulatory et al.. Periventricular leukomalacia is com- neurotrophic factor autocrine loop in cul- arrest. II. Hypoxic versus free radical injury. mon after neonatal cardiac surgery. J Tho- tured hippocampal neurons. J Neurochem J Thorac Cardiovasc Surg 1999; 117: 1172– rac Cardiovasc Surg 2004; 127:692–704. 2005; 94: 713–722. 1179. Erratum in: J Thorac Cardiovasc Surg. 2004 33 Levine ES, Kolb JE. Brain-derived neuro- 46 Anand KJ, Hickey P. Halothane-morphine Sep; 128(3):498. trophic factor increases activity of NR2B- compared with high dose sufentanil for 57 Miller SP, McQuillen PS, Vigneron DB containing N-methyl-d-aspartate receptors anesthesia and postoperative analgesia in et al. Preoperative brain injury in newborns in excised patches from hippocampal neu- neonatal cardiac surgery. N Engl J Med with transposition of the great arteries. Ann rons. J Neurosci Res 2000; 62: 357–362. 1992; 326: 1–9. Thorac Surg 2004; 77: 1698–1706. 34 Ming Z, Griffith BL, Breese GR et al. 47 Reddy SC, Panigrahy A, White WF et al. 58 Sanchez RM, Jensen FE. Maturational Changes in the effect of isoflurane on N- Developmental changes in neurotransmitter aspects of epilepsy mechanisms and conse- methyl-D-aspartic acid-gated currents in cul- receptor binding in the human periaqueduc- quences for the immature brain. Epilepsia tured cerebral cortical neurons with time in tal gray. J Neuropathol Exp Neurol 1996; 2001; 42: 577–585. culture: evidence for subunit specificity. 55: 409–418. 59 Ben-Ari Y. Basic developmental rules and Anesthesiology 2002; 97: 856–867. 48 Tripathi A, Khurshid N, Kumar P et al. their implications for epilepsy in the imma- 35 Fagel DM, Sibereis J, Ebbitt T et al. Corti- Expression of delta- and mu-opioid recep- ture brain. Epileptic Disord 2006; 8: 91–102. cal neurogenesis is enhanced by chronic tors in the ventricular and subventricular Review. perinatal hypoxia. Exp Neurol 2006; 199: zones of the developing human neocortex. 60 Romijn HJ et al. At what age is the 77–91. Neurosci Res 2008; 61: 257–270. developing cerebral cortex of the rat 36 Fraser CD, Andropoulos D. Principles of 49 Steine-Martin A, Knapp PE, Martin K comparable to that of the full-term newborn antegrade cerebral perfusion during arch et al. Opioid system diversity in developing human baby? Early Hum Dev 1991; 26: reconstruction in newborns/infants. Semin neurons, astroglia and oligodendroglia in 61–67. Thorac Cardiovasc Surg Pediatr Card Surg the subventricular zone and striatum: 61 Fiumelli H, Cannceda L, Poo MM. Modu- Annu 2008; 61–68. impact on gliogenesis in vivo. Glia 2001; 36: lation of GABAergic transmission by activ- 37 Kurth CD, Priestley M, Golden J et al. 78–88. ity via postsynaptic Ca2+-dependent Regional patterns of neuronal death after 50 Wang H, Cuzon VC, Pickel VM. Ultra- regulation of KCC2 function. Neuron 2005; deep hypothermic circulatory arrest in new- structural localization of delta-opioid recep- 48: 773–786. born pigs. J Thorac Cardiovasc Surg 1999; tors in the rat caudate-putamen nucleus 62 Ikonomidou C, Mosinger JL, Salles KS 118: 1068–1077. during postnatal development: relation to et al. Sensitivity of the developing rat brain 38 Bottiger BW, Schmitz B, Wiessner C et al. synaptogenesis. J Comp Neurol 2003; 467: to hypobaric/ischemic damage parallels sen- Neuronal stress response and neuronal cell 343–353. sitivity to N-methyl-D-aspartate neurotoxic- damage after cardiocirculatory arrest in rats. 51 Kivell BM, Day DJ, McDonald FJ et al. ity. J Neurosci 1989; 9: 2809–2818. J Cereb Blood Flow Metab 1998; 18: 1077– Developmental expression of mu and delta 63 Zhan X, Fahlman CS, Bickler PE. Isoflura- 1087. opioid receptors in the rat brainstem: evi- ne neuroprotection in rat hippocampal slices 39 Leist M, Jaattela M. Four deaths and a fun- dence for a postnatal switch in mu isoform decreases with aging. Anesthesiology 2006; eral: from caspases to alternative mecha- expression. Brain Res Dev Brain Res 2004; 104: 995–1003. nisms. Nat Rev Mol Cell Biol 2001; 8: 589– 48: 185–196. 64 Gee CE, Benquet P, Raineteau O et al. 598. 52 Sargeant TJ, Day DJ, Miller JH et al. Acute NMDA receptors and the differential ische- 40 Vento M, Asensi M, Sastre J et al. Oxida- in utero morphine exposure slows G2/M mic vulnerability of hippocampal neurons. tive stress in asphyxiated term infants resus- phase transition in radial glial and basal Eur J Neurosci 2006; 23: 2595–2603. citated with 100% oxygen. J Pediatr 2003; progenitor cells in dorsal telencephalon of 65 Creeley CE, Olney JW. The young: neuroa- 142: 240–246. the E15.5 embryonic mouse. Eur J Neurosci poptosis induced by anesthetics and what to 41 Ramji S, Ahuja S, Thirumpuram S et al. 2008; 28: 1060–1067. do about it. Anesth Analg 2010; 110: 442– Resuscitation of asphyxic newborn infants 53 Lin Cs, Tao PL, Jong YJ et al. Prenatal 448. with room air or 100% oxygen. Pediatr Res morphine alters the synaptic complex of 66 Warner DS. Isoflurane neuroprotection. 1993; 34: 809–812. postsynaptic density 95 with N-methyl-D- Anesthesiology 2000; 92: 1226–1228. 42 Zwemer CF, Whitesall SE, D’Alecy LG. aspartate receptor subunit in hippocampal 67 Zhao P, Zuo Z. Isoflurane preconditioning Cardiopulmonary-cerebral resuscitation with CA1 subregion of rat offspring leading to induces neuroprotection that is inducible

558 Pediatric Anesthesia 21 (2011) 554–559 ª 2011 Blackwell Publishing Ltd L. Wise-Faberowski and A. Loepke Anesthesia and congenital heart disease

nitric oxide synthase-dependent in neonatal 69 Creeley CE, Olney JW. The young: neu- newborn pigs. Anesthesiology 2001; 95: rats. Anesthesiology 2004; 101: 695–703. roapoptosis induced by anesthetics what to 959–964. 68 Engelhard K, Werner C, Reeker W et al. do about it. Anesth Analg 2010; 110: 442– 71 Loepke AW, Priestley MA, Schultz SE et al. Desflurane and isoflurane improve neuro- 448. Desflurane improves neurologic outcome logical outcome after incomplete cerebral 70 Kurth CD, Priestly M, Watzman HM et al. after low-flow cardiopulmonary bypass in ischemia in rats. Br J Anesth 1999; 83: Desflurane confers neurologic protection newborn pigs. Anesthesiology 2002; 97: 415–421. for deep hypothermic circulatory arrest in 1521–1527.

Pediatric Anesthesia 21 (2011) 554–559 ª 2011 Blackwell Publishing Ltd 559 Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Pediatric pacemakers and ICDs: how to optimize perioperative care Manchula Navaratnam1 & Anne Dubin2

1 Department of Anesthesia, Stanford University and Hospitals, Stanford, CA, USA 2 Department of Pediatric Cardiology, Stanford University and Hospitals, Stanford, CA, USA

Keywords Summary pediatric; pacemakers; ICDs; perioperative care An increasing number of pediatric patients with permanent pacemakers and implantable cardioverter defibrillators (ICDs) require cardiac and non- Correspondence cardiac surgery. It is critical that the anesthesiologist caring for these Dr Manchula Navaratnam, patients understands the management of the device and the underlying Department of Anesthesia, Stanford heart disease. Children with these devices are more vulnerable to lead fail- University Medical Center, 300 Pasteur ure and inappropriate shocks compared with the adult population. Pre- Drive, H3587 Stanford, CA 94305, USA Email: [email protected] operative assessment and appropriate reprogramming of the device, in addition to minimizing sources of electromagnetic interference, are key- Section Editor: Chandra Ramamoorthy stones in the perioperative care of these patients. Prior consultation with qualified programmers is recommended to enable timely optimization of Accepted 18 February 2011 the device. Magnets may be used in emergency situations but it is impor- tant to appreciate the limitations of magnet use on different models of doi:10.1111/j.1460-9592.2011.03562.x pacemakers and ICDs. Safe and successful perioperative care is dependent upon a well-organized and coordinated multidisciplinary team approach.

Introduction Pacemaker overview As pacemaker and implantable cardioverter defibrilla- Permanent pacemaker systems consist of an impulse tors (ICDs) become smaller, and surgical complexity generator and leads. The leads may be attached to the increases, more pediatric patients are undergoing pace- endocardium via a transvenous approach or to the epi- maker and ICD therapy. This is reflected in the num- cardium via a surgical approach. Transvenous systems ber of patients with these devices presenting for both are usually implanted via the left subclavian vein with cardiac and noncardiac surgery. Anesthesiologists the lead progressing into the right atrium and/or right caring for these patients need to have a sound under- ventricle. The pulse generator is then implanted in the standing of the patient’s underlying heart disease and left prepectoral region. the intended device function, in addition to a practical The decision for a surgical approach is dependent algorithm for device optimization during the perioper- upon several factors, including age and size of patient ative period. as well as cardiac anatomy. Surgical placement of This article reviews the nomenclature and basic leads on the epimyocardial surface is usually via a function of permanent pacemakers and ICDs and indi- minithoracotomy or subxiphoid approach, and the cations for implantation in children and patients with leads are then tunneled and connected to a pulse gen- congenital heart disease (CHD). A practical approach erator that is commonly implanted subcutaneously in to the perioperative management of these devices is the abdominal area. presented including potential complications and Although the morbidity associated with epicardial adverse events associated with them. Pacemakers and lead placement is low (1), the leads are vulnerable ICDs will be collectively referred to as cardiac rhythm to fracture and dislodgement (2) and are at risk management devices (CRMD). of developing increasing pacing thresholds over

512 Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd M. Navaratnam and A. Dubin Pediatric pacemakers and ICDs time (3,4). Lead survival is superior in transvenous lead. Unipolar sensing produces a large potential dif- systems (2). ference because of the increased interelectrode distance. Regardless of the system used, it is generally In bipolar systems, both the anode and cathode are sit- believed that the greatest risk factor for damage to uated on the lead with only a short distance between implanted leads is age at time of implant and the pres- them. The advantage of the bipolar system is that less ence of CHD (2,5,6). energy is required to produce a smaller potential dif- Children and infants have higher resting and peak ference. Electrical signals from outside the heart are heart rates than adults, which may increase battery uti- less likely to be sensed, and thus, there will be a lization and significantly impact the longevity of pulse reduced susceptibility to electromagnetic interference generators (5). In addition, limits to maximum track- (EMI). Recent years have seen a trend towards using ing rates from some devices may result in a substantial bipolar leads. decrease in exercise performance (7). The location and function of pacemakers are Pacemakers may be single chamber, dual chamber, or described using an internationally recognized code that trichamber (resynchronization pacemakers) depending was developed by The North American Society of Pac- on patient size and indication for implantation. Leads ing and Electrophysiology (NAPSE) and the British are typically placed in the right atrium and/or right ven- Pacing and Electrophysiology group (Table 1) (8). tricle. In resynchronization systems, leads are usually The five-position code shown in Table 1 is often placed in both the right and left ventricle. For endo- shortened to the first three positions, e.g., single cham- cardial systems, the LV lead is usually placed within the ber pacemakers can pace the atrium (AAI) or the ven- coronary sinus but atypical configurations have been tricle (VVI). described for patients with CHD. In patients with single Dual chamber (DDD) mode is the most sophisti- ventricle physiology, a resynchronization system may cated and commonly used mode. Position IV describes signify two epicardial leads on a single ventricle. rate modulation and is incorporated into most modern Single chamber devices will sense intrinsic electrical pacemakers. This function is commonly used in activity from the corresponding chamber within a pre- patients who do not have intact sinus node function, set time limit. This will either inhibit or trigger pacing and it enables the pacemaker to automatically increase of that chamber depending on how the device is the heart rate to meet metabolic demands. Most programmed. patients with permanent pacemakers are programmed Dual chamber devices can sense and pace both in to the DDDR mode. The rate response can involve the atrium and in the ventricle and maintain atrioven- either atrial or ventricular pacing depending on the tricular (AV) synchrony in patients without intact AV mode and underlying conduction abnormality. For node function. AV synchrony is important in optimiz- instance, a patient with complete heart block and a ing cardiac output by maintaining the atrial ‘kick’ single ventricular lead can have intact sinus node func- which increases end diastolic volume resulting in tion but still require rate response in a VVIR mode. increased stroke volume. AV synchrony can also lower The most commonly used sensors are those that detect atrial pressures (and hence congestive symptoms) as it bodily accelerations because of motion (9). Sensors, eliminates ‘cannon’ a waves or atrial contraction which determine minute ventilation via changes in against a closed AV valve. thoracic impedance, are also in use (9). Pacemakers sense intrinsic cardiac depolarization by It is important to be aware of how rate modulation measuring changes in the electrical potential of myo- is accomplished, and as discussed later, it may be nec- cardial cells between the anode and cathode. essary to disable this function preoperatively. In unipolar leads, the pulse generator functions as Multisite pacing (position V) refers to either the the anode with the cathode located at the tip of the presence of more than one lead in a single cardiac

Table 1 Generic pacemaker code NAPSE/ BPEG. Adapted from Ref. (8) Positions/categories I/chamber(s) II/chambers III/response IV/rate V/multisite paced sensed to sensing modulation pacing

O = None O = None O = None O = None O = None A = Atrium A = Atrium T = Triggered R = Rate modulation A = Atrium V = Ventricle V = Ventricle I = Inhibited V = Ventricle D = Dual D = Dual D = Dual D = Dual (A + V) (A + V) (T + I) (A + V)

Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd 513 Pediatric pacemakers and ICDs M. Navaratnam and A. Dubin chamber or biventricular pacing. Biventricular pacing procedure or total cavo-pulmonary connection (Fon- (also called cardiac resynchronization therapy) is an tan) for a single ventricle (12). The clinical significance evolving technique of using simultaneous or near of sinus bradycardia is age dependent, frequently simultaneous pace activation of one or both ventricles asymptomatic, and rarely requires pacing. Current rec- to improve ventricular dysynchrony and cardiac func- ommendations are displayed in Table 2. tion and is discussed later. Biatrial pacing is currently Bradycardia–tachycardia syndrome, an arrhythmia being investigated in clinical trials as a way to suppress where intra-atrial re-entrant rhythm alternates with atrial fibrillation but is not yet commercially available severe bradycardia, is considered a Class IIa indication (9,10). in patients following congenital heart surgery. This rhythm disturbance has been associated with signifi- cant morbidity and mortality in this population Indications for permanent pacemaker (13,14). Permanent pacing has been shown to prevent implantation these episodes (15,16). In 2008, the American College of Cardiology, Ameri- Third-degree AV block can be classified as either can Heart Association, and Heart Rhythm Society acquired or congenital in the pediatric population and updated the guidelines for pacemaker implantation is the most common indication for pacing in children. and included recommendations for pediatrics and The commonest cause of acquired heart block in chil- CHD (11). The main indications for pacing in children dren is postoperative AV block and is usually a result are sino-atrial (SA) node disease (including brady– of operations involving the interventricular septum tachy syndrome and symptomatic sinus bradycardia) around the area of the AV node. The current reported and complete AV block (11). incidence stands at 1–3% (17,18). Patients with L-TGA An increasing number of pediatric patients are sur- or heterotaxy are also prone to developing complete viving palliative congenital heart surgery with impaired AV block even without surgical intervention. A poor hemodynamics. Signs and symptoms that may not outcome has been described in patients with perma- require pacing in an adult may have a significant nent postsurgical AV block who do not receive pace- impact on a child with borderline hemodynamics. maker therapy (19). Thus, pacing in the postoperative Sinus node dysfunction is not common in children patient with AV block lasting more than 7 days is con- with structurally normal hearts. It is more likely after sidered a Class I indication (11). Congenital complete certain types of palliation, e.g., late complication of AV block is usually associated with maternal tetralogy of fallot repair, after a Senning or Mustard lupus antibodies and presents as fetal or neonatal

Table 2 Indications for implantation of permanent pacemaker in children and adolescents. Adapted from Refs (5,11). Source: American Heart Association, Inc.

Class 1: Pacemaker implantation indicated • Advanced second/third-degree AV block with symptomatic bradycardia, ventricular dysfunction, or low cardiac outputs or that persists at least 7 days postcardiac surgery • Symptomatic age inappropriate sinus bradycardia • Congenital third-degree AV block with wide QRS, complex ventricular ectopy, or ventricular dysfunction ) ) • Congenital third-degree AV block in an infant with HR <55 bÆmin 1 or HR <70 bÆmin 1 in infant with CHD Class IIa: Pacemaker implantation reasonable • Patients with CHD, sinus bradycardia, and intra-atrial reentry tachycardia • Asymptomatic sinus bradycardia in child with complex CHD with resting HR <40 bÆmin)1 or pauses >3 s • Patient with CHD and impaired hemodynamics because of sinus bradycardia and loss of AV synchrony • Unexplained syncope in repaired CHD patient complicated by transient AV block with residual fasicular block Class IIb: Pacemaker implantation may be considered • Transient postop AV block that reverts to sinus rhythm with residual bifasicular block • Congenital third-degree AV block in asymptomatic child/adolescent with acceptable rate, narrow QRS complex, and normal ventricular function ) • Asymptomatic sinus bradycardia after biventricular repair of CHD with resting HR <40 bÆmin 1 or pauses >3 s Class III: Pacemaker implantation not effective/contraindicated • Transient postop AV block that reverts to normal AV conduction • Asymptomatic Type I AV block ) • Asymptomatic sinus bradycardia in adolescent with resting HR >40 bÆmin 1 and longest R–R interval <3 s • Asymptomatic postop bifasicular block in the absence of transient complete AV block

CHD, congenital heart disease; AV, atrioventricular.

514 Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd M. Navaratnam and A. Dubin Pediatric pacemakers and ICDs bradycardia. This condition is associated with high postnatal mortality rates, and the bradycardia may result in a secondary dilated cardiomyopathy. Pace- maker implantation and normalization of heart rates are associated with a decrease in heart size and improvement in systolic function in these patients (20). Cardiac resynchronization therapy is a way of improving mechanical efficacy and favors remodeling in hearts with ventricular dysynchrony. It aims to improve the timing and completeness of contraction in the right and left ventricle (for two ventricle patients) or to improve synchronized contraction within a single ventricle. There is emerging data, which indicates that long-term pacing of the right ventricle can cause sec- ondary dilated cardiomyopathy and adverse remodel- ing of the left ventricle in some patients with congenital third degree AV block (21–23). The most recent recommendation for resynchroniza- tion device implantation is adult guidelines (11) and include ejection fraction (EF) £35%, QRS >0.12 ms and New York Heart Association (NYHA) Class 3–4. Figure 1 CXR showing epicardial dual chamber pacemaker and implantable cardioverter defibrillator (ICD). Note bipolar pacemaker Although there are no specific indications for children leads in right atrium and right ventricle and ICD defibrillator coil and CHD patients, resynchronization therapy has been (thick loop) on the lateral surface of the left ventricle attached to used successfully in pediatric patients with poor func- the . tion and electromechanical dysynchrony, who do not strictly meet the adult criteria. A retrospective multi- center study of 103 patients from 22 institutions pub- started. The device chooses antitachycardia pacing or lished in 2005 (24) demonstrated the short-term shock depending on the presentation and programming. efficacy of resynchronization therapy in pediatrics and Most ICDs will also start pacing when the R–R CHD patients. This cohort experienced an improve- interval is too long (antibradycardia function). All the ment in the mean EF from 26% to 40%. tachycardia sensors in current use for ICDs are taken from the ECG. About 20% of ICD patients have devices that incor- Pediatric ICD overview porate sophisticated dual and biventricular pacing Implantable cardioverter defibrillators have been in use modes for permanent pacing (26). for the past 20 years for primary and secondary pre- Implantable cardioverter defibrillators also have an vention of sudden cardiac death. The ICD has all the international generic code (Table 3). For complete capabilities of a pacemaker, with the additional poten- identification, position IV is expanded into its full tial for defibrillation of tachyarrhythmias (usually ven- pacemaker code. For example, most devices with a tricular). It is also possible to program the defibrillator rate responsive pacemaker and ICD will be identified to pace terminate tachyarrhythmias such as mono- as VVE-DDDR. morphic ventricular tachycardia. Implantable cardioverter defibrillators are usually Indications for ICD implantation in children implanted transvenously using a lead that allows for sensing, pacing, and defibrillation. However, in small Implantable cardioverter defibrillators are primarily children and patients with CHD, it may not be possi- implanted in patients who suffer a sudden cardiac ble to use the transvenous route. In that situation, arrest or those at high risk of malignant ventricular novel configurations have been used, including placing arrhythmias. The majority of patients with ICDs are the coil in the pericardial or subcutaneous space (25) adult patients with ischemic heart disease and poor (Figure 1). EF. The published guidelines on the implantation of Implantable cardioverter defibrillators measure each these devices in pediatrics are derived primarily from cardiac R–R interval, and when enough short R–R adult randomized clinical trials (11) (Table 4). In con- intervals are detected, an antitachycardia event is trast to adults, there are currently no large prospective

Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd 515 Pediatric pacemakers and ICDs M. Navaratnam and A. Dubin

Table 3 Generic defibrillator code NAPSE/ Position I Position II Position III Position IV BPEG. Adapted from Ref. (8) Shock Antitachycardia Tachycardia Antibradycardia chamber(s) pacing chamber(s) detection pacing chamber(s)

O = None O = None E = Electrogram O = None A = Atrium A = Atrium R = Rate modulation A = Atrium V = Ventricle V = Ventricle V = Ventricle D = Dual D = Dual D = Dual (A + V) (A + V) (A + V) studies examining the efficacy of implantable defibrilla- QT syndrome, Brugada syndrome, and catecholamine- tors in children, and <1% of devices are implanted in induced polymorphic ventricular tachycardia (VT). the pediatric population. However, as technology The authors reported that 26% of patients received advances, the use of ICDs in the pediatric and CHD appropriate shock therapy and 21% had inappropriate population has expanded. shocks. The main reasons for inappropriate shocks In 2008, Berul et al. (27) published the largest were lead failure (14%), sinus/atrial tachycardia (9%), reported series of pediatric and congenital heart dis- and oversensing (4%). ease ICD patients. They conducted a retrospective This relatively high rate of inappropriate shocks is review from four pediatric centers in the USA and col- similar to that seen in other published pediatric series lated information for all implants between 1992 and (28,29), and the authors concluded that higher inappro- 2004. A total of 443 patients were included with a priate shock rates in children vs adults with CHD sup- median age of 16 years. In 52% of patients, the ICD ports the hypotheses that continued growth and activity was implanted for primary prevention; 69% of patients place a strain on the ICD leads. In another study (29), had structural hearts disease such as Tetralogy of Fal- growth was strongly associated with lead failure. lot (21%) and hypertrophic cardiomyopathy (14%). Thus, while ICDs in children may be life saving, The remaining 31% of patients had structurally nor- there are a high number of inappropriate shocks and mal hearts with primary electrical diseases such as long lead failures seen in this population. This should be

Table 4 Indications for implantable cardiac defibrillator in children and patients with CHD. Adapted from Refs (5,11). Source: American Heart Association, Inc.

Class I: ICD implantation indicated • Cardiac arrest because of VF/VT and reversible cause excluded • Spontaneous sustained VT in CHD patients • Symptomatic sustained VT in CHD patients who have had EPS • Unexplained syncope in CHD patients with EPS inducible sustained, hemodynamically significant VT Class IIa: ICD implantation reasonable • Recurrent unexplained syncope in CHD patients with LV dysfunction or EPS-induced ventricular arrhythmias • Sustained VT with near normal ventricular function • Long QT or catecholamine-induced VT with recurrent syncope or VT despite B blockers • HCM or arrythmogenic right ventricular dysplasia with one or more risk factors for sudden cardiac death • Brugada syndrome and syncope or VT that has not resulted in cardiac arrest • Nonhospitalized patients awaiting cardiac transplantation Class IIb: ICD implantation may be considered • Complex CHD patients with recurrent syncope and advanced ventricular dysfunction where thorough investigations failed to define cause • Long QT syndrome and risk factors for sudden cardiac death • Familial cardiomyopathy associated with sudden death • Patients with LV noncompaction Class III: ICD implantation not effective/contraindicated • VT or VF from arrhythmias amenable to catheter ablation, e.g., WPW • Unexplained syncope with structurally normal heart and no inducible ventricular tachyarrythmias • Ventricular tachyarrythmias because of reversible cause in absence of CHD • Incessant VT/VF • Patients who do not have reasonable expectation of survival/functional status of 1 year

ICD, implantable cardioverter defibrillator; CHD, congenital heart disease; EPS, .

516 Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd M. Navaratnam and A. Dubin Pediatric pacemakers and ICDs

The preoperative assessment should include a focused history and examination to determine the indi- cation for CRMD, coexisting cardiovascular patho- logy, pocket generator location, and any cardiac conditions that may influence defibrillator pad posi- tion, e.g., Dextrocardia. Ideally, patients will have a device information card containing the make, model, and serial number of the device. Patient medical records or old operative notes may also have this important information and may contain the most recent interrogation report. A patient history of symptomatic bradycardia or syncope or successful AV nodal ablation prior to device implantation may indicate pacemaker depen- dency. A chest x-ray (CXR) can help in determining the type of device (ICD vs pacemaker) and whether it is single, dual, or biventricular (Figures 1 and 2). It can also be used to determine the number, position, and integrity of leads as well as any unusual configura- Figure 2 CXR showing transvenous dual chamber pacemaker tions of lead/generator placement and tunneling that implantable cardioverter defibrillator. Note pace/sense bipolar elec- may impact the surgical approach. trode in right atrium and defibrillator coil in right ventricle. A recent 12-lead surface ECG should also be obtained to determine whether the pacer is being used and which chamber is being sensed and/or paced. kept in mind when managing these patients during the Ultimately, the only reliable method of assessing perioperative period. battery status, lead placement, and adequacy of pace- maker/ICD settings is direct interrogation with a pro- grammer. This is usually performed by dedicated Preoperative assessment and preparation electrophysiology staff, cardiologist, or an industry Adequate preoperative assessment and preparation are representative. Pacemakers should be routinely interro- essential in the pediatric patient with a CRMD. gated once a year and ICDs every 3–4 months (9,26). The ASA practice advisory for the perioperative It is recommended that all devices with ICD function management of patients with a CRMD recommends should be assessed within 30 days of surgery (36) but that the patient should have a preop evaluation focus- there is no agreed consensus on acceptable timing for ing on three main areas (30): those with only pacemaker function. A copy of the 1. Type of device most recent interrogation report should be obtained, 2. Dependency on device for antibradycardia pacing which should include the patient’s underlying rate and and underlying escape rhythm rhythm and the need for intraoperative back up pacing 3. Device function support. It may also contain information on whether a Several investigators have shown that an incomplete magnet mode is present and the preset magnet preoperative evaluation may lead to adverse outcomes response of the device. (31–33). An observational study of 60 adult patients For elective surgery, it is always prudent to discuss scheduled for noncardiac surgery showed that preoper- the case in advance with a cardiologist or qualified ative interrogation detected pacemaker dysfunction in programmer informing them of the surgery date, type, 12% of patients (34), thus highlighting the importance and sources of EMI so that they have adequate time of a preoperative evaluation. for device optimization. In adults, battery malfunction is the most common Figure 3 summarizes a practice algorithm that may cause of generator device failure with ICD malfunction be used for optimization of devices during the peri- rate substantially higher than pacemakers (35). operative period. Occasionally emergency surgery will Children are considered to have a higher risk for preclude a full interrogation of the device. In those sit- lead failure and fracture than adults. Lead failure uations, it is still important to use as many of the remains the most common cause of inappropriate strategies depicted as possible depending on the time shocks and ICD complications in children (27). constraints of the emergency procedure.

Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd 517 Pediatric pacemakers and ICDs M. Navaratnam and A. Dubin

Indication for CRMD Co-existing cardiovascular pathology Focused history and Manufacturer’s CRMD card examination Determine generator location Supplemental resources

CXR +12 Lead ECG Pre-op Recent electrolytes Diagnostic evaluation ICD Interrogation <30 days ? Recent Interrogation report

Underlying rate/rhythm Consultation with Surgery date/Type cardiologist/qualified Sources of EMI programmer Need for reprogramming/backup pacing Determine Magnet response

No Is EMI likely? Proceed with surgery

Yes

Is patient pacer No dependent or Consider magnet use surgery < 15cm of generator? Intra-op Yes Reprogram device to asynchronous mode rate > patient’s intrinsic rate

Yes Consider increasing rate to optimize oxygen delivery Place defib pads in AP Disable anti-tachycardia function for ICD position away from generator Disable all rate enhancements Monitor cardiac rhythm and peripheral pulse Use bipolar diathermy where possible If monopolar use cutting Maintain electrolyte and metabolic normality mode with short bursts

Post-op interrogation and re-set CRMD Determine optimum heart rate to meet post op metabolic demands Post-op Figure 3 Practice algorithm for peri-opera- Set pacing parameters to appropriate pacing/sensing threshold tive management of cardiac rhythm man- Monitor heart rhythm and rate agement device (CRMD).

There is also a risk that a current can be induced in Electromagnetic interference and reprogramming the leads resulting in electrical discharge to the myo- Safe perioperative management of pacemakers is cardium, which may produce arrhythmias or even a dependent on understanding and effectively managing burn. EMI in the operating room. Modern devices are less susceptible to EMI because Electromagnetic interference, radio-frequency waves of the use of bipolar leads, improved filters and circuit in the 50–60 Hz range, can be generated in the periop- shields, which insulate the internal electronics from the erative setting in a multitude of ways. Electrocautery, metal casing device, and improved noise protection transthoracic defibrillation, therapeutic radiation, algorithms, which filter signals outwith the range of radio-frequency ablation, extracorporeal shock wave normal cardiac noise. However, the risk of EMI inter- lithotripsy, MRI, nerve stimulators, fasciculations, ference still exists even with modern, sophisticated shivering, and large tidal volumes have all been shown devices. to cause EMI (9,30). The ASA task force recommends reprogramming to Electromagnetic interference can cause either inap- an asynchronous paced mode in patients who are pace- propriate triggering or inhibition of paced output as maker dependent when there is a significant risk of well as reversion to an asynchronous (noise suppres- EMI (30). Reprogramming to an asynchronous mode sion) mode (9). Asynchronous pacing may compete at a rate higher than the patients’ intrinsic rate will with a spontaneous rhythm and lead to an arrhythmia. overcome potential oversensing or undersensing from For ICDs, inappropriate triggering may lead to inap- EMI. If possible, qualified personnel should be con- propriate shock or antitachycardia therapy. tacted before the procedure to allow time for optimal

518 Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd M. Navaratnam and A. Dubin Pediatric pacemakers and ICDs device management. Reprogramming should be per- according to the device manufacturer and model. For formed prior to skin preparation and draping of the example, Medtronic magnet rate is usually 85 bÆmin)1, patient so that access to the device is not hindered. In whereas the St Jude rate is usually 100. It is important emergency situations, where time may preclude repro- to appreciate that a preset magnet rate may not be suf- gramming by qualified personnel, a magnet may be ficient to meet the metabolic demands of a pediatric placed over the device to revert to asynchronous pac- patient. The preset magnet rate may also vary accord- ing. Occasionally switching to asynchronous pacing ing to the battery voltage (i.e., life) of the generator, may cause hemodynamic disturbances; therefore, it is e.g., for Boston Scientific devices, the preset magnet important to have all the patient monitors in place rate is usually 100 bÆmin)1 but may decrease to 80 before changing modes. when there is only 3 months left on the battery life. Special algorithms such as rate adaptive functions Magnet placement on a device at the end of service life should be disabled (30). can cause cessation of pacing altogether. It is also recommended that the ICD antitachyar- Although most ICDs will suspend antitachycardia rhythmia function should be disabled (30) and a quali- therapy when a magnet is placed over the device, like fied programmer should do this. Although a magnet pacemakers, magnet response varies according to the should not be used routinely to disable ICDs (30), it manufacture and device program. There are also dif- can be used in an emergency situation. Patients with ferences between manufacturers with regard to where disabled ICDs should have transthoracic defibrillator to position the magnet (36). pads in place. Only an interrogation with a programmer can reveal Reprogramming a device will not protect it from current magnet response settings. When a magnet is damage or reset caused by EMI. Between 1984 and used for CRMDs with combined pacemaker/ICD func- 1997, the US FDA was notified of 456 adverse events tion, the ICD function will be suspended but the associated with pulse generators and at least 50% were device will not be forced to revert to an asynchronous related to diathermy (26). pacing model. Patients with these devices who are Adequate monitoring remains vital, and it is prudent pacemaker dependent need to have their device repro- to reduce EMI as much as possible. This includes use grammed to an asynchronous mode preoperatively if of bipolar diathermy or ultrasonic especially if there is significant risk of EMI such as diathermy. surgery is within 15 cm of the generator. If monopolar In general, details about magnet response can be diathermy must be used, then cutting (minimal power obtained from the manufacturer’s helpline. It is impor- setting) rather than coagulation current and short tant before starting a procedure for a patient with a intermittent irregular bursts <1 s duration should be CRMD to know where the nearest magnet is situated used (37). and to locate the patient’s generator position in case it Good skin contact with the diathermy grounding needs to be accessed in an emergency. pad should be ensured, and it should be placed far away from the device, in a way that current flow will Intraoperative management not intersect the pacing system (26). For example, for head and neck surgery, the grounding pad should be It is essential that ECG monitoring of the patient placed on the shoulder contralateral to the device and includes the ability to continuously detect pacemaker not on the thigh. activity. A perfused peripheral pulse should be moni- tored with a waveform display, e.g., pulse oximetry or invasive pressure monitoring. Other modalities that Magnet use confirm pulse generation during pacing include manual The appropriate role of magnets in the perioperative palpation and auscultation via a precordial or esopha- period remains controversial. Magnet activated geal stethoscope. One should have a low threshold for switches were incorporated into pacemakers to demon- beat to beat arterial pressure monitoring especially for strate remaining battery life and sometimes pacing long procedures with significant fluid shifts and in threshold safety factors (26). They are not specifically patients with resynchronization devices, heart failure, designed to treat pacemaker emergencies or to treat or hypertrophic cardiomyopathy. Temporary pacing EMI effects. Placing a magnet over the center of the and defibrillation equipment should be immediately pacemaker will usually cause it to revert to an asyn- available before, during, and after the procedure. chronous pacing mode (AOO, VOO, or DOO) where It is recommended that transthoracic defibrillator no intrinsic cardiac activity is sensed. The pacemaker pads should be placed in an anterior–posterior position will pace at a preset ‘magnet rate’ which will vary and as far away from the pulse generator as possible.

Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd 519 Pediatric pacemakers and ICDs M. Navaratnam and A. Dubin

(9,26,36). Transthoracic defibrillation may result in hypercarbia have also been shown to independently reprogramming of the pulse generator, damage to the increase the myocardial stimulation threshold (38). device circuits or myocardial burns. If an external defi- brillator has been used during a procedure, it is essen- Postoperative care tial to interrogate the device postop to assess for any damage. Any pacemaker that was reprogrammed for the sur- Although the ASA task force believes that anesthetic gery should be reset at the end of the procedure. Most techniques do not influence CRMD function (30), it is manufactures advise re-interrogation of all devices important to pay attention to choice of anesthetics in postop especially if monopolar diathermy, significant patients with CRMD devices. Drugs that suppress the fluid or blood component administration, or external AV or SA node (e.g., dexmedetomidine, potent opi- defibrillation have been used (30). Pacing and sensing ates) may render the patient pacer dependent. Inhala- thresholds should be measured, and the device should tion agents such as Isoflurane or Sevoflurane may be programmed to optimize hemodynamics in the exacerbate long QT syndrome and cause arrhythmias. postoperative setting. Pediatric patients with DDD Three main reasons for pacemaker failure are failure pacing who develop sinus tachycardia may reach their to capture, lead failure, and generator failure. The last Wenckebach or 2 : 1 block point, so it is helpful to two are rare in a system that has been adequately know what these are for each patient and try to avoid interrogated preop. If lead failure does occur in a extreme sinus tachycardia. ICD patients should be patient who is pacemaker dependent, alternative meth- monitored until the antitachycardia therapy is ods of pacing include external restored. and temporary transvenous pacing. Pacemakers may exhibit acutely elevated lead thres- Conclusion holds and failure to capture with metabolic perturba- tion or ischemia (26,38). Electrolyte and metabolic Pediatric patients with permanent pacemakers and abnormalities, especially hyperkalemia, metabolic aci- ICDs present unique management challenges during dosis and alkalosis, hypothermia, and hyperglycemia both cardiac and noncardiac surgery. It is critical for all significantly increase the pacing threshold (38). the anesthesiologist caring for these patients to be Therefore, large fluid shifts, dehydration, and signifi- adept at assessing and managing patients with these cant bleeding during the perioperative period may in devices in the perioperative period. turn influence pacemaker function. Hypoxemia and

References 1 Aellig NC, Balmer C, Dodge-Khatami A population. Pacing Clin Electrophysiol 2009; abnormalities: a report of the American et al. Long-term follow-up pacemaker 32: S71–S74. College of Cardiology/American Heart implantation in neonates and infants. Ann 7 Mathony U, Schmidt H, Groger C et al. Association Task Force on Practice Guide- Thorac Surg 2007; 83: 1420–1423. Optimal maximum tracking rate of dual lines (Writing Committee to Revise the 2 Fortescue EB, Berul C, Cecchin F et al. chamber pacemakers required by children ACC/AHA/NASPE2002 Guideline Update Patient, procedural and hardware factors and young adults for a maximal cardiorespi- for Implantation of Cardiac Pacemakers associated with pacemaker lead failures in ratory performance. Pacing Clin Electro- and Antiarrhythmia Devices) Developed in pediatrics and congenital heart disease. physiol 2005; 28: 378–383. Collaboration With the American Associa- Heart Rhythm 2004; 1: 150–159. 8 Bernstein AD, Daubert J-C, Fletcher RD tion for Thoracic Surgery and Society of 3 Cohen MI, Bush DM, Vetter VL et al. Per- et al. NASPE POSITION STATEMENT. Thoracic Surgeons. Circulation 2008; 117: manent epicardial pacing in pediatric The revised NASPE/BPEG generic code for E350–E408. patients: seventeen years of experience and antibradycardia, adaptive-rate, and multisite 12 Batra AS, Balaji S. Pacing in adults with 1200 outpatient visits. Circulation 2001; 103: pacing. Pacing Clin Electrophysiol 2002; 25: congenital heart disease. Expert Rev Cardio- 2585–2590. 260–264. vasc Ther 2006; 4: 663–670. 4 Silvetti MS, Drago F, Grutter G et al. 9 Stone ME, Apinis A. Current perioperative 13 Garson A Jr, Bink-Beolkens M, Hesslein PS Twenty years of pediatric cardiac pacing: management of the patient with a cardiac et al. Atrial flutter in the young: a collabo- 515 pacemakers and 480 leads implanted rhythm management device. Semin Cardio- rative study of 380 cases. J Am Coll Cardiol in 292 patients. Europace 2006; 8: 530– thorac Vasc Anesth 2009; 13: 31–43. 1985; 6: 871–878. 536. 10 Knight BP, Gersh BJ, Carlson MD et al. 14 Gelatt M, Hamiltion RM, McCrindle BW 5 Chun T. Pacemakers and defibrillator ther- Role of permanent pacing to prevent atrial et al. Arrhythmias and mortality after the apy in pediatrics and congenital heart fibrillation. Circulation 2005; 111: 240–243. mustard procedure: a 30-year single center disease. Future Cardiol 2008; 4: 469–479. 11 Epstein AE, Di Marco JP, Elenborgen KA experience. J Am Coll Cardiol 1997; 29: 6 Shah MJ. Implantable cardioverter defibril- et al. ACC/AHA/HRS2008 guide lines for 194–201. lator-related complications in the pediatric device-based therapy of cardiac rhythm

520 Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd M. Navaratnam and A. Dubin Pediatric pacemakers and ICDs

15 Silka MJ, Manwill JR, Kron J et al. Brady- 23 Kim JJ, Friedman RA, Eidem BW et al. of patients with cardiac rhythm manage- cardia mediated tachyarrhythmias in con- Ventricular function and long-term pacing ment devices: pacemakers and implantable genital heart disease and responses to in children with congenital complete cardioverter defibrillators: a report by the chronic pacing at physiological rates. Am J atrioventricular block. J Cardiovasc Electro- American Society of Anesthesiologists Task Cardiol 1990; 65: 488–493. physiol 2007; 18: 373–377. Force on Perioperative Management of 16 Stephenson EA, Casavant D, Tuzi J et al. 24 Dubin AM, Janousek J, Rhee E et al. Patients with Cardiac Rhythm Management Efficacy of atrial antitachycardia pacing Resynchronization therapy in pediatric and Devices. Anesthesiology 2005; 103: 186–198. using the Medtronic AT500 pacemaker in congenital heart disease patients: an interna- 31 Mychaskiw G, Eichhorn JH. Interaction of patients with congenital heart disease. Am J tional multicenter study. J Am Coll Cardiol an implanted pacemaker with a transesoph- Cardiol 2003; 92: 871–876. 2005; 46: 2277–2283. ageal atrial pacemaker: report of a case. 17 Andersen HO, de Leval MR, Tsang VT 25 Stephenson EA, Batra AS, Knilans TK J Clin Anesth 1999; 11: 669–671. et al. Is complete heart block after surgical et al. A multicenter experience with novel 32 Prday JP, Towey RM. Apparent pacemaker closure of ventricular septum defects still an implantable cardioverter defibrillator config- failure caused by activation of ventricular issue? Ann Thorac Surg 2006; 82: 948–956. urations in the pediatric and congenital threshold test by a magnetic instrument dur- 18 Gross GJ, Chiu CC, Hamilton RM et al. heart disease population. J Cardiovasc ing general anaesthesia. Br J Anaesth 1992; Natural history of postoperative heart block Electrophysiol 2006; 17: 41–46. 69: 645–646. in congenital heart disease: implications for 26 Rozner MA. The patient with a cardiac 33 Kellow NH. Pacemaker failure during tran- pacing intervention. Heart Rhythm 2006; 3: pacemaker or implanted defibrillator and surethral resection of the prostate. Anesthe- 601–604. management during anesthesia. Anesthesia sia 1993; 48: 136–138. 19 Lillehei CW, Sellers RD, Bonnabeau RC and medical disease. Curr Opin Anaesthesiol 34 Pili-Floury S, Farah E, Samain E. Perioper- et al. Chronic post surgical complete heart 2007; 20: 261–268. ative outcome of pacemaker patients under- block with particular reference to prognosis, 27 Berul CI, VanHare GF, Kertesz MJ et al. going noncardiac surgery. Eur J management and a new P-wave pacemaker. Results of a multicenter retrospective Anaesthesiol 2008; 25: 514–516. J Thorac Cardiovasc Surg 1963; 46: 436– implantable cardioverter-defibrillator regis- 35 Maisel WH. Pacemaker and ICD generator 456. try of pediatric and congenital heart disease reliability. JAMA 2006; 295: 1929–1934. 20 Breur JM, Udink Ten Cate FE, Kapusta L patients. J Am Coll Cardiol 2008; 51: 36 Joshi GP. Perioperative management of et al. Pacemaker therapy in isolated congen- 1685–1691. outpatients with implantable cardioverter ital complete atrio-ventricular block. Pacing 28 Silka MJ, Kron J, Dunnigan A et al. Sud- defibrillators. Curr Opin Anaesthesiol 2009; Clin Electrophysiol 2002; 25: 1685–1691. den cardiac death and the use of implant- 22: 701–704. 21 Moak JP, Hasbani K, Ramwell C et al. able cardioverter defibrillators in pediatric 37 Rozner MA. Review of electrical interfer- Dilated cardiomyopathy following RV pac- patients. Circulation 1993; 87: 800–807. ence in implanted cardiac devices. Pacing ing for AV block. J Cardiovasc Electrophysiol 29 Alexander ME, Cecchin F, Walsh EP et al. Clin Electrophysiol 2003; 26: 923–925. 2006; 10: 1072–1073. Implications of implantable cardioverter 38 Dohrmann ML, Goldschlager NF. Myocar- 22 Udink ten Cate FEA, Breur JMPJ, Cohen defibrillator therapy in congenital heart dis- dial stimulation threshold in patients with MI et al. Dilated cardiomyopathy in iso- ease and pediatrics. J Cardiovasc Electro- cardiac pacemakers: effect of physiologic lated congenital complete atrioventricular physiol 2004; 15: 72–76. variables, pharmacologic agents, and lead block: early and long-term risk in children. 30 Zaidan JR, Atlee JL, Belott P et al. Practice electrodes. Cardio Clinic 1985; 3: 527–537. J Am Coll Cardiol 2001; 37: 1129–1133. Advisory for the perioperative management

Pediatric Anesthesia 21 (2011) 512–521 ª 2011 Blackwell Publishing Ltd 521 Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Cerebral oximetry for pediatric anesthesia: why do intelligent clinicians disagree? Nicholette Kasman1 & Ken Brady2

1 Department of Pediatric Anesthesia, Stanford University School of Medicine, Stanford, CA, USA 2 Department of Pediatrics and Anesthesia, Texas Children’s Hospital, Baylor College of Medicine, Houston, TX, USA

Keywords Summary near-infrared spectroscopy; pediatrics; congenital heart disease Reflectance near-infrared spectroscopy has been used to measure cortical tissue oximetry for more than 30 years. In that time, many centers have Correspondence adopted the routine use of the cerebral oximeter for children having repair Ken Brady, of congenital heart lesions, while some prominent academic centers have Texas Children’s Hospital, 6621 Fannin, WT resisted routine use of these monitors citing lack of definitive evidence for 17-417B, Houston, TX 77030, USA outcome benefit. In this review, we provide an overview of the method Email: [email protected] used to measure cerebral oximetry, as well as validation and clinical out- Section Editor: Chandra Ramamoorthy come data that have accrued from the use of cerebral oximeters. We dis- cuss the peculiarities of evidentiary review for monitoring devices, and the Accepted 19 January 2011 confounding errors that occur when a monitor is evaluated as a therapeutic intervention. We outline the physiologic basis of cerebral desaturation and doi:10.1111/j.1460-9592.2011.03549.x the shifts in practice that have occurred with implementation of NIRS monitoring.

of care of the critically ill child, providers will have no Introduction room to maneuver effectively. While the academic Cerebral oximetry with near-infrared spectroscopy intransigent appears to lack common sense, common (NIRS) is a noninvasive, continuous assessment of sense can also be misleading, harm patients, and drive brain oxygen delivery and utilization (1). NIRS-based up medical costs (4,5). Considering that none of the cerebral oximeters quantitate a venous-weighted ratio monitors currently used in the operating room have of oxygenated and deoxygenated hemoglobin in the been shown to improve long-term outcome, how can region of cerebral cortex underlying the sensors, usu- the operative team evaluate the cost/benefit ratio of ally placed on the forehead. Initially hailed as the ideal cerebral oximetry monitoring? neonatal neuromonitor, currently marketed cerebral There may not be an accessible way to delineate the oximeters have gained more traction in the operating value or harm of a monitoring device, as precedent room functioning more as hemodynamic monitors, attempts with other monitors have failed to do so where they are practically suited for use during cardio- (6,7). Therapeutic interventions for critically ill chil- pulmonary bypass. There is no standard of care for dren are often performed with imperfect evidentiary the use of NIRS-based cerebral oximetry in Pediatric data, commonly extrapolated from adult studies, or Anesthesia, although it has its greatest presence in absent entirely (8). Given the latitude of clinical prac- pediatric cardiac anesthesia. Proponents of NIRS-mon- tice resulting from data paucity, it is the responsibility itoring cite validation data, physiologic principles, and of the pediatric intensivist and anesthesiologist to be common sense (2), while opponents of the use of familiar with the data that are available. We will pro- NIRS cite a lack of rigorously demonstrated outcome vide a focused overview of NIRS technology, a brief improvement and cost (3). Evidence-based medicine review of specific validation data for NIRS use, clinical distinguishes medicine as a science from mysticism. outcome data relevant to pediatric practice, and a pro- However, if strict evidence is required for every aspect posed clinical treatment algorithm for desaturation

Pediatric Anesthesia 21 (2011) 473–478 ª 2011 Blackwell Publishing Ltd 473 Cerebral oximetry for pediatric anesthesia N. Kasman and K. Brady events detected with NIRS. We will discuss the reflectance arc (including arterial, venous and capillary changes in practice that have occurred temporally with hemoglobin), the resulting number is biased toward the use of NIRS monitoring. the larger venous hemoglobin mass, which is consis- tently higher than, but correlated with jugular venous oximetry (10,11). Where the pulse oximeter is a useful Reflectance NIRS trend of pulmonary function and the a-A gradient, Frans Jobsis first described the measurement of cerebral cerebral oximetry trends the ratio of regional oxygen oxyhemoglobin content using reflectance NIRS in 1977 delivery and utilization to detect cerebral ischemia. (9). Near-infrared light with a wavelength between 650 and 950 nm penetrates biologic tissue without harmful Validation data ionization and is absorbed by a limited number of chro- mophores found in human tissue – notably hemoglobin The first validation of any proposed monitoring device

(Hb), oxyhemoglobin (Hb-O2), and cytochrome aa3. is to determine its accuracy. There is no gold standard Further, these chromophore species absorb light differ- of tissue oxyhemoglobin content, so jugular venous entially across the near-infrared bandwidth, so the use oximetry is often used as the standard for comparison. of multiple wavelengths of light permits quantitation of In piglet studies measuring both cerebral and jugular the relative concentrations of Hb and Hb-O2 separately. oximetry during cardiopulmonary bypass and circula- Cytochrome aa3 is the last enzyme in the mitochondrial tory arrest, a strong correlation between cerebral and oxygen transport chain, so there has been interest in jugular oximetry was reported (r = 0.91) across a full measuring the redox-state of this molecule as a marker range of jugular saturations (12). of cellular oxygen availability, but this technique has Clinical replication of this animal data has been less been limited to investigational work. consistent. Children and infants have been studied under Although the marketed cerebral oximeters and pulse conditions of cardiopulmonary bypass, extracorporeal oximeters use proprietary algorithms to determine the membrane oxygenation, and cardiac catheterization. In relative concentrations of Hb and Hb-O2, the method these clinical scenarios, the jugular oxyhemoglobin con- is best understood with the basic equation of spectros- tent is accessible, so comparisons have been made with copy, the simplified Beer–Lambert equation: cerebral oximetry. An early study of 40 children having I A ¼log ¼ ekLC, where A is absorbance, I and I0 a mixture of catheterization and bypass-requiring proce- I0 are recovered and incident light intensities, respec- dures reported modest correlation (r = 0.69) and a bias tively, ek is the molar absorptivity of the absorbing of high cerebral oximetry at low jugular oximetry (13). molecule being measured, at the wavelength of light Subsequent studies found higher correlations, similar to used, L is the pathlength of the light and C is the con- the animal data, including a report of 30 children having centration of the molecule. The reflectance technique catheterization and 17 neonates on extracorporeal mem- does differ from the precise analytical technique brane oxygenation support (10,14). applied to cuvettes in a spectrophotometer. Instead of Cerebral oximetry is a purported measure of oxida- passing through the tissue for complete recovery on tive substrate availability to the brain, so an alternative the opposite side, the near-infrared light is reflected at validation standard was the use of mass spectroscopy tissue interfaces in a banana-shaped arc that is recover- to measure tissue phosphate energy stores. Reductions able on the ipsilateral surface with sufficient intensity in cerebral oximetry have been correlated with reduc- to trend the concentration of the chromophores being tions in tissue adenosine triphosphate and phosphocre- measured. It is not hard to imagine that imperfections atine (15). There has been little evidence to challenge in the estimation of pathlength and scatter introduced the claim that cerebral oximetry detects an ischemic by reflectance are canceled by expressing the resultant milieu in the cerebral cortex. Clinical relevance has concentrations as a ratio: been the subject for most of the debate regarding NIRS monitoring. ½Hb O2 Hb O2saturation ¼ ½þHb ½Hb O2 Clinical outcome data By comparison to pulse oximetry, cerebral oximeters trend venous-weighted measurements because the A comprehensive, systematic, and academically rigor- entire returned signal is measured rather than just the ous review of the evidence for the use of cerebral pulsatile measurements that make pulse oximetry spe- oximetry was performed by Hirsch et al., concluding: cific to arterial blood oxygen saturation (6). Because ‘Although near-infrared spectroscopy has promise for cerebral oximetry interrogates all hemoglobin in the measuring regional tissue oxygen saturation, the lack

474 Pediatric Anesthesia 21 (2011) 473–478 ª 2011 Blackwell Publishing Ltd N. Kasman and K. Brady Cerebral oximetry for pediatric anesthesia of data demonstrating improved outcomes limits the monitoring from care. How have other monitoring support for widespread implementation’ (16). Note- devices been evaluated by evidence-based medicine worthy studies included in this evaluation are summa- standards? A comprehensive review was made for the rized below. Cochrane database, evaluating the more familiar and Austin et al. performed a nonrandomized, prospec- uncontestedly standard of care pulse oximeter. Studies tive study of multimodal neuromonitoring (including of pulse-oximetry inclusion and exclusion in periopera- cerebral oximetry), in 250 children during cardiac sur- tive care included data from more than 22 000 gery. Monitoring events were strictly defined and patients. Although considered a standard of care by recorded, as well as interventions prompted by the the ASA, the Cochrane database review ‘found no evi- events. When electroencephalogram, transcranial dence that pulse oximetry affects the outcome of anes- Doppler ultrasonography, and cerebral oximetry were thesia for patients’ (7). used together, cerebral oximetry accounted for the Similar stories can be told for other commonly used majority (58%) of monitoring events, with desatura- monitoring devices: The infant cardiorespiratory moni- tion defined as a 20% reduction from baseline NIRS tor has not been shown to decrease sudden infant measurements. Postoperative neurologic sequelae were death syndrome, the fetal heart rate monitor has not detected in 26% of patients who had any monitoring been shown to change the incidence of cerebral palsy, event that did not provoke a response. Neurologic and the intracranial pressure monitor has not been sequelae were seen in 6% of patients who had moni- shown to decrease secondary neurologic injury toring events that were treated, similar to 7% of (6,23,24). As with pulse oximetry, there will not likely patients who had no monitoring event (17). be a study to conclusively demonstrate that the appli- Dent et al. performed MRI before and 9 days after cation of cerebral oximetry improves patient outcomes. Norwood procedure in 22 infants with hypoplastic left Nor will there be a study adequately demonstrating heart syndrome. Seventy-three per cent of patients had ‘lack of benefit’ to cause rejection of NIRS-based mon- new lesions, or extension of existing lesions, and this itoring. Part of the difficulty in demonstrating the was associated with cerebral oximetry readings <45% value of a monitoring device is that monitoring results for >180 min (18). McQuillen et al. performed pre- can prompt a variety of interpretations and clinical and postoperative MRI in 53 neonates with congenital responses (17). It becomes practically prohibitive to heart disease requiring surgery. Overall brain injury recruit a sample size sufficient to account for this vari- was seen in 56% of patients, and low cerebral oxime- ability. Given this limitation, it is more instructive to try during cardiopulmonary bypass was associated examine the effect of interventions prompted by moni- with new lesions (19). Phelps et al. studied 50 neonates toring technologies, than to study deployment of the with hypoplastic left heart syndrome postoperative to monitor itself. Norwood stage I repair and found that length of ICU stay, need for ECMO, and death as a combined out- Goal-directed therapy and NIRS come were associated with 48-h mean cerebral oxime- try measurements (20). Cerebral oximetry provides a target for support of A more recent study of 67 infants requiring cardio- oxygen delivery to the brain that is similar to strategies pulmonary bypass for heart surgery that included pre- for shock treatment, collectively termed ‘goal-directed and postoperative MRI found no association between therapy’ (2,25). Therapies promoted by NIRS monitor- cerebral desaturation and new neurologic injuries (21). ing of the brain are best understood by examining the However, this study was performed at a center that determinants of oxygen delivery to the brain. For most has embraced a ‘goal-directed therapy’ with strategies organs, oxygen delivery is a function of cardiac output of bypass and anesthetic care to optimize cerebral and arterial oxygen saturation. The uneven application oximetry (22). The relatively low incidence of neuro- of systemic vasoconstriction makes the brain relatively logic injury in the study population (36%) cannot be immune to decrements in cardiac output, so cerebral compared to strategies without these adaptations in blood flow is a function of cerebral perfusion pressure, the absence of randomization. not cardiac output (26). Oxygen delivery to the brain CPPr4½Hb %Sat None of this evidence meets ‘level 1’ requirements to can thus be expressed as: O2Del / g , prompt a standard of care, because it is observational, where CPP is cerebral perfusion pressure, r is the theo- unrandomized, and does not specifically measure the retical collective arterial resistance vessel radius, [Hb] impact of cerebral oximetry monitoring on patient out- is the arterial hemoglobin concentration, % Sat is the come, which would require an inconceivably blinded percentage oxygen saturation of arterial hemoglobin, randomization to include or exclude cerebral oximetry and g is blood viscosity. Table 1 shows examples of

Pediatric Anesthesia 21 (2011) 473–478 ª 2011 Blackwell Publishing Ltd 475 Cerebral oximetry for pediatric anesthesia N. Kasman and K. Brady

Table 1 Goal-directed therapy guided by NIRS

Variable Clinical scenario causing cerebral desaturation Intervention

O2 delivery CPP Hypotension, elevated ICP, elevated central Maintain ABP above the lower limit of autoregulation. venous pressure Check venous drainage cannula r4 Hypocarbia, vasospasm, malpositioned Decrease minute ventilation, pH-stat management. arterial cannulae Check aortic cannula position [Hb] Anemia Transfusion

% Sat Cyanosis Lung recruitment maneuvers, increase FiO2, manage

QP/QS ratio g Polycythemia, sickle-cell disease Partial exchange transfusion, permissive anemia

O2 consumption Fever, seizure, arousal Cooling, sedation

CPP, cerebral perfusion pressure; ICP, intracranial pressure; ABP, arterial blood pressure; r, resistance vessel radius; [Hb], blood concentra- tion of hemoglobin; % Sat, arterial oxygen saturation; g, blood viscosity. how the understanding of oxygen delivery and con- ment that such data are unlikely to be obtained with- sumption can prompt interventions when cerebral out a radical breakthrough in study design methods. oximetry is low. Intelligent clinicians have room to disagree on the util- Goal-directed therapy guided by cerebral oximetry ity of cerebral oximetry monitoring in anesthesiology monitoring is expected to result in higher pump flow practice. Those against routine use of cerebral oxime- rates during bypass, higher CO2, a tendency for pH- try in their clinical practice point out the burden of stat management, and more blood transfusions (27). increased patient care costs without level I evidence of Further, centers that favor the use of cerebral oximetry clinical benefit. Adopters of NIRS-based cerebral mon- are more likely to adopt regional cerebral perfusion itoring are not likely to abandon their practice, even techniques as an alternative to deep hypothermic circu- understanding the limitations of the supporting data, latory arrest (22,28). These shifts in practice have been but this position is consistent with our use of other, shaped, in part, by the inclusion of cerebral oximetry more familiar monitoring devices. monitoring in the congenital cardiac operating suite The Brain Trauma Foundation faced a similar over the last decade as part of a more global effort at conundrum when declaring standards for the use of goal-directed therapy specific to pediatric cardiopulmo- the intracranial pressure monitoring devices. Despite a nary bypass. The impact of these changes on neuro- lack of rigorous objective data to demonstrate that the logic outcome is the subject of ongoing trials that intracranial pressure monitor changes clinical out- require long-term neurologic follow-up. comes, the foundation made a level II recommendation for use of the monitor in cases of severe traumatic brain injury. Further, the foundation indicated in the Conclusion guideline that efforts to study this question are not NIRS-based cerebral oximetry is a noninvasive and likely to be successful and therefore made a relatively easily applied monitoring technology that has been strong recommendation without this evidence (6). shown to trend the oxygen saturation of hemoglobin No drug becomes a recommended therapy without in the frontal cortex. Based on physiologic principles evidence of both safety and efficacy. Why are monitors of oxygen delivery and utilization in the brain, we have adopted, and even made standard of care absent this outlined and discussed the impact that cerebral oxime- requirement? The answer seems to be more practical try monitoring is having on the care of children with than academic. Monitors prompt interventions and congenital heart disease. We have presented evidence can impact the cost of patient care. It is not the moni- that the cerebral oximeter can detect clinically signifi- tor itself that changes the patient outcome, but the cant desaturation events that are related to acute brain interventions that result from application of the moni- injury during congenital heart surgery, but similar tor. There has not yet been a study design that can events in other clinical scenarios for adults, children, treat a monitoring device as a standard intervention and infants remain unclear. The limited data that we and measure its effect within constraints of a reason- have reviewed in this manuscript do not demonstrate able sample size. A study of pulse oximetry, for conclusively that the application of cerebral oximetry instance, might show benefit if it prompts low-stretch changes patient outcomes, and we have made the argu- and high end-expiratory pressure ventilation strategies

476 Pediatric Anesthesia 21 (2011) 473–478 ª 2011 Blackwell Publishing Ltd N. Kasman and K. Brady Cerebral oximetry for pediatric anesthesia in patients with acute respiratory distress syndrome pediatric cardiac care centers, and the preponderance of (ARDS). However, if the same monitor is studied centers which have adopted the monitor are unlikely to without a concomitant understanding of the mecha- randomize patients to not receive NIRS monitoring. It nism of ventilator-associated lung injury, then it is would be more productive to examine the changes in likely to show significant morbidity. While the pulse care introduced by NIRS monitoring than to suggest oximeter can serve as a guide to ventilator manage- yet another expensive, underpowered, and inevitably ment, it is the ventilator management strategy, and not inconclusive study of cerebral oximetry monitoring the pulse oximeter, that ultimately impacts patient out- devices. come. The application of cerebral oximetry is similarly less relevant than the interventions prompted by cere- Disclosures bral oximetry monitoring in relation to outcome. We are not prepared to stop using the pulse oximeter Dr. Brady has consulted for Somanetics, Inc. in a rela- to guide care for patients with critical respiratory illness. tionship that was disclosed and managed by the com- However, examination of practice patterns guided by mittee for outside interests at The Johns Hopkins that monitor is instructive long after it became standard University School of Medicine. This relationship has of care. Cerebral oximetry monitoring has dichotomized also been disclosed to the Baylor College of Medicine.

References 1 Brazy JE, Lewis DV, Mitnick MH et al. 10 Rais-Bahrami K, Rivera O, Short BL. Vali- 17 Austin EH 3rd, Edmonds HL Jr, Auden Noninvasive monitoring of cerebral oxygen- dation of a noninvasive neonatal optical SM et al. Benefit of neurophysiologic moni- ation in preterm infants: preliminary obser- cerebral oximeter in veno-venous ECMO toring for pediatric cardiac surgery. J vations. Pediatrics 1985; 75(2): 217–225. patients with a cephalad catheter. J Perina- Thorac Cardiovasc Surg 1997; 114(5): 707– 2 Tweddell JS, Ghanayem NS, Hoffman GM. tol 2006; 26(10): 628–635. 15, 17; discussion 15–6. Pro: NIRS is ‘‘standard of care’’ for postop- 11 Yoxall CW, Weindling AM, Dawani NH 18 Dent CL, Spaeth JP, Jones BV et al. Brain erative management. Semin Thorac Cardio- et al. Measurement of cerebral venous oxyhe- magnetic resonance imaging abnormalities vasc Surg Pediatr Card Surg Annu 2010; moglobin saturation in children by near-infra- after the Norwood procedure using regional 13(1): 44–50. red spectroscopy and partial jugular venous cerebral perfusion. J Thorac Cardiovasc 3 Hirsch JC, Charpie JR, Ohye RG et al. occlusion. Pediatr Res 1995; 38(3): 319–323. Surg 2006; 131(1): 190–197. Near infrared spectroscopy (NIRS) should 12 Abdul-Khaliq H, Troitzsch D, Schubert S 19 McQuillen PS, Barkovich AJ, Hamrick SE not be standard of care for postoperative et al. Cerebral oxygen monitoring during et al. Temporal and anatomic risk profile of management. Semin Thorac Cardiovasc neonatal cardiopulmonary bypass and deep brain injury with neonatal repair of congeni- Surg Pediatr Card Surg Annu 2010; 13(1): hypothermic circulatory arrest. Thorac Car- tal heart defects. Stroke 2007; 38(suppl 2): 51–54. diovasc Surg 2002; 50(2): 77–81. 736–741. 4 McGuire W. Parachute approach to evi- 13 Daubeney PE, Pilkington SN, Janke E et al. 20 Phelps HM, Mahle WT, Kim D et al. Post- dence based medicine: arguments are easily Cerebral oxygenation measured by near- operative cerebral oxygenation in hypoplas- refuted. BMJ 2006; 333(7572): 807; discus- infrared spectroscopy: comparison with jug- tic left heart syndrome after the Norwood sion -8. ular bulb oximetry. Ann Thorac Surg 1996; procedure. Ann Thorac Surg 2009; 87(5): 5 Potts M, Prata N, Walsh J et al. Parachute 61(3): 930–934. 1490–1494. approach to evidence based medicine. BMJ 14 Abdul-Khaliq H, Troitzsch D, Berger F 21 Andropoulos DB, Hunter JV, Nelson DP 2006; 333(7570): 701–703. et al. Regional transcranial oximetry with et al. Brain immaturity is associated with 6 Watzman HM, Kurth CD, Montenegro LM near infrared spectroscopy (NIRS) in com- brain injury before and after neonatal car- et al. Arterial and venous contributions to parison with measuring oxygen saturation in diac surgery with high-flow bypass and cere- near-infrared cerebral oximetry. Anesthesiol- the jugular bulb in infants and children for bral oxygenation monitoring. J Thorac ogy 2000; 93(4): 947–953. monitoring cerebral oxygenation. Biomed Cardiovasc Surg 2010; 139(3): 543–556. 7 Pedersen T, Moller AM, Hovhannisyan K. Tech 2000; 45(11): 328–332. 22 Andropoulos DB, Stayer SA, McKenzie Pulse oximetry for perioperative monitoring. 15 Nollert G, Jonas RA, Reichart B. Optimiz- ED et al. Novel cerebral physiologic moni- Cochrane Database Syst Rev 2009; (4): ing cerebral oxygenation during cardiac sur- toring to guide low-flow cerebral perfusion CD002013. gery: a review of experimental and clinical during neonatal aortic arch reconstruction. 8 Natale JE, Joseph JG, Pretzlaff RK et al. investigations with near infrared spectropho- J Thorac Cardiovasc Surg 2003; 125(3): Clinical trials in pediatric traumatic brain tometry. Thorac Cardiovasc Surg 2000; 491–499. injury: unique challenges and potential 48(4): 247–253. 23 American College of O, Gynecologists. responses. Dev Neurosci 2006; 28(4–5): 16 Hirsch JC, Charpie JR, Ohye RG et al. ACOG Practice Bulletin. Clinical Manage- 276–290. Near-infrared spectroscopy: what we know ment Guidelines for Obstetrician-Gynecolo- 9 Jobsis FF. Non-invasive, infra-red monitor- and what we need to know – a systematic gists, Number 70, December 2005 (Replaces ing of cerebral O2 sufficiency, bloodvolume, review of the congenital heart disease litera- Practice Bulletin Number 62, May 2005). HbO2-Hb shifts and bloodflow. Acta Neurol ture. J Thorac Cardiovasc Surg 2009; 137(1): Intrapartum fetal heart rate monitoring. Scand Suppl 1977; 64: 452–453. 154–159, 9e1–12. Obstet Gynecol 2005; 106(6): 1453–1460.

Pediatric Anesthesia 21 (2011) 473–478 ª 2011 Blackwell Publishing Ltd 477 Cerebral oximetry for pediatric anesthesia N. Kasman and K. Brady

24 Committee on F, Newborn. American 26 Schwartz AE, Sandhu AA, Kaplon RJ heart surgery. Anesth Analg 2004; 99(5): Academy of P. Apnea, sudden infant death et al. Cerebral blood flow is determined 1365–1375; table of contents. syndrome, and home monitoring. Pediatrics by arterial pressure and not cardiopulmo- 28 Ghanayem NS, Hoffman GM, Mussatto 2003; 111(4 Pt 1): 914–917. nary bypass flow rate. Ann Thorac KA et al. Perioperative monitoring in 25 Rivers E, Nguyen B, Havstad S et al. Early Surg 1995; 60(1): 165–169; discussion 9– high-risk infants after stage 1 palliation of goal-directed therapy in the treatment of 70. univentricular congenital heart disease. J severe sepsis and septic shock. N Engl J 27 Andropoulos DB, Stayer SA, Diaz LK et al. Thorac Cardiovasc Surg 2010; 140(4): 857– Med 2001; 345(19): 1368–1377. Neurological monitoring for congenital 863.

478 Pediatric Anesthesia 21 (2011) 473–478 ª 2011 Blackwell Publishing Ltd Pediatric Anesthesia ISSN 1155-5645

REVIEW ARTICLE Mortality as an outcome measure following cardiac surgery for congenital heart disease in the current era Ravi R. Thiagarajan1,2 & Peter C. Laussen1,2

1 Cardiac Intensive Care Unit, Department of Cardiology, Children’s Hospital Boston, Boston, MA, USA 2 Department of Pediatrics, Harvard Medical School, Boston, MA, USA

Keywords Summary cardiac surgery; congenital heart surgery; outcomes; mortality Mortality following cardiac surgery for congenital heart disease is rare in the current times. In this review article, we explore current mortality rates, Correspondence factors associated with mortality, and pitfalls in the use of mortality as a Ravi R. Thiagarajan, measure for assessing outcomes following congenital heart surgery. Associate Professor of Pediatrics, Harvard Medical School and Senior Associate in Cardiology, Children’s Hospital Boston, 300, Longwood Avenue, Boston, MA 02115, USA Email: [email protected]

Section Editor : Chandra Ramamoorthy

Accepted 7 March 2011 doi:10.1111/j.1460-9592.2011.03580.x

current mortality rates and factors associated with Introduction mortality for children undergoing cardiac surgery in Survival for children with congenital heart disease the current era. We will also discuss issues regarding requiring operative correction has improved over the the interpretation and current relevance of mortality as last decade because of advances in diagnostic modali- an outcome measure when comparing institutional ties, surgical and cardiopulmonary bypass support performance for children undergoing cardiac surgery techniques, and postoperative management (1). for congenital heart disease. Although mortality is an important outcome, current low rates across institutions make it difficult to use Mortality following surgery for congenital heart mortality as a measure of performance for quality disease – where are we today? improvement projects and as an outcome measure for research protocols aimed at developing new manage- Admissions and mortality rates from 1992 to 2009 in ment strategies (2). Decreasing mortality has lead to the Cardiac Intensive Care Unit at Children’s Hospi- the increased use of in-hospital morbidity as an out- tal, Boston, are shown in Figure 1 as an example of come measure following cardiac surgery for congenital improving survival for children with heart disease heart disease (3). Morbidity acquired during hospital- requiring intensive care. The mortality rate shown in ization after cardiac surgery may compromise long- the figure reflects the combined death rates for patients term functional status, decrease quality of life, and after cardiac surgery (currently 1.8%) and for critically result in continued resource utilization. Thus, critical ill patients with congenital or acquired heart disease evaluation of morbidity acquired during hospitaliza- who have not undergone surgery (currently 4.2%). tion after congenital cardiac surgery is essential to Current published estimates of mortality rates for chil- compare performance, improve quality of care, and dren and infants undergoing cardiac surgery vary from optimize resource utilization. In this review, we explore 3.7% to 4.3% (2,4,5). Variation in the reported rates

604 Pediatric Anesthesia 21 (2011) 604–608 ª 2011 Blackwell Publishing Ltd R.R. Thiagarajan and P.C. Laussen Cardiac surgical outcomes

CICU admissions and mortality 1992–2009 1600 6.0% Total admissions % Mortality 1400 5.0%

1200

4.0% 1000

800 3.0%

600 2.0% No. of admissions 400

1.0% 200

0 0.0% 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

Figure 1 Cardiac Intensive Care Unit, Children’s Hospital Boston, surgical and nonsurgical admissions and mortality during 1992–2009. may be as a result of the data source (single institution procedures performed (case-mix), comparison of vs multi-institutional data), type of data set (adminis- mortality rates between institutions requires the use of trative or cardiac surgery-specific data sets), duration procedural complexity-based ‘risk adjustment’ methods of data collection, and definition of the timing of to adjust for differences in case-mix among institutions mortality (30-day mortality vs survival to hospital to make interinstitutional comparisons valid. Proce- discharge). dural complexity is thus the basis of the currently used Many advances in imaging technology, surgical tech- risk stratification systems such as Risk Adjustment in niques, cardiopulmonary bypass, postoperative care, Congenital Heart Surgery-1 (RACHS-1) and Aristotle and use of extracorporeal membrane oxygenation to methods (8,9). manage refractory cardiorespiratory failure and aid Other risk factors for mortality following congenital cardiopulmonary resuscitation have all contributed to heart surgery include neonatal age group, particularly reduced mortality rates (1,6,7). Survival following all those with birthweight <2.5 kg and/or prematurity, types of congenital heart surgery should be expected major noncardiac structural anomalies and genetic or despite the increasing complexity of surgical proce- nongenetic syndromes, the need for re-operation for a dures performed. However, it is important to note the residual cardiac defect, and complications that arise relative plateau in mortality rate over recent years during the postoperative period (Table 1) (8,10–12). In (Figure 1), which begs the question as to whether a single institutional analysis of causes of death in chil- ‘zero’ mortality is attainable. Several patient and non- dren undergoing congenital heart surgery, Ma et al. medical factors make the goal of ‘zero’ mortality (12) identified that patients who died were commonly elusive.

Table 1 Factors associated with mortality after congenital heart Patient factors that increase mortality in children surgery undergoing cardiac surgery Patient factors Nonpatient factors An important determinant of mortality for children undergoing cardiac surgery for congenital heart disease Prematurity Center volume Neonates Surgeon experience is the complexity of a surgical procedure undertaken. and volume Mortality increases as complexity of the surgical proce- Birthweight <2.5 kg Human factors dure performed increases (8,9). Mortality rates for Noncardiac structural anomalies Cognitive issues individual cardiac procedures based on type and proce- Genetic and nongenetic syndromes Diagnostic errors dural complexity have been elegantly compiled by Procedural complexity Technical performance O’Brien et al. (4) and range from 0% to 40%. Because Perioperative complications Provider fatigue mortality rates vary widely depending on procedural Staffing and bed occupancy Communication complexity, and institutions vary widely in the type of

Pediatric Anesthesia 21 (2011) 604–608 ª 2011 Blackwell Publishing Ltd 605 Cardiac surgical outcomes R.R. Thiagarajan and P.C. Laussen neonates and infants <1 year of age, premature new- ences in case-mix among the centers studied. Another borns, and children with a diagnosis of single ventricle report by Welke et al. (2) using data from 32 413 disease. Complications during postoperative care such patients from 48 institutions in the Society of Thoracic as need for cardiopulmonary resuscitation and need Surgeons Congenital Heart Surgery Database reported for extracorporeal membrane oxygenation support more recently (2002 through 2006) demonstrated that were also common occurrences in children who died in center volume was inversely related to mortality rate this study. when all procedures were compared together. Mortal- The previously mentioned advances in the field of ity was shown to be considerably lower in large pediatric cardiac surgery that have improved mortality volume centers performing ‡350 cases per year for have also allowed for complex corrective operations in complex procedures (Aristotle difficulty >3.0) neonates with critical heart disease. As a group, neo- compared to centers performing <350 cases per year nates have a higher mortality risk because of their size (14.8% vs 8.4%). However, for low-complexity proce- and increased vulnerability owing to immaturity of dures (Aristotle difficulty £3.0), there was no difference organ systems (13–15). In a report of 30-day mortality in mortality among centers. Thus, it is possible that in 14 843 neonates undergoing cardiac surgery from increased experience and improved perioperative care the European Association for in low volume centers may have largely neutralized the database, the mortality rate was reported as 9.1% (15). effect of institutional case volume on mortality for In these patients, lower body weight, higher-complexity less-complex surgical operations. The influence of high surgical procedures, longer support times, and single level of scrutiny from regulatory agencies, public ventricle diagnosis were associated with increased mor- reporting of center-specific survival rates, and tality. Similarly, Dorfman et al. (10), in a single-center consumer expectation of improved performance for report of a group of neonates with critical heart dis- low volume centers may have led to improved out- ease undergoing cardiac surgery report a hospital mor- comes. However, the significance and influence of these tality rate of 7%, also higher compared to the average issues on improved outcomes in low volume centers mortality rate of all children undergoing cardiac sur- are difficult to estimate. gery for congenital heart disease. Other nonmedical, Whether an upper threshold exists in high volume but patient-related factors that may influence mortality centers beyond which operative mortality may actually include race and insurance status (16). increase is another area for evaluation in the future. Increasing case volume can exceed the resources avail- able in a system, resulting in decreased quality of care Nonpatient factors that increase mortality from overscheduling, need for rapid turnover resulting Along with the patient-related factors discussed earlier, in increased re-admission rates, and increased errors several other factors such as institution type, surgical related to human factors such as fatigue and poor volume, and surgeon experience may influence survival communication. Defining the threshold beyond which in children undergoing cardiac surgery (Table 1) quality of care may decrease in high volume centers (17,18). The relative importance of these nonpatient may be important in guiding future regionalization, factors compared to patient factors is not clearly more appropriate distributing work load, improving known and has been the subject of many recent inqui- quality of care, and sustaining reduction in the mortal- ries. Institutional volume and surgeon experience have ity rate for congenital heart surgery in the future. been extensively examined in several studies as possible In addition to volume and center characteristics, factors influencing mortality following surgery for technical performance of the surgeon and adequacy of congenital heart disease. In the mid-1990s, Jenkins repair after an operation have also been shown to et al. and Hannan et al. demonstrated that a volume influence mortality for children undergoing complex of cases performed at a center and surgeon experience cardiac surgery (20–22). In a recent report, Karamich- influenced mortality after congenital heart surgery. alis et al. (22) assessed the influence of a technically Since then, several other investigators have evaluated optimal operation compared to a technically adequate the volume and mortality relationship in children or inadequate operation on mortality following the undergoing cardiac surgery. Using data from the stage 1 palliation of hypoplastic left heart syndrome. National Inpatient Sample (NIS), a multi-institutional They showed that a technically optimal procedure data set containing data on 55 164 surgical procedures neutralized the effect of poor preoperative physiologi- during 1988 through 2005, Welke et al. (5,19) showed cal status and contributed to reduced hospital mortal- that large volume centers (>200 cases per year) out- ity. In another report, Karamlou et al. evaluated the performed all other centers after adjusting for differ- influence of institutional and surgeon experience,

606 Pediatric Anesthesia 21 (2011) 604–608 ª 2011 Blackwell Publishing Ltd R.R. Thiagarajan and P.C. Laussen Cardiac surgical outcomes patient management, and patient demographic factors centers. Adding to these factors, low morality rates on mortality following congenital heart surgery for make comparison of institutional performance based four groups of neonates undergoing palliative or repar- on mortality difficult. ative procedures for a transposition of the great arter- Welke et al. (2), in a 2010 report, using data from ies, pulmonary atresia with intact ventricular septum, 21 709 operations from the NIS data set demonstrated interrupted aortic arch, or the Norwood operation that pediatric surgical volume in centers performing (23). They demonstrated that for some procedures cardiac surgery was too small to effectively show (Norwood and pulmonary atresia with ventricular differences in mortality rate for assessing institutional septal defect), institution or surgeon experience did not performance. Furthermore, the authors correctly influence mortality; rather patient and postoperative conclude that the absence of differences based on low management issues were important. However, in the volume may falsely reassure small-volume centers that group of patients undergoing repair of transposition of their performance is no different than the mean. the great arteries, institution and surgeon experience Because of the many limitations of mortality as an rather than patient factors influenced mortality. Thus, outcome measure as outlined here, the use of other the influence of institution and surgeon factors on morbid events that occur with higher frequency, such mortality may vary based on the procedure and may as postoperative complications, neurological outcomes, not affect outcomes for all surgical procedures. and length of hospitalization to better reflect institu- Furthermore, this study also demonstrates that in tional performance should be considered in the future some cases mortality may be strongly influenced by (24). These morbid events may decrease functional patient and postoperative management factors rather status and quality of life for those surviving to hospital than by technical outcomes of the procedure itself. discharge after congenital heart surgery and thus also Although research has focused on surgeon perfor- influence long-term outcomes and resource utilization. mance, the contribution of the performance of other Future research directed at studying the use of morbid caregivers including anesthesiology, cardiology, and events as an outcome measure following congenital intensive care teams toward errors in diagnosis, deci- heart surgery may help better define an optimal sion making, and communication has not been evalu- measure that is relevant and useful for designing inter- ated and should be the focus of future research. ventions that can help promote quality of care as well as long-term outcomes. Mortality as an indicator of quality of cardiac surgical programs Summary In-hospital mortality is frequently cited as a measure Mortality for children undergoing cardiac surgery is to compare the performance of cardiac surgical low in the current era. Neonates, children with major programs for purposes of research, as well as by noncardiac structural anomalies, and children undergo- external agencies and regulatory bodies (2). The use of ing high-complexity cardiac surgical procedures may mortality as a discriminator of quality stems from the have increased mortality. Studies aimed at improving adult cardiac surgical experience where mortality has mortality for cardiac surgery in children with heart dis- been shown to accurately reflect quality. This is ease should potentially target these populations. directly related to the availability of a large volume of Although it is necessary to continue to monitor mor- adult cases for comparison, the relative uniformity of tality in children undergoing cardiac surgery, low mor- adult cardiac surgical procedures, and the availability tality rates and need for improved quality of life and of robust risk adjustment systems. In comparison, functional outcomes in survivors demand that we fewer children undergo cardiac surgery, and center vol- define other relevant outcomes measures for the future. ume and case complexity can vary widely between

References 1 Wessel DL, Laussen PC. Intensive care unit. of quality differences between pediatric car- 4 O’Brien SM, Clarke DR, Jacobs JP et al. In: Keane JF, Lock JE, Fyler DC, eds. diac surgical programs. Ann Thorac Surg An empirically based tool for analyzing Nadas’ Pediatric Cardiology, 2nd edn. 2010; 89: 139–144; discussion 145–136. mortality associated with congenital heart Philadelphia, PA: Saunders Elsevier, 2006: 3 Jacobs JP, Wernovsky G, Elliott MJ. Analy- surgery. J Thorac Cardiovasc Surg 2009; 303–323. sis of outcomes for congenital cardiac 138: 1139–1153. 2 Welke KF, Karamlou T, Ungerleider RM disease: can we do better? Cardiol Young 5 Welke KF, O’Brien SM, Peterson ED et et al. Mortality rate is not a valid indicator 2007; 17(Suppl 2): 145–158. al. The complex relationship between pedi-

Pediatric Anesthesia 21 (2011) 604–608 ª 2011 Blackwell Publishing Ltd 607 Cardiac surgical outcomes R.R. Thiagarajan and P.C. Laussen

atric cardiac surgical case volumes and 13 Costello JM, Polito A, Brown DW et al. 2005. Ann Thorac Surg 2008; 86: 889–896; mortality rates in a national clinical data- Birth before 39 weeks’ gestation is associ- discussion 889–896. base. J Thorac Cardiovasc Surg 2009; 137: ated with worse outcomes in neonates with 20 Bacha EA, Larrazabal LA, Pigula FA et al. 1133–1140. heart disease. Pediatrics 2010; 126: 277–284. Measurement of technical performance in 6 Kolovos NS, Bratton SL, Moler FW et al. 14 Curzon CL, Milford-Beland S, Li JS et al. surgery for congenital heart disease: the Outcome of pediatric patients treated with Cardiac surgery in infants with low birth stage I Norwood procedure. J Thorac extracorporeal life support after cardiac weight is associated with increased mortal- Cardiovasc Surg 2008; 136: 993–997, 997 surgery. Ann Thorac Surg 2003; 76: 1435– ity: analysis of the Society of Thoracic e991–992. 1441; discussion 1441–1432. Surgeons Congenital Heart Database. J 21 Larrazabal LA, del Nido PJ, Jenkins KJ 7 Thiagarajan RR, Laussen PC, Rycus PT Thorac Cardiovasc Surg 2008; 135: 546– et al. Measurement of technical performance et al. Extracorporeal membrane oxygenation 551. in congenital heart surgery: a pilot study. to aid cardiopulmonary resuscitation in 15 Kansy A, Tobota Z, Maruszewski P et al. Ann Thorac Surg 2007; 83: 179–184. infants and children. Circulation 2007; 116: Analysis of 14,843 neonatal congenital heart 22 Karamichalis JM, Thiagarajan RR, Liu H 1693–1700. surgical procedures in the European Associ- et al. Stage I Norwood: optimal technical 8 Jenkins KJ, Gauvreau K, Newburger JW ation for Cardiothoracic Surgery Congenital performance improves outcomes irrespective et al. Consensus-based method for risk Database. Ann Thorac Surg 2010; 89: 1255– of preoperative physiologic status or case adjustment for surgery for congenital heart 1259. complexity. J Thorac Cardiovasc Surg 2010; disease. J Thorac Cardiovasc Surg 2002; 123: 16 Benavidez OJ, Gauvreau K, Jenkins KJ. 139: 962–968. 110–118. Racial and ethnic disparities in mortality 23 Karamlou T, McCrindle BW, Blackstone 9 Lacour-Gayet F, Clarke D, Jacobs J et al. following congenital heart surgery. Pediatr EH et al. Lesion-specific outcomes in The Aristotle score: a complexity-adjusted Cardiol 2006; 27: 321–328. neonates undergoing congenital heart method to evaluate surgical results. Eur J 17 Hannan EL, Racz M, Kavey RE et al. Pedi- surgery are related predominantly to patient Cardiothorac Surg 2004; 25: 911–924. atric cardiac surgery: the effect of hospital and management factors rather than 10 Dorfman AT, Marino BS, Wernovsky G and surgeon volume on in-hospital mortal- institution or surgeon experience: A et al. Critical heart disease in the neonate: ity. Pediatrics 1998; 101: 963–969. Congenital Heart Surgeons Society Study. presentation and outcome at a tertiary care 18 Jenkins KJ, Newburger JW, Lock JE et al. J Thorac Cardiovasc Surg 2010; 139: center. Pediatr Crit Care Med 2008; 9: 193– In-hospital mortality for surgical repair of 569–577 e561. 202. congenital heart defects: preliminary obser- 24 Newburger JW, Wypij D, Bellinger DC 11 Benavidez OJ, Gauvreau K, Del Nido P et al. vations of variation by hospital caseload. et al. Length of stay after infant heart Complications and risk factors for mortality Pediatrics 1995; 95: 323–330. surgery is related to cognitive outcome at during congenital heart surgery admissions. 19 Welke KF, Diggs BS, Karamlou T et al. age 8 years. J Pediatr 2003; 143: 67–73. Ann Thorac Surg 2007; 84: 147–155. The relationship between hospital surgical 12 Ma M, Gauvreau K, Allan CK et al. Causes case volumes and mortality rates in pediat- of death after congenital heart surgery. Ann ric cardiac surgery: a national sample, 1988– Thorac Surg 2007; 83: 1438–1445.

608 Pediatric Anesthesia 21 (2011) 604–608 ª 2011 Blackwell Publishing Ltd