Stage I palliation

In 1979, Norwood et al. reported a neonatal palliative operation for HLHS that stabilized native HLHS physiology and eliminated ductal dependence. He later described a Fontan palliation as the second stage operation, resulting in the first long‐term survivors of HLHS [16]. The Norwood stage I palliation involves reconstruction of the diminutive ascending aorta, usually with a cryopreserved homograft patch, along with the creation of outflow from the systemic RV through the pulmonary valve, which also provides blood flow to the native aortic root and coronary arteries. The pulmonary valve becomes the “neo‐aortic valve” (Figure 30.3) [17, 18]. The main PA is transected, and the proximal portion of the PA is incorporated into the aortic repair while the underside of the PA bifurcation is oversewn. In addition, an atrial septectomy is performed to provide unrestricted left‐to‐right blood flow across the atrial septum. Pulmonary blood flow is then delivered via a polytetrafluoroethylene (PTFE) shunt, either via the creation of a systemic‐to‐pulmonary shunt from the right innominate or subclavian artery to right PA (also termed modified Blalock–Thomas‐Taussig shunt [mBTTS]) or by the creation of an RV–PA shunt (also termed the Sano shunt). It is crucial that the area of the anastomosis between the pulmonary valve and the aortic valve is free of obstruction, as it can affect the already tenuous coronary blood flow. The complete removal of any residual ductal tissue to prevent later coarctation of the aorta is also critical. During aortic arch reconstruction, cardiopulmonary bypass (CPB) techniques must provide a bloodless surgical field. Deep hypothermic circulatory arrest (DHCA) or antegrade regional cerebral perfusion (ACP) may be utilized for this purpose. Preference for a particular approach is largely surgeon and institution dependent. Details of each technique along with short‐ and long‐term outcomes are discussed in Chapter 11.

Systemic‐to‐pulmonary and RV‐to‐pulmonary shunts

The systemic‐to‐pulmonary artery shunt is a 3.0‐ to 4.0 mm polytetrafluoroethylene (PTFE) graft that provides continuous flow into the PA. Originally described as a modified “Blalock‐Taussig” shunt after surgeon Alfred Blalock and cardiologist Helen Taussig, more recently, it has become identified as a modified “Blalock‐Thomas‐Taussig shunt” to recognize the primary contribution of Vivien Thomas, a surgical technologist, to its successful development and implementation [19, 20]. Shunt flow is inversely proportional to the length of the shunt and the fourth power of its radius. It is directly proportional to the pressure drop across the shunt and therefore influenced by the ratio of PVR and systemic vascular resistance (SVR). Kinking or distortion of either the shunt or PA can occur and interfere with the flow. Because there is continuous “run‐off” of flow from the systemic circulation into the PA, these patients remain at risk for steal from the systemic circulation as PVR decreases. This is proposed as the major pathophysiological etiology for the frequent instability observed in the early postoperative period in these patients. In 2003, Sano et al. reported improvement in stage I survival in a small number of patients at one institution when pulmonary blood flow (PBF) was provided by a 5–6 mm PTFE graft from the RV to the PA, rather than via an mBTTS (Figure 30.4) [18, 21]. Because the RV–PA shunt arises proximal to the neo‐aortic valve, there is no flow from the aorta into the PAs during diastole, limiting diastolic run‐off and coronary ischemia. The resultant higher diastolic blood pressure has been hypothesized as the physiologic reason for improved survival.

The SVRT was designed to test the hypothesis that stage I survival would improve with the utilization of the RV–PA shunt, with an endpoint of death or transplantation in the first 12 months [7]. In all, 555 infants were randomized to receive either an mBTTS or RV–PA shunt. Approximately 10% of subjects required cross‐over from their randomly assigned shunt type to the other shunt type because of anatomic considerations. At 12 months, 36.4% of the mBTTS subjects had died or been transplanted, compared with 26.3% for the RV–PA shunt (P = 0.010). However, overall complication rates in the RV–PA shunt group were higher (72 vs. 46 per 100 infants, P = 0.03), and the number of infants experiencing one or more complications was also greater with the RV–PA shunt (39 vs. 28%, P = 0.03). After 12 months, the hazard ratio for death or transplantation did not differ between the two types of shunts. Follow‐up after 6 years demonstrated slightly higher transplant‐free survival (5%) but an increased need for catheter re‐interventions in the RV‐PA shunt patient group; however, there were no statistically significant differences in the hazard ratio for death or transplantation as well as catheter re‐interventions [22]. A recent meta‐analysis of 32 studies by Cao et al. assessed long‐term survival in HLHS with either mBTTS (n = 1348) or RV‐PA shunt (n = 1258). Although the RV‐PA shunt group had significantly higher in‐hospital survival and inter‐stage survival between stage I and II palliation until 2 years post‐stage I palliation, no significant differences in long‐term outcomes between stage II and stage III palliation were identified [23]. In contemporary practice, the shunt type used depends largely on the surgeon and institutional preference.

In a publication specifically examining the rate and causes of 30‐day in‐hospital mortality in SVRT subjects, 16% of patients died during the initial hospitalization, with 12% mortality in the first 30 days after Norwood stage I palliation. Although after multivariable analysis shunt type was not identified as a significant risk factor for mortality, the presence of an mBTTS was an independent risk factor for cardiopulmonary resuscitation (CPR). Independent risk factors for both 30‐day and total in‐hospital mortality included low birth weight, genetic abnormality, post‐procedure need for extracorporeal membrane oxygenation (ECMO), and open sternum. Longer duration of DHCA was an independent risk factor for hospital mortality. Independent risk factors for major morbidity (renal failure, sepsis, increased duration of ventilation, and hospital stay) included genetic abnormality, lower center/surgeon volume, open sternum, and additional post‐Norwood operations [24].

Inter‐stage mortality occurring after the initial hospitalization for stage I palliation and during the 3–6 months prior to Stage II superior cavopulmonary connection (SCPC) has been a persistent issue. The physiology of the HLHS patients remains tenuous after the initial palliation, as a parallel circulation with all systemic and pulmonary output passing through the single systemic ventricle still persists. The Qp: Qs remains labile, and thus, coronary ischemia continues to be a risk. Significant hemodynamic perturbations may easily result from dehydration due to viral gastroenteritis, upper respiratory infections, or micro‐aspiration episodes caused by a recurrent laryngeal nerve injury that frequently accompanies stage I palliation. In the SVRT, of 426 subjects discharged from the hospital, 50 died inter‐stage (12%) [25]. The inter‐stage death rate was greater for mBTTS patients (18%) than for RV–PA shunt patients (6%, P < 0.001). Risk factors for inter‐stage death included gestational age <37 weeks, Hispanic ethnicity, AA/MA, a greater number of post‐Norwood complications, census block poverty level, and moderate to severe AVVR. Ahmed et al. recently developed a validated risk scoring system to predict mortality or transplant in single‐ventricle patients during the inter‐stage period between stage I and II palliation. Between 2008 and 2015, data from over 2,000 patients from the National Pediatric Cardiology Quality Improvement Collaborative Registry were analyzed to create a weighted scoring system. This novel system, known as the NEONATE score, assessed the following independent risk factors: Norwood type (mBTTS or hybrid vs. Sano shunt), postoperative ECMO, opioids required at discharge, no digoxin required at discharge, arch obstruction, ≥ moderate Tricuspid regurgitation, and extra oxygen plus ≥ moderate tricuspid regurgitation at discharge. During the inter‐stage period, the strongest predictor of death or transplant was ≥ moderate tricuspid regurgitation, which worsened if oxygen was required [26]. With a combined 28% mortality observed in the SVRT in the inter‐stage period between stages I and II, the challenges of caring for this group of patients are evident. Death after stage II SCPC in the first year of life is less common; the stabilizing effect on the pathophysiology of HLHS (see later) is the most significant factor.

Preoperative considerations

The diagnosis of HLHS is most frequently made in utero. After an initial assessment, immediate treatment in the delivery room for HLHS patients includes initiation of a PGE₁ infusion, placement of umbilical arterial and venous catheters, maintenance of spontaneous ventilation on room air, and transfer to a neonatal or cardiac intensive care unit (ICU). A postnatal transthoracic echocardiogram is performed to confirm the initial diagnosis and provide other crucial anatomic information, including LV size, presence of AVVR, degree of atrial septal restriction, or the presence of anomalous pulmonary venous return. Other less common but important details such as the presence of coronary artery fistulae or sinusoids to the LV should also be assessed. Patients with a restrictive atrial septum may require urgent or emergent intervention, requiring either balloon atrial septostomy or early surgical intervention. Cardiac catheterization is rarely performed for diagnostic purposes, and additional imaging (i.e., computed tomography [CT] or magnetic resonance imaging [MRI]) is seldom required unless anatomic questions remain. Certain anatomic and physiologic features have been identified as risk factors that may affect outcomes (see Table 30.1).

The chest radiograph is followed daily to assess the heart size and status of the pulmonary vasculature. Common preoperative laboratory tests include complete blood count, coagulation studies, and blood type and cross‐match, along with the measurement of serum electrolytes, blood urea nitrogen, creatinine, and glucose. In a recent evaluation of standard preoperative laboratory testing for pediatric cardiac surgery patients in the US and Canada by the Congenital Cardiac Anesthesia Society Hemostasis Interest Group, 11 of 12 centers surveyed reported routinely obtaining preoperative testing. However, 83% agreed that only a type and screen and complete blood count were necessary. No centers reported changes in management due to preoperative testing, suggesting that preoperative testing requirements are controversial and subject to institutional preferences [54].

Ideally neonatal HLHS patients “self‐regulate” their Qp:Qs prior to stage I palliation, spontaneously breathing with a slightly elevated PaCO₂. Diuretics are often used preoperatively to limit blood volume recirculating through the lungs. Nasal continuous positive airway pressure (CPAP) may be helpful to provide positive pressure to relieve respiratory distress and elevate the PVR while maintaining spontaneous respiration. Tracheal intubation and mechanical ventilation are reserved for the most severe cases of respiratory insufficiency or for patients with anatomic variants (e.g., restrictive intra‐atrial communication) that markedly diminish PBF. Early surgical intervention before severe derangements in Qp:Qs ensue is essential to prevent complications.

Intraoperative considerations

Neonates presenting for stage I reconstruction (i.e., Norwood procedure with mBTTS or Sano shunt) receive an intravenous (IV) induction of anesthesia with meticulous attention to the hemodynamic consequences of the anesthetic agents selected. Phenylpiperidine‐based synthetic opioids, such as fentanyl, are generally preferred because they blunt the endogenous catecholamine response to noxious stimuli at doses that are usually tolerated hemodynamically [55–57]. However, even with these hemodynamically “neutral” agents, larger doses may result in significant cardiovascular changes, such as bradycardia and hypotension. These observations suggest that some endogenous catecholamine release is required in an infant with HLHS to sustain satisfactory hemodynamics. Unfortunately, the threshold narcotic dose that is “sufficient” and not “excessive” varies between patients, necessitating careful titration.

Although clinical research conducted in the early 1990s popularized the view that very high doses of opioid analgesics should be administered perioperatively for stress hormone suppression in neonates undergoing cardiac surgery [58], efforts to duplicate these findings have not confirmed the original results [59]. The latter demonstrated that very high doses of fentanyl did not completely suppress the release of endogenous catecholamines, even in combination with benzodiazepine infusions. Additionally, outcome measures were no different in any of the study groups. Emerging evidence has demonstrated potential benefits of the use of dexmedetomidine in pediatric cardiac surgery, including neuroprotection as well as improved morbidity and mortality [60]. According to data from the Congenital Cardiac Anesthesia Society and Society of Thoracic Surgeons Cardiac Anesthesia Database Collaboration, dexmedetomidine use has steadily increased since 2010 [61]. Strategies vary significantly between institutions, with total intraoperative fentanyl doses ranging between 10–100 μg/kg, and the addition of low‐dose volatile anesthetics, benzodiazepines, ketamine, and dexmedetomidine in various combinations described as components of a balanced anesthetic approach to initial palliations for HLHS [62]. Most institutions now utilize fentanyl dosages in the lower range, generally <25 μg/kg.

Management of the airway and ventilation assumes great importance during induction of anesthesia. Given the propensity for most neonates with HLHS to have excessive PBF, the anesthesiologist must take care not to employ ventilatory maneuvers that lower PVR, such as hyperventilation or utilization of high inspired concentrations of oxygen (FiO₂). In an infant with typical HLHS physiology, manual ventilation should be initiated with room air or a low concentration of supplemental oxygen. The extent to which FiO₂ is increased before laryngoscopy should depend upon the magnitude of hemodynamic response to the initiation of controlled ventilation. Infants who demonstrate a significant reduction in systemic arterial pressure despite low FiO₂ may not tolerate prolonged exposure to a higher FiO₂ without deleterious hemodynamic consequences.

Intraoperative monitoring consists of standard cardiovascular, respiratory, and temperature monitors and continuous direct arterial pressure monitoring. An umbilical arterial line, if present, may be used to preserve other arterial access sites for future interventions. Most centers utilize central venous access as well. An umbilical venous catheter positioned in the orifice of the superior vena cava (SVC) at the time of surgery serves as a valuable monitor of mixed venous oxygen saturation, enabling more precise assessment of systemic CO and Qp: Qs [6, 63, 64]. Some institutions may place direct transthoracic atrial lines instead of percutaneous jugular or subclavian central venous pressure catheters due to concern regarding thrombosis in the central thoracic veins in SV patients. However, other institutions have failed to demonstrate the significant risk associated with the use of upper body central venous catheters even in SV patients [65]. Peripherally inserted central catheters (PICCs) are increasingly utilized as well. In a 2020 retrospective review of PICC use in pre‐stage II SV patients, only 2 out of 56 patients with PICCs developed detectable thrombus (4%) versus 3 out of 17 patients with central venous catheters (18%) [66]. Most centers utilize cerebral and somatic near‐infrared spectroscopy (NIRS) as additional non‐invasive monitors which correlate with SvO₂. Cerebral NIRS can be used to guide ACP as it has been shown to detect alterations in CBF in neonates undergoing Norwood stage I palliation [67, 68]. The use of cerebral and somatic NIRS in the immediate postoperative period has also been shown to predict early mortality and to highlight the need for ECMO in HLHS patients [69]. Transesophageal echocardiography (TEE) is often utilized depending on the size of the neonate and institutional preference. Epicardial echocardiography after CPB can also be utilized and is particularly useful to assess AVVR and neo‐aortic regurgitation.

Measures to clear airway secretions and assure complete re‐expansion of the pulmonary parenchyma are performed in the terminal phases of rewarming on bypass. The magnitude of reduction in the mean systemic arterial pressure that occurs with the opening of the shunt during the terminal phase of CPB can provide qualitative insights into what PVR might be expected in the early post‐CPB period. Initial ventilatory support should be adjusted accordingly. It is best to begin with a FiO₂ between 0.6 and 1 and a pattern of ventilation designed to result in low‐normal PaCO₂, recognizing that further adjustments will be indicated by the individual infant’s physiology. Assuming a technically satisfactory repair without unusual risk factors, PVR typically falls within the first few hours after surgery. Until then, some patients may potentially require high levels of supplemental oxygen in the immediate post‐CPB period.

While the physiologic goals post‐CPB remain identical to those of the preoperative period, the variables affecting the achievement of these goals differ. The pulmonary circulation now resides either at the distal end of a restrictive prosthetic systemic‐to‐pulmonary shunt or an RV–PA shunt. A variety of subtleties, including technical execution of the shunt, relief of atrial obstruction, and the pre‐existing condition of the pulmonary circulation, can render the ultimate physiology unpredictable. When a Sano shunt has been performed, technical issues of graft insertion, proper graft size (particularly in the neonate <2.0 kg), and the ability to delineate the etiology of insufficient PBF at the termination of CPB can introduce additional complexities.

Even with a perfect technical result, the Norwood operation does not reduce the volume or pressure burden on the SV. Supporting and maintaining the physiology of parallel systemic and pulmonary circulations with a Qp: Qs of 1.0 remain the objective throughout the post‐CPB period. Postoperatively, the heart has incurred the cost of insults related to CPB, cessation of coronary perfusion, and DHCA, which may contribute to the cardiovascular frailty often exhibited in the early postoperative period. However, in the absence of major deficiencies in myocardial protection or persistent anatomic residua, such as significant aortic arch obstruction, coronary compromise, or valvular insufficiency, myocardial dysfunction can usually be addressed with relatively modest doses of inotropic agents and vasodilation. Low dose epinephrine (0.03–0.05 mcg/kg/min) and milrinone (0.25–0.5 mcg/kg/min) are commonly utilized after CPB. Vasopressin infusion may be added as well if necessary. While some clinicians advocate routine use of perioperative milrinone in Norwood procedures, others propose a targeted approach, citing the relatively weak inotropic and lusitropic effects of milrinone compared to its vasodilating properties. There is also conflicting evidence regarding the efficacy of perioperative milrinone in reducing mortality after the Norwood procedure [70–72]. In many neonates and young infants undergoing major cardiac interventions, myocardial performance may deteriorate in the first 6–12 hours postoperatively before it begins to improve. As a result, measures to reduce metabolic demands by providing continuous neuromuscular blockade and opioid infusions may be utilized in the immediate postoperative period. The sternum may be left open, covered by Silastic^TM or other synthetic material, and stabilized with a plastic strut. This may improve hemodynamic stability as mediastinal edema is frequent in the first several postoperative days. The potentially deleterious effects of positive pressure ventilation are also mitigated. The open sternum also provides access for postoperative mediastinal exploration for bleeding, or the institution of ECMO. Sternal closure can be accomplished at the bedside 24–72 hours postoperatively.

A variety of ultrafiltration techniques, including conventional, modified, zero‐balance, and combination ultrafiltration, may be utilized during and after CPB to improve postoperative hemodynamics. Several studies have demonstrated the beneficial effects of ultrafiltration upon hematocrit, hemodynamics, hemostasis, pulmonary function, and central nervous system recovery [73–78]. Perioperative weight gain is significantly reduced, as are the levels of specific inflammatory mediators. Occasionally, the position of the bypass cannulae or the continuous flux of blood through the modified ultrafiltration (MUF) circuit results in unfavorable hemodynamic changes precluding completion of the ultrafiltration. In 99 consecutive patients undergoing stage I palliation between September 2000 and August 2002, all tolerated the hemodynamic perturbations of MUF [79]. Clinical outcomes, however, have not consistently been improved and no single method of ultrafiltration has proven superior over others [73, 80]. Some institutions routinely place peritoneal dialysis (PD) catheters in the OR, and utilize PD in the early postoperative period for removal of excess interstitial fluid and potentially to remove inflammatory mediators.

Rapid restoration of normal hemostasis is an essential objective in the post‐CPB period, particularly as the creation of the neo‐aorta in Stage I reconstruction requires substantial arterial suture lines, and cyanotic heart disease is a predictor of postoperative blood loss [81]. Following the completion of MUF and confirmation of satisfactory technical and physiologic repair, heparin effect is reversed with protamine. Despite the theoretical advantages of fresh whole blood, studies have not demonstrated benefit from this strategy relative to the use of reconstituted whole blood [82]. Additionally, fresh whole blood is not available at most institutions. Gruenwald et al. demonstrated that a reconstituted fresh whole blood bypass prime, consisting of packed red blood cells (PRBCs), fresh frozen plasma, and platelets from the same donor, <7 days old, followed by a second such unit given after bypass, reduced bleeding in infants undergoing surgery compared with standard blood component therapy [83]. Should these measures fail to achieve adequate hemostasis despite the elimination of most surgical bleeding sites, laboratory testing (platelet count and function, fibrinogen level, and viscoelastic testing) should be conducted to direct component therapy.

Data specifically examining procoagulant medications in infants undergoing stage I reconstruction are sparse. Antifibrinolytic therapies have demonstrated efficacy in infants and children following heart surgery, but data in neonates continues to be very limited [84–87]. Although aprotinin, a serine protease inhibitor that acts as an antifibrinolytic agent with anti‐inflammatory properties, demonstrated effectiveness comparable to other antifibrinolytics in infants and children [88]; this agent was voluntarily removed from the market when it was linked to greater 5‐year mortality following coronary bypass grafting [89]. The lysine analogs epsilon aminocaproic acid and tranexamic acid appear to have comparable efficacy [84, 87], and extensive experience implying safety, making them appropriate substitutes for aprotinin [90].

Recombinant activated factor VII (rFVIIa) has been used as rescue therapy in infants and children following cardiac surgery, including neonates undergoing stage I reconstruction when conventional correction of postoperative coagulopathy with blood product transfusions fails to control hemorrhage [91, 92]. While rFVIIa can aid in controlling intractable bleeding, caution is warranted, as a recent retrospective review of 149 infants and children who received rFVIIa during cardiac surgery demonstrated an increased risk of thrombosis compared to control patients (20% vs. 8%) [93]. Another retrospective review of 414 neonates who underwent cardiac surgery found a 22% frequency of thrombosis in patients who received rFVIIa compared to a 6% incidence in patients who did not receive it. In neonates who underwent the Norwood procedure, those who received rFVIIa and developed thrombus had significantly higher mortality and incidence of requiring ECMO [94]. Recombinant factor VIIa may be capable of producing a hypercoagulable state that poses risks to certain surgical repair elements, including prosthetic shunts. Prothrombin complex concentrates (PCCs) may pose a similar thrombotic risk and should be used cautiously in patients undergoing stage I palliation. The 2019 patient blood management guidelines from the Network for the Advancement of Patient Blood Management, Haemostasis, and Thrombosis (NATA) for neonates and children undergoing cardiac surgery recommend against the use of PCCs in pediatric cardiac surgery except in clinical trials due to insufficient efficacy and safety data [95].

Point‐of‐care viscoelastic tests and transfusion algorithms are increasingly used in pediatric cardiac surgery [95–97]. Cyanotic patients, including those with HLHS, are known to have abnormal thromboelastography parameters compared to normal children, with these coagulation imbalances predisposing to both bleeding and thrombosis [96, 98]. The development of institution‐specific, target‐population transfusion algorithms may lead to reduced intraoperative transfusions and improved outcomes. However, additional data regarding patients undergoing stage I palliation is needed [96, 97, 99]. As PRBC transfusion has been associated with a longer duration of mechanical ventilation and ICU stay in stage I palliation patients [100], the Pediatric Critical Care Transfusion and Anemia Expertise Initiative released recommendations regarding red blood cell transfusion in patients with CHD. These recommendations state transfusions should be avoided “solely based upon hemoglobin, if hemoglobin is >9.0 g/dL in infants undergoing staged palliative procedures with stable hemodynamics” [101]. Further details of coagulation and bleeding management are discussed in Chapter 16.

Postoperative considerations

Excessive hypoxemia represents one of the more commonly encountered problems in the early post‐CPB period. Although inadequate Qp:Qs is often the cause, factors that impair systemic oxygen delivery and reduce mixed venous oxygen saturation can also occur [63, 64, 102, 103]. A progressive increase in systemic oxygen saturation is often observed during MUF, likely due to the effect of hemoconcentration and the resulting increased oxygen delivery on mixed venous oxygen saturation. Post‐CPB, measures directed at maintaining the hematocrit above 40–45% may alleviate excessive demands placed upon the recovering heart to increase systemic output. The distinction between systemic hypoxemia due to low Qp:Qs, low pulmonary venous oxygen saturation, or low mixed venous oxygen saturation (SvO₂) is crucial as the therapies for each differ. Measures designed to reduce PVR will impose a further volume load on a heart already struggling to provide marginal systemic perfusion. Patients demonstrating low SvO₂ are better served with therapies that promote systemic output, such as inotropic agents or vasodilators (see Figure 30.6) [104].

Similarly, infants with low pulmonary venous oxygen saturation require a strategy of ventilatory support designed to reduce atelectasis and promote gas exchange in impaired alveoli. Although pulmonary vein desaturation is not the most common cause of persistent systemic hypoxemia after stage I palliation, one series found pulmonary vein desaturation in as many as 30% of patients [105]. Unfortunately, this diagnosis is rarely made definitively in the OR or cardiac ICU due to the logistical challenges of blood sampling from the pulmonary veins. Intraoperatively, expectant measures directed at expansion of the lungs and maintenance of normal functional residual capacity (FRC) usually suffice to avoid pulmonary vein desaturation.

When systemic hypoxemia is suspected to be due to low Qp: Qs, other manifestations can provide supporting evidence. A trial opening of the systemic‐to‐pulmonary artery shunt during the latter phases of rewarming on CPB may have failed to demonstrate a significant drop in the mean systemic arterial pressure, and early post‐CPB hemodynamics may have revealed a relatively narrow pulse pressure and/or high diastolic pressure. A substantial discrepancy may exist between arterial and end‐tidal CO₂ measurements. These suggestive pieces of inferential evidence can be confirmed by aortic Doppler flow analysis or calculation of a Fick ratio using oxygen saturation determinations. Most commonly, diminished PBF may reflect a subtle technical aspect of the arch reconstruction, innominate artery dimension, or shunt placement. However, specific patient subsets can exhibit profound abnormalities in the pulmonary vasculature that result in excessive PVR elevations. Neonates born with HLHS and severe pulmonary venous obstruction due to intact atrial septum without alternative decompressing veins routinely demonstrate extremely high and volatile PVR. Even the typical HLHS anatomic constellation is associated with marked abnormalities in the number and muscularization of the pulmonary vasculature by pathologic examination. Hypotheses attribute these changes to chronic fetal pulmonary venous obstruction [106, 107]. One can speculate that these changes become more extreme in the context of the significant obstruction caused by HLHS with an intact atrial septum. Fetal echocardiography has confirmed alteration in pulmonary venous flow pattern in relation to the magnitude of restriction at the atrial septum [108].

In the context of hypoxemia due to low Qp:Qs, interventions fall into three categories: surgical modification, pulmonary vasodilation measures, and systemic vasoconstriction. In the subgroup of patients expected to have unusually elevated PVR, surgical modifications might entail placement of a larger shunt or placement of a central shunt (PTFE shunt from the aorta to the pulmonary artery). Potential pulmonary vasodilator therapies include the strategies one might employ in any patient demonstrating elevated PVR, such as supplemental oxygen, moderate hyperventilation, normothermia, alkali, and inhaled nitric oxide (iNO) [109–111]. Additionally, the use of higher doses of inotropic infusions or even vasoconstrictors to increase the driving pressure across the shunt may be considered. The latter necessitates careful monitoring to avoid jeopardizing perfusion to other vital organs.

The depressed myocardial function represents another potential problem in the early post‐CPB period. Some degree of myocardial dysfunction typically occurs following stage I palliation, as no hemodynamic benefit is achieved that offsets the cost of CPB and an ischemic interval. When this dysfunction becomes more significant than usual, specific causes should be sought. Even in the context of the typical conduct of stage I reconstruction, the presence of AA can make routine myocardial protection measures, such as the infusion of cardioplegia solutions, challenging, and thus inadequate myocardial preservation represents one potential cause for persistent or excessive myocardial depression.

Technical considerations represent the predominant cause of myocardial dysfunction following this complex intervention. One of the most complex aspects of this procedure is the reconstruction of an aortic arch so that the small ascending aorta, which principally serves to provide coronary flow, is not compromised. Problems with retrograde flow to the coronary arteries may not become evident until the cardiac volume is restored in anticipation of terminating CPB. Residual hemodynamic derangement represents another potential cause of myocardial dysfunction. Given that under the best circumstances, the Norwood operation does not provide appreciable hemodynamic benefit, one would expect that a surgical result with newly imposed volume or pressure loads would be poorly tolerated. Examples of such findings include residual aortic arch obstruction, AV valve dysfunction, and/or semilunar valve obstruction or regurgitation.

Metabolic disturbances can also result in significant myocardial dysfunction. A fragile RV struggling to cope with significantly increased volume output demands at systemic pressure is perhaps more susceptible to what might otherwise be modest metabolic disturbances. As such, variables that can impact myocardial performance, such as ionized calcium levels and the presence of lactic acidosis, should be monitored closely and observed for trends. The rapid administration of blood products containing calcium‐binding drugs, high levels of potassium, and lactic acid, along with other vasoactive mediators can result in an acute, profound deterioration in cardiac performance in the early postoperative period. When the arterial pH falls below 7.3, myocardial performance may deteriorate and contribute to a further reduction in Qs. The administration of IV bicarbonate, calculated to eliminate the base deficit completely, often exerts a beneficial effect on both myocardial performance and Qs. In addition to the inherent cardiac sensitivity, certain anatomic factors accentuate this vulnerability to metabolic disturbances. Blood carrying the transfused products from the systemic venous circulation enters the RV and is directed immediately to the reconstructed aorta, which has as its first branch the coronary arterial circulation. Thus, constituents of the transfused blood (e.g., citrate, potassium, lactate) infused into the venous circulation arrive at the coronary arteries with greater speed and more concentration than might have occurred had they been dissipated throughout the pulmonary vasculature before entering the aorta. This effect is further accentuated if central venous catheters are employed to infuse the blood product. It is best to utilize either washed packed cells or blood that is <7 days of age.

Arrhythmias most commonly occur as manifestations of the problems described previously. When they occur early in the rewarming process on CPB, coronary insufficiency represents the most common cause, particularly if the arrhythmia is ventricular in origin. Metabolic disturbances produce the same qualitative rhythm changes seen in normal hearts, although the manifestations might be more extreme. Given the predominantly extracardiac nature of the Norwood procedure, acquired heart block rarely follows this operation unless it existed preoperatively. On rare occasions, a patient presents with HLHS and a primary arrhythmia, such as Wolff–Parkinson–White syndrome. A 2021 retrospective review of 120 patients with HLHS after the Norwood procedure demonstrated a 26% incidence of postoperative arrhythmias. Of these, the majority (32.4%) were ectopic atrial tachycardia [112].

Excessive PBF may complicate the early postoperative period; however, this diagnosis should be entertained cautiously. In many instances, the apparent excess PBF reflects a relative imbalance with respect to significantly diminished systemic CO (Qs). The latter should be specifically excluded and addressed before invoking extreme measures to restrict PBF. Of course, subtle technical differences in the conduct of the operation can result in an anatomic propensity to an excessive Qp: Qs, which can, in turn, jeopardize systemic perfusion. Such patients typically exhibit an extremely wide pulse pressure or low diastolic pressure, reflecting pulmonary “run‐off.” If myocardial performance otherwise appears robust, specific measures employed to increase PVR are appropriate in this setting. In most patients, this condition dissipates as the infant recovers from surgery. Should the problem persist beyond the first postoperative day, cardiac catheterization should be considered to evaluate the need for further surgical intervention to diminish PBF.

The subset of patients who demonstrate either an inability to separate from CPB or refractory cardiac dysfunction postoperatively may benefit from ECMO utilization as a support strategy. Ungerleider et al. have advocated the routine use of mechanical ventricular assistance in all patients following stage I reconstruction, reporting 89% survival to hospital discharge [113]. According to data from the National Pediatric Cardiology Quality Improvement Initiative database, 66 of 931 (7.1%) HLHS patients required ECMO during admission for stage I palliation. Patients requiring ECMO had a significantly higher risk of death or transplant (18%) compared to those who did not require ECMO (9%) [114]. Risk factors described for post‐Norwood ECMO requirement include birth weight <2.5 kg, longer CPB time, peak serum lactate >9 mmol/L, and the vasoactive inotropic score of >27 [115]. An analysis from the Extracorporeal Life Support Organization of 738 HLHS patients requiring ECMO after stage I palliation demonstrated a 31% survival rate. Longer duration of mechanical ventilation before ECMO, longer duration of ECMO support, and ECMO complications such as renal failure or neurologic injury, all increased the risk of mortality [116, 117].

The risk of cardiac arrest in neonates with HLHS remains a concern in stage I postoperatively and the inter‐stage period until SCPC is performed [118]. According to the Society of Thoracic Surgeons Congenital Heart Surgery Database, there was a 12.7% incidence of cardiac arrest among stage I palliation patients between 2007 and 2012. CPR in this population is also difficult, with a 62.3% mortality rate compared to 12.5% mortality in patients who did not experience an arrest [119]. The 2020 American Heart Association Pediatric Advanced Life Support guidelines include a specific section regarding the resuscitation of patients with SV physiology [120]. In stage I palliated patients, chest compressions provide both pulmonary and systemic blood flow, while chest recoil increases filling to the SV. Due to the parallel circulation, systemic blood flow will be lower and may have a low oxygen content, leading to coronary ischemia, systemic hypoperfusion, and end‐organ damage [118]. Cardiac arrest and resuscitation are discussed in greater detail in Chapter 25.

The European Association for Cardio‐Thoracic Surgery (EACTS) and the Association for European Paediatric and Congenital Cardiology (AEPC) recently developed a set of guidelines for managing neonates and infants with HLHS. These guidelines include suggested management for diagnosis, imaging, preoperative assessment, surgical and anesthetic management, and inter‐stage monitoring. These recommendations allow practice standardization and identify potential targets for practice improvement at individual institutions [121]

Pulmonary artery banding

Neonates with pulmonary overcirculation and varying degrees of CHF may present for PAB placement to protect the pulmonary vascular bed from high pressure to facilitate successful progression to SCPC and Fontan physiology. Cyanosis is mild in these patients; they are generally tachypneic and often receive diuretic therapy. Preoperative evaluation should include a thorough history and physical examination, a review of the chest radiograph for evaluation of heart size and degree of pulmonary overcirculation, and baseline pulse oximetry. A current hemoglobin level will aid in the assessment of the degree of chronic cyanosis, and serum electrolyte values can evaluate abnormalities due to chronic diuretic therapy. A review of the most recent transthoracic echocardiogram should reveal precise details of the anatomy and functional characteristics, including any mitral valve regurgitation. Although cardiac rhythm disturbances are unusual at this stage, a 12‐lead ECG is helpful to exclude arrhythmias.

After placement of standard monitors, an IV induction of anesthesia is most commonly performed, utilizing fentanyl in combination with incremental doses of either ketamine or propofol. After administration of a neuromuscular blocking agent and tracheal intubation arterial access is achieved for monitoring. Institutional viewpoints vary regarding the need and best location for placement of central venous access. Some centers advocate the use of femoral venous cannulation, due to concern for potential thrombus formation in the SVC, as it will be necessary later for successful SCPC. Others may choose to place a transthoracic RA line. If appropriate peripheral IV access exists some centers elect not to use central venous access for this procedure. The patient’s preoperative condition, along with institutional preferences and the need for postoperative vascular access in the ICU, will assist in decision‐making.

Steady‐state anesthetic and ventilatory conditions should be maintained during the procedure in order to optimize PAB adjustments. The FiO₂ and PaCO₂ at the time of PAB placement should approximate baseline conditions for the patient; communication between the surgeon and anesthesiologist regarding the desired conditions is essential. A FiO₂ of 0.30 or less is generally utilized, and PaCO₂ is often in the mid‐40s mmHg due to the metabolic alkalosis from diuretic therapy and the mild state of respiratory insufficiency from CHF. Many centers use TEE guidance during PAB placement.

The operation is usually performed through a median sternotomy without CPB. After dissection, the surgeon places a thin cotton tape or strip of Gore‐Tex around the main PA, using care not to impinge on the branch PAs or pulmonary valve. A formula derived by Trusler advocated an initial band circumference of 24 mm + 1 mm/kg body weight for infants with bidirectional mixing disorders, but this was subsequently modified for patients with univentricular physiology, narrowing the circumference further to 22 mm + 1 mm/kg [132, 133]. Various methods may be used to assess and adjust PAB tightness. The primary method is to assess the change in pulse oximeter saturation and the measured arterial PaO₂ saturation after band placement, with a SpO₂ of 75–85% and a PaO₂ of approximately 40–45 mmHg customarily targeted. Transesophageal echocardiography helps measure peak velocity (V) across the PA band; 3 m/s is a frequent goal and corresponds to a pressure gradient of 36 mmHg using the modified Bernoulli principle (peak velocity = 4V²). Additionally, echocardiography can reassure that no encroachment on the pulmonary valve or distortion of the pulmonary arteries has occurred. Finally, direct pressure measurement in the surgical field is utilized by some institutions; a PA pressure distal to the PAB that is about one‐third of systemic arterial pressure (range, 25–50%) is a frequent goal. The PAB is adjusted temporarily with surgical clips, blood gas measurements can be obtained after steady‐state conditions are achieved, and the band is then permanently secured with non‐absorbable sutures. An overly tight PAB will produce profound cyanosis and bradycardia, while one that is too loose will produce little change from baseline conditions. The final PAB circumference is often determined by surgical judgment, hence the importance of achieving and maintaining steady‐state anesthetic and ventilatory conditions.

Although usually not necessary, inotropic support should be readily available as afterload will increase after placement of the PAB once the pulmonary circuit no longer provides a low resistance “pop‐off.” Placement of a PAB mechanically increases the resistance to PA flow, and therefore, more blood flow is directed out through the aorta. Systemic arterial pressure often increases after placement of a PAB, and occasionally a vasodilator such as sodium nitroprusside is necessary to control blood pressure. After surgery, tracheal extubation may occur in the OR or within the first few hours of ICU admission.

As the patient emerges from anesthesia, CO changes, along with the relative tightness of the band. As infants are in a period of rapid somatic and cardiac growth, they often “outgrow” the PAB, and cyanosis increases. As nearly all of these infants undergo an SCPC within a few months, this issue is not usually problematic in the infant SV population.

Systemic‐PA shunt and ductal stenting

Patients with TA and either pulmonary atresia or significant PS who require PGE₁ and maintenance of ductal patency for PBF may present for either surgical placement of a mBTTS or placement of a ductal stent in the cardiac catheterization laboratory to establish a stable source of PBF. Occasionally an older infant with profound cyanosis and a closed ductus arteriosus may also present for placement of an mBTTS. Although an mBTTS can be performed via right thoracotomy, the median sternotomy approach is most commonly utilized. Unless necessary due to hemodynamic instability or significant hemoglobin‐oxygen saturations during the procedure, CPB is not utilized. The anesthetic considerations for placement of an mBTTS are presented in detail in Chapter 28.

Percutaneous transcatheter stenting of the ductus arteriosus was first described in 1992 and has continued to increase in prevalence as an alternative palliative approach to secure PBF in patients with duct‐dependent lesions [134]. Ductal stenting (DS) has proven to be a reasonable alternative to the mBTTS, as it is associated with similar or improved symmetrical PA growth, fewer procedural complications, a shorter ICU and hospital length of stay, and improved survival compared to surgical shunt [135, 136]. In comparing the two procedures, cardiac arrest, the need for extracorporeal circulatory support, arrhythmias requiring therapy, and bleeding were observed only in the mBTTS group, while access site injury was observed in the DS group [137]. As the experience continues to grow and early technical challenges are overcome, DS is emerging as a preferred alternative to the mBTTS. Further detail about ductal stenting is presented in Chapter 34.

Unplanned reinterventions to treat cyanosis before SCPC may be required in patients with either mBTTS or DS. These interventions most commonly involve either balloon or stent angioplasty of the PDA stent or mBTTS, or surgical revision [136]. Planned reinterventions may also occur in DS patients, with reexpansion of the stented ductus allowing adjustment of PBF over time. A study comparing pre‐SCPC hemodynamics and PA growth in SV patients palliated with either DS or mBTTS found no significant differences between the two [138].

Damus‐Kaye‐Stansel operation and palliative ASO

Management of patients with unrestrictive PBF and systemic ventricular outflow tract obstruction (TA with associated TGA) remains challenging. Systemic blood flow must pass through the VSD, or bulboventricular foramen, to a rudimentary subaortic outlet chamber. The VSD may become restrictive, limiting blood flow, or ventricular hypertrophy after placement of a PAB may further increase outflow obstruction. While early relief of obstruction is essential, surgical management of this issue has not been uniform. Although subaortic stenosis may be relieved directly by means of VSD and/or subaortic chamber enlargement, this incurs the risks of heart block, recurrent stenosis, and ventricular dysfunction due to the ventriculotomy. As noted earlier, the DKS operation may be performed for TA variants with significant subaortic or aortic obstruction or with a restrictive VSD. The anesthetic considerations are very similar to those for Norwood stage I palliation, as discussed earlier. Because the single functional ventricle is an LV, and because the aortic valve is not atretic and the coronary artery circulation not as tenuous as in HLHS, these patients are generally expected to have a more stable pathophysiologic condition after DKS, compared with the Norwood stage I palliation. The palliative ASO has many anesthetic considerations in common with the standard ASO for d‐TGA with two ventricles (discussed in Chapter 29). The major difference is that TA remains, and adequate PBF needs to be assured after the ASO, often requiring enlargement of the VSD or possibly placement of an mBTTS.