Abstract:
Biventricular repair requires two ventricles, each capable of supporting the systemic and pulmonary circulation. Septation has been interpreted as the pursuit of biventricular circulation rather than single-ventricle track. Although successful biventricular repair is preferable, the improvement in neonatal staged palliation has improved single-ventricle prognosis. Recent studies have shown the two-ventricle options that subject both the child and family to multiple surgical procedures and lengthy hospitalizations, as well as low cardiac output, should be avoided. Others argue that an aggressive approach to biventricular or one-and-one-half-ventricle management should be the goal with two ventricles and more complex intraventricular anatomy. We review diagnosis-specific issues for decision making and palliative procedures. Many of the studies regarding decision making give a specific surgical approach to a specific anatomy and physiology, along with decision-making guidelines for selecting between a biventricular and single-ventricle repair. The risk factors that have an impact on early mortality or survival, which are outlined in the chapter, are helpful for application of practice at an individual center. However, there are no data to support a particular pathway that is applicable to all centers. Multi-institutional studies that identify guidelines or calculators may be more broadly applicable. The final decision for univentricular versus biventricular is patient specific and rests on the experience of the surgeon and the institutional resources.
Key Words
univentricular, biventricular repair, palliation, systemic outflow obstruction
Overview
Biventricular repair requires two ventricles, each fully capable of supporting the systemic and pulmonary circulation. Septation has been interpreted as the pursuit of biventricular circulation rather than single-ventricle track. Although successful biventricular repair is preferable, the improvement in neonatal staged palliation has improved single-ventricle prognosis. Recent studies have shown the two-ventricle options that subject both the child and family to multiple surgical procedures and lengthy hospitalizations, as well as low cardiac output, should be avoided. Others argue that an aggressive approach to biventricular or one-and-one-half-ventricle surgical management should be the goal with two ventricles and more complex intraventricular anatomy.
Atrioventricular valve size, ventricular size and volume, chordal configuration, systemic and pulmonary venous return, location of ventricular septal defect(s), and anatomy of intraventricular connections are all components that go into the decision. There are several tools and calculators to help with determining which pathway a patient should ultimately follow based on anatomy and physiology (see Chapter 51 on left ventricular [LV] outflow tract obstruction).
Even after the decision has been made and the patient placed into a single-ventricle pathway, some of these patients may be converted to a one-and-one-half-ventricle or biventricular repair. Although it is preferable to accurately triage appropriate patients to biventricular repair, later conversion is an option. This chapter reviews the decision making for clinical pathway single-ventricle versus biventricular repair, as well as the initial palliative procedures that are requisite for single-ventricle repair. Importantly, the decision-making process described within this chapter addresses the choice of single ventricle versus biventricular based only on the adequacy of the anatomy. The discussion is not relevant to those cases in which the patient can clearly be a biventricular candidate but is felt to be too complex surgically to be accomplished. A patient with two adequate ventricles and two adequate atrioventricular (AV) valves, in the absence of absolute contraindications, such as significant AV valve straddle, should be a candidate for biventricular repair.
Decision Making
Perhaps the most important and likely the very first decision that needs to be made for patients who present with congenital heart disease is “whether the patient will be a complete (two ventricle) repair or need to be staged to palliation (single ventricle—Fontan—or possibly one-and-one-half-ventricle reconstruction). Morphologic information that influences this decision includes concerns about ventricular imbalance (mainly related to the size of the ventricles), septal malalignment, valvar morphology, and the presence of the components of the ventricles. This ventricular imbalance at the ends of the spectrum represent a relatively simple surgical decision. The challenges concerning decision making arise for those subtypes that fall within the borderline position on the spectrum, often resulting in a choice between a “straightforward” single-ventricle palliation or a more complex two-ventricle repair. Decision making can be further complicated by the fact that some patients present with extremely complex anatomy and numerous diagnostic studies may be required to inform the best decisions. Fortunately, it is usually possible to stabilize patients with complex congenital heart disease using a variety of tools, including prostaglandins (to preserve pulmonary or systemic blood flow in ductal dependent lesions), ventilation, and inotropic support in dedicated intensive care units (ICUs) with consolidated and collaborative expertise in managing critical heart disease. This often provides additional time for critical decision making, and the need to make hasty determinations is less usual in the current era. Palliation in unstable or critical patients is often tricky and carries high risk, and the advent of a sophisticated multidisciplinary team approach to these patients has created great benefit. Amid these challenges it is important to remember that, although the outflow obstruction can be related to ventricular imbalance, it is the inflow in the ventricular cavity that may represent the biggest challenge to determining the feasibility of two-ventricle repair.
We will review diagnosis-specific issues for decision making. Many of the following studies give a specific surgical approach to a specific anatomy and physiology, along with decision-making guidelines for selecting between a biventricular or single-ventricle repair. The risk factors that have an impact on early mortality or survival, which are outlined later, are provided with the hope that they might be helpful for application to practice at an individual center. However, there are no data to support a particular pathway that is applicable to all centers. Multi-institutional studies that identify guidelines or calculators may be more broadly applicable. The final decision for univentricular versus biventricular is patient specific and also depends on the experience of the surgeon and the institutional resources.
Left Ventricular Hypoplasia and Outflow Obstruction
The crux for neonates born with systematic outflow obstruction is whether or not the LV is adequate to support systemic circulation. The relationship between LV hypoplasia and outflow obstruction has been related to flow, as well as shear stress at an intraventricular level. At one end of the spectrum there is such severe hypoplasia of the LV, as in hypoplastic left heart syndrome (HLHS), that transplant or a single-ventricle pathway with Fontan completion is required for survival. At the other end of the spectrum the hypoplasia can be mild enough and the LV capable of supporting the systemic circulation. However, it is the gray zone with moderate LV hypoplasia that includes the “borderline” patients where the decision making can be a challenge. The decision making with this anatomy is further complicated by the need to proceed with therapy when the systemic perfusion may be ductal dependent and the pulmonary blood flow unrestricted. It is difficult to determine, based on hemodynamics, if the LV is adequate while the ductus is patent. Consideration is given not only to the size of the LV, but also to the size of the mitral valve, the presence of any secondary obstruction across the mitral valve (as might occur in some forms of parachute mitral valve or other congenital mitral anomalies), the presence of significant endocardial fibroelastosis, the size of the LV outflow tract, the anatomy and size of the aortic valve, and the presence of an aortic coarctation. In addition, the presence of intracardiac shunts at the atrial and/or ventricular level can influence hemodynamics, and this impact needs to be considered when making critical determinations regarding potential for two-ventricle versus one-ventricle pathways. Once the ductus is closed, the values of the cardiac index, LV end-diastolic pressure, and LV outflow gradient can be considered in the determination.
Currently there is no set of defined criteria to identify which patients are more likely to benefit from biventricular repair. Systemic cardiac output obstruction can occur at different levels, including at the level of the mitral valve, LV, LV outflow tract obstruction (LVOTO), aortic valve, or aorta. Several scores and calculators have been developed to try to identify preoperative risk factors for choosing the biventricular pathway (see Chapter 51 ). The Rhodes score analyzes the LV size, aortic valve area, and mitral valve size. Its application is only for those patients with obstruction at the level of the aortic valve. The Rhodes score evaluation was subsequently updated by Colan et al., with the important caveat that, as with the Rhodes score, it is specifically for critical aortic stenosis and not for the borderline LV. The Colan modification also assumes a mitral valve annulus z score of less than 2, assuming that any mitral valve of that size likely requires univentricular palliation.
The Congenital Heart Surgeons’ Society (CHSS) conducted a large prospective multi-institutional trial looking at preoperative characteristics to determine biventricular versus single-ventricle repair. The question they set out to answer was which demographic, morphologic, and functional factors help predict survival benefit for a biventricular repair pathway versus univentricular palliation in neonates with LVOTO. The initial intended biventricular repair with either balloon dilation or surgical valvotomy was indicated in 116 patients. An initial Norwood procedure was performed in 179 patients. The survival at 5 years for the biventricular repair and Norwood were 70% and 60%, respectively. The study used multivariable hazard models for survival with each of the two pathways. Incremental risk factors for death in patients undergoing biventricular repair included lower z score of aortic valve at sinuses, younger age at entry, and higher-grade endocardial fibroelastosis. Risk factors for death in patients with Norwood as their initial procedure were lower ascending aorta diameter and moderate to severe tricuspid valve regurgitation. Based on these risk factors, a prediction of survival could be calculated for all patients. The predictions for survival for the biventricular repair versus the Norwood pathway were compared with the differences representing predicted survival benefit. The investigators found that although discordant treatment selection (i.e., a patient predicted to benefit from a univentricular repair who underwent a biventricular repair and vice versa) resulted in a worse survival for both univentricular and biventricular repair, the cost of a discordant biventricular repair in a patient who would have been predicted to benefit from a univentricular approach was much more significant.
Conclusions drawn from the CHSS study included the following: mortality is high for neonates with critical LV outflow obstruction, survival is improved with more appropriate selection of the repair pathways, and the pathway predicting survival benefit relates to the adequacy of the left-sided structures. The final result of this study was a multiple linear regression equation that predicted both magnitude and direction of the survival benefit for the optimal pathway based on characteristics of individual patients. The calculator can be found at www.chssdc.org . Follow-up studies have shown reintervention in patients selected inappropriately, as determined by the CHSS score, for biventricular repair tended to have poorer outcomes. This suggests that in borderline cases, univentricular palliation may be the safer strategy based on early to midterm follow-up.
Left ventricular outflow obstruction can occur at multiple levels, allowing for numerous and different options to provide systemic outflow, as well as to address the aortic valve. The Yasui operation, for example, is designed for the unusual patient with arch hypoplasia, severe aortic stenosis, and two well-formed ventricles, two well-formed inlet (AV) valves, and a large ventricular septal defect (VSD). When this combination of lesions is present, the Yasui procedure has shown (whether performed in one or two stages) a lower mortality rate than the Norwood operation. This procedure includes a Damus-Kaye-Stansel connection (detailed later in the chapter), a baffle of the VSD to the systemic outflow, and a conduit for right ventricle (RV) to pulmonary artery (PA) ( Fig. 62.1 ) continuity. Another option for a patient with the combination of defects described earlier is the Ross-Konno procedure, which has been used for enlarging the LV outflow tract and repairing the aortic valve annulus in neonates. Many of the studies aimed at addressing LV outflow obstruction in the neonate have developed parameters to help with the decision making. Some of these criteria include LV outflow tract dimension (subaortic area and valve) greater than or equal to the weight of the patient in kilograms to proceed with standard repair of aortic arch hypoplasia and VSD. Additional criteria include aortic annular dimension greater than 4.5 mm or z score greater than −5.
In combination with any of these procedures the resection of endocardial fibroelastosis can be helpful with long-term ventricular compliance. Both the Yasui and the Ross-Konno are procedures to provide a two-ventricle outcome for complex congenital heart disease. However, in some institutions, patients with the defects described earlier may be staged to a single-ventricle pathway, and unique features of the anatomy (disease-specific factors) and of the patient (patient-specific factors) may have critical influence on which pathway is chosen—this field is, unfortunately, that complicated!
Another study looked at echocardiographic parameters to determine independent risk factors for failure of biventricular repair for patients with multiple left heart lesions. By multivariate analysis the predictors of failure identified included moderate/large VSD, unicommissural aortic valve, and lower mitral valve lateral dimension z score. In the subset of patients with mitral valve involvement (including those with stenosis, hypoplasia, or parachute morphology) the univariate analysis showed failure of biventricular repair was associated with lower LV mass z score and lower mean LV/RV long-axis ratio. Hypoplastic mitral valve annulus was an independent risk factor for failure of the biventricular approach. One caution in determining the degree of mitral stenosis must be taken as it pertains to the transmitral Doppler velocity. The gradient may be underestimated in the presence of an atrial septal defect with left-to-right shunting. It may in fact be the mitral valve that constitutes the bigger difficulty when deciding if biventricular repair is feasible in a patient with multiple left heart obstructive lesions. In addition, following such biventricular repairs such as a Ross-Konno procedure, it is often the mitral valve that contributes significantly to long-term morbidity.
Unbalanced Atrioventricular Septal Defect
Decision making for an unbalanced complete AV septal defect (UCAVSD) is dependent on which ventricle has the majority of the inflow, the percentage of unbalance, and the adequacy of the AV valves. Several studies have identified echocardiographic data aimed at determining parameters that are successful at biventricular repair. Cohen et al. expressed the AV valve index (AVVI) as the ratio of the smaller AV valve component of the divided common AV valve over the larger component. Thus the numerator and denominator could be left or right AV valve area depending on the dominance. More recently to simplify the usage and understanding, the modified AVVI is a ratio of the relative area of the left AV valve in relation to the entire common AV valve. Therefore an AAVI between 0.4 and 0.6 identifies the balanced range of the spectrum. Patients with an AVVI between 0.2 and 0.4 with unbalance to the left (meaning that the RV is larger than the LV) who were treated with a multitude of operative management strategies, including single-ventricle palliation, were found to have poor outcomes. Additional studies have suggested the use of a standard criteria of at least 40% of the common AV valve should overlie the LV to favor a biventricular approach. These data suggest that at the extreme end of the spectrum, patients with severe imbalance and straddling chords are better managed by following the Fontan track. This also holds true when the unbalance is to the right (with a smaller RV compared with the LV) because when one ventricle (and often the size of the potential inlet valve) is too small, it may be better to choose the single-ventricle pathway over a two-ventricle repair. In some circumstances in which the RV is marginal, a one-and-one-half-ventricle pathway could be selected in which the RV is used for the inferior vena cava (IVC) return and the superior vena cava (SVC) return is diverted into the pulmonary arteries directly with a bidirectional Glenn procedure. It is also reasonable to consider a two-ventricle repair, leaving a fenestrated atrial septal defect that can be evaluated and closed at a later time in the catheterization laboratory.
Other reports argue that there is high risk associated with single-ventricle palliation in this particular anatomic condition. It is suggested that a more aggressive surgical approach to parachute mitral valve or closely spaced papillary muscles in right-dominant UCAVSD patients via the splitting of a single papillary muscle head and the elongation of a papillary muscle with release off the LV free wall can be used for biventricular repair. For a left-dominant UCAVSD the tethering attachments to the anterior papillary muscle can be mobilized and the sinus/trabecular portion of the right ventricle enlarged to allow for biventricular repair. More recent studies have examined the inflow angle between the right and left atrioventricular valves and the septum (right ventricle/left ventricle inflow angle) angle in determining right ventricular dominance, which showed promise in guiding early operative decision-making. MRI is extremely helpful for better analysis of ventricular volumes to help guide decision on treatment and should be considered as critical for many of these patients. A major discriminator for proceeding with a biventricular circulation was a measured ventricular volume of the hypoplastic (left) ventricle of greater than 20 mL/m2 by any imaging (two or three dimensional echocardiography, cardiac MRI, and cardiac catheterization) modalities. That discriminator contrasts with prior reports that argued a volume less than 40 mL/m2 was considered hypoplastic for the LV. The authors caution that the volumes must be measured in the face of equal ventricular pressures and with the septum in the midline position. They expected and observed that the septum shifts rightward in two-ventricular septation in right-dominant lesions. With left-dominant lesions an increase in RV volume was thought to be secondary to papillary muscle mobilization and increased RV inflow. It has been noted that patients with trisomy 21 with severe ventricular hypoplasia as a subgroup have been reported to have poor outcomes with single-ventricle palliation, regardless of whether they have a dominant right or LV, and therefore may benefit from a more aggressive approach at achieving a biventricular repair.
Pulmonary Atresia With Intact Ventricular Septum
The tricuspid valve is the key determinant for successful biventricular repair of pulmonary atresia with an intact ventricular septum, which would generally include an initial procedure of a transannular patch and a systemic to PA shunt. The size of the tricuspid valve is linked to development of a tripartite RV. Outcomes following biventricular repair are significantly worse when the tricuspid valve z score is less than −4. Continued evaluation of the adequacy of the right heart is necessary to determine if the biventricular approach can be successful long term. If the right heart is determined to be inadequate for borderline cases, the one-and-one-half-ventricle pathway, which places a superior cavopulmonary connection while leaving antegrade pulmonary blood flow through the pulmonary valve from the IVC, has been successful. However, the benefit of a one-and-one-half-ventricle repair is theoretical because follow-up studies suggest that exercise capacity and cardiac reserves are similar to univentricular patients.
Biventricular repair with tricuspid valve z score greater than −2.5 and a tripartite RV have improved outcomes with or without a shunt in the early neonatal period. The decision-making process for pulmonary atresia with an intact ventricular septum can be further complicated by the presence of coronary fistulae, particularly in the setting of RV dependence of the coronary circulation ( Fig. 62.2 ). When there are two of three main coronary arteries dependent on the high-pressure RV for coronary supply, it is not prudent to decompress the RV and pursue a biventricular approach. In these patients single-ventricle palliation also carries a higher risk, not only initially, but also during subsequent staged palliation. Primary transplantation has been suggested as a preferred option for this population.
Congenitally Corrected Transposition of the Great Arteries
The anatomic correction of congenitally corrected transposition of the great arteries with the “double switch” operation has been well studied. The aim of this approach is to allow the LV to function as the systemic ventricle. More recently there have been some risk factors identified, such as the need for LV retraining, whereby single-ventricle palliation may lead to better outcomes than a two-ventricle approach. If the VSD is restrictive, then LV pressure can be less than the RV (systemic) pressure at the time of presentation, depending on the presence of important subpulmonary stenosis (which is often present in this lesion. The LV outflow tract connects to the PA, and the combination of LV outflow obstruction—pulmonary stenosis—and VSD is common). When there is no LV outflow obstruction, the LV will need training to help prepare it to accommodate the systemic circulation. If the patient is older, as in a teenager or young adult, the results of training followed by the double switch procedure have been disappointing. Other studies have shown that age greater than 12 years is associated with a greater probability of LV failure and higher operative mortality at anatomic correction. Survival in subsequent studies was 84.9% at 7 years. Twelve percent had new LV dilation or impaired systolic ventricular function, and 14% developed aortic regurgitation after the double switch. Patients in that series who required LV training had moderate to severe LV dysfunction (39% compared with 6% who required no LV training).
Anatomic factors complicating a double switch include dextrocardia (especially when a Senning/Rastelli procedure is required to complete the conversion to anatomic repair, and this risk is most likely secondary to conduit compression under the sternum), anomalous coronaries, and predominant inlet rather than conoventricular VSD. Although the majority can be overcome to allow an anatomic repair, a physiologic repair can be considered. However, the outcomes of a traditional physiologic repair of congenitally corrected transposition of the great arteries with VSD closure, leaving the RV as the systemic ventricle, have been disappointing. This traditional approach of VSD closure versus Fontan procedure was evaluated. The best long-term outcomes were found with the Fontan procedure. Risk factors for death at any time were RV end-diastolic pressure greater than 17 mm Hg, complete heart block after surgery, subvalvar pulmonary stenosis, and Ebstein malformation of the tricuspid valve.
Heterotaxy Syndrome
The complex ventricular relationships and positioning associated with heterotaxy syndrome may complicate biventricular repairs in this patient population. These can be complex operations, and venous abnormalities (for both the systemic and pulmonary venous return) have been identified as risk factors for repair (either univentricular or biventricular) with the need for complex intraatrial baffles (in two-ventricle repairs) or the presence of progressive pulmonary venous stenosis (in one-ventricle pathways) carrying a late risk of obstruction. In general it is our opinion that essentially any patient with two adequate ventricles and two adequate AV valves should be considered a candidate for a biventricular repair. Despite that goal, many patients in this category are still treated with single-ventricle palliation, and the following information is included for completeness. The single-ventricle pathway results in this population have been less than ideal, although they have improved more recently. Hemodynamics are critical to consider in decision making for these patients if a single-ventricle palliation is undertaken. Many of these patients may have abnormal and obstructed pulmonary venous drainage, and in these cases the single-ventricle pathway is not likely to lead to a satisfactory outcome. A single-ventricle pathway may be applied to patients with good ventricular compliance and function, well-developed pulmonary arteries, and low pulmonary resistance (implying unobstructed pulmonary venous drainage).
Palliative Procedures
Once the decision is made to proceed with a single-ventricle pathway, the most critical factors include both when the patient should receive the initial palliation and what that palliation should be. The four keys to optimal palliation include (1) a precise diagnosis before palliation, (2) an initial operation that prevents potential late problems (such as progressive ventricular outflow obstruction), (3) adequate, controlled pulmonary blood flow, and (4) timely ventricular unloading with either bidirectional superior cavopulmonary anastomosis (bidirectional Glenn) or hemi-Fontan procedure.
Accurate preoperative evaluation is essential to planning the optimal initial palliation. Initial procedures may have to be adapted in the presence of complicating factors, such as systemic outflow obstruction (intracardiac, aortic arch, coarctation) and pulmonary venous obstruction (anomalous pulmonary venous connection and restrictive atrial septal defect). The first palliation is performed with the eventual goal of normalization of the volume and pressure work of the functional ventricle, while supplying blood adequately saturated with oxygen to the systemic circulation, regardless of the underlying cardiac anatomy. Pulmonary vascular resistance (PVR) in the early neonatal period precludes obtaining this goal in one step; therefore the palliation needs to be provided in stages that are all compatible with an ultimate plan for each individual patient.
The initial postnatal palliation requires that the pulmonary and systemic circulation remain in parallel, while pulmonary blood flow is controlled, allowing for proper development and maturation of the pulmonary vascular bed. The following stages of the palliation are to create a gradual conversion to Fontan physiology, first with superior venocavopulmonary connection and subsequently to complete cavopulmonary connection (Fontan), which then alters the circulation from being in parallel to one connected in series.
The early postnatal medical management is directed at determination of the anatomy and management of the volume and pressure load on the ventricle while maintaining adequate oxygen delivery. Underlying the stabilization is the delicate balance between the systemic and pulmonary vascular resistance with the understanding that over the first 4 to 6 weeks the PVR will fall and adjustments may have to be made accordingly. This initial time period is one of continued evaluation and adjustment to maintaining the patient’s clinical stability. The evaluation is directed at answering the following questions: (1) Is there a reliable source of systemic blood flow? (2) Is there a reliable source of pulmonary blood flow? (3) Is there any impediment to pulmonary venous return? and (4) Is there an appropriate balance between the systemic and pulmonary circulations? For complex patients, and many of these patients present complex anatomic challenges, there may be obstruction (or potential for late development of obstruction) to systemic or pulmonary blood flow. For these patients the ductus arteriosus can often be maintained with prostaglandin to support both systemic and pulmonary blood flow while decisions are made regarding the most appropriate surgical staging.
Prostaglandin E 1 (PGE 1 ) was initially described by Olley and colleagues in 1976. Patients benefiting from PGE 1 can be divided into ductal dependent pulmonary (i.e., pulmonary atresia with intact ventricular septum, or tricuspid atresia) and systemic (HLHS) circulation. Both of these groups require PGE 1 soon after birth to maintain adequate systemic oxygen delivery. When ductal patency is part of the management, the balance between systemic and pulmonary blood flow can initially be challenging. When the circulations are in parallel (meaning ventricular output can distribute through the ductus to either the systemic or pulmonary bed), the relative blood flow to each circulation depends predominantly on relative balance between pulmonary and systemic vascular resistance. Because pulmonary resistance is generally much lower than systemic resistance, flow across the ductus, presuming “normal” vascular beds to the pulmonary and systemic circulation, is typically increased to the pulmonary vascular bed, including diastolic “runoff” from the systemic circulation into the pulmonary circulation. Manipulation of the inspired gases can help manage the balance of systemic and pulmonary circulation by changing the vascular resistance in the pulmonary bed. Oxygen is a potent pulmonary vasodilator. Hypercarbia is a significant vasoconstrictor and cerebral vasodilator. Even with optimal medical management, there may still be an imbalance between the pulmonary and systemic blood flow. This relative imbalance can ultimately adversely affect the patient from a hemodynamic and systemic oxygen delivery standpoint. Thus earlier surgical palliation may be necessary when medical manipulations are inadequate to overcome the resulting imbalance in circulation.
The goals of initial palliation are to provide unobstructed systemic blood flow, controlled pulmonary blood flow with well-balanced pulmonary and systemic circulations, and unobstructed pulmonary and systemic venous return. The primary surgical options for initial palliation include (1) PA band, (2) systemic to PA shunt, (3) Damus-Kaye-Stansel (DKS) procedure, (4) Norwood/hybrid Norwood, and (5) early bidirectional Glenn or hemi-Fontan procedure. The optimal procedure to accomplish these goals will depend on the anatomy (disease-specific factors) and on the age, size, and condition of the patient at the time of presentation (patient-specific factors). Timing of initial palliation depends on the severity of the flow imbalance at baseline with regard to the pulmonary circulation. In patients with reduced pulmonary blood flow the degree of cyanosis (<70%) is the best indicator to proceed with surgical palliation with systemic to PA shunting or other appropriate form of palliation. Excessive pulmonary blood flow and the onset of signs and symptoms consistent with volume overload and heart failure (growth failure, tachycardia, tachypnea, and the need for mechanical ventilation) suggest the need for intervention. These signs will worsen as the PVR decreases. Common methods to control pulmonary blood flow include PA banding or PA ligation and a systemic to PA shunt. Control of excessive pulmonary blood flow will also be important to control the pressure and volume overload to the pulmonary circulation and to protect the lungs from the development of pulmonary vascular obstructive disease. When the patient has obstruction (or potential for obstruction) to systemic outflow, conversion of systemic outflow to the pulmonary outflow tract with a DKS type of procedure (including a Norwood reconstruction) should be considered. In these circumstances in which the pulmonary outflow tract is converted to become the systemic outflow, a source of pulmonary blood flow (either an aortopulmonary shunt or an RV-PA conduit) may need to be incorporated into the plan.
Historically, congenital heart surgery was linked to palliation because complete repair of heart defects in infancy was technically challenging. Therefore the early history of palliation included staged repair of heart defects. The lessons learned from and the techniques that evolved from these procedures that were intended to keep patients alive until more definitive surgery could be performed have greatly influenced the current methods for palliation.
Pulmonary Artery Banding
Neonatal presentation with excessive pulmonary blood flow and signs of heart failure require control of the pulmonary blood flow to allow adequate somatic growth, to eliminate the volume load on the ventricle, and to protect the pulmonary vascular bed from the development of pulmonary vascular obstructive disease as the normal post- decrease in PVR occurs. The strategy of PA banding is an attempt to optimize the pulmonary-to-systemic blood flow ratio (Q p :Q s ) to avoid the potential for resultant multiorgan system dysfunction. PA banding, first described by Muller and Danimann in 1952, is the creation of “controlled” pulmonary stenosis to prevent pulmonary hypertension and excessive pulmonary blood flow. The approach previously included left thoracotomy or median sternotomy. We more commonly perform PA banding through a median sternotomy to allow for better positioning of the band and to minimize distortion of the branch pulmonary arteries. Furthermore, median sternotomy provides excellent and flexible access to virtually any anatomic arrangement, and it also becomes the only incision many of these patients need, even for future repairs, thus limiting the number of incisions. A variety of materials (polytetrafluoroethylene [PTFE], umbilical tape, silk ligature) can be used to make the band at the discretion of the surgeon.
The Trusler formula for PA banding is a calculation used to estimate the starting band circumference: 24 mm +1 mm/kg in weight for a total admixture defect (single ventricle). The formula has been modified by Baslaim to place a band that is 2.25 mm more narrow in the single-ventricle patient population. This affords effective protection of the PA bed so that the PVR at the time of the stage II and Fontan procedures is as low as possible.
After circumferential dissection the PA is encircled by a subtraction technique or direct around the PA. The band is then tightened with clips or sutures sequentially placed through the band. Historically PA bands were commonly used to control excessive PA blood flow in infants with symptomatic VSDs until they grew larger and were more amenable to surgical VSD closure. When banding for patients with VSDs (who would ultimately become candidates for a two-ventricle repair), the goal of the band was to drop the PA pressure to approximately one-half the systemic pressure. A tighter band often created less pulmonary blood flow and significant hypoxemia. The goal for single-ventricle patients is somewhat different because it is essential to protect the pulmonary vascular bed from high pressure to enable conversion to cavopulmonary physiology in a timely fashion. For this reason we generally band the PA as tightly as we can—often dropping the PA pressure distal to the band to a “normal” PA pressure (20 mm Hg). We pay attention to the arterial oxygen saturation that results and aim to have the PA pressure as low as possible with oxygen saturations in the 75% to 85% range. When placing a band, the systemic pressures may increase. However, concomitant with this is an increase in afterload because the ventricle no longer has the low-resistance pulmonary circuit as an unobstructed “pop off,” so some of these patients may require brief (1 to 2 day) therapy with low-dose inotropes. Postoperatively it is important to make sure that the patient has adequate systemic oxygen delivery (meaning the patient has adequate systemic oxygen saturations), as noted previously, along with adequate systemic cardiac output, which can usually be monitored by following the lactic acid levels. In some cases the heart cannot manage this change in afterload, and the band may need to be adjusted in the few days following initial banding.
The PA pressure can be measured distally by direct pressure catheter to assess band tightness. Transesophageal echocardiography is helpful to determine placement and ensure no encroachment on the valve and minimal distortion of the branch pulmonary arteries. Echocardiography can also provide a Doppler flow measurement across the band site, which can be useful in measuring PA pressure. A more normal PA pressure facilitates interpretation of the PA pressure distal to the band, as well as the gradient measured by echocardiography. Pulmonary artery banding is more commonly used in single-ventricle patients with tricuspid atresia, double-inlet LV, and unbalanced AV septal defect with unrestricted pulmonary blood flow.
Caution should be used in placing PA bands in single-ventricle patients with discordant ventricular-arterial connections and reliance on a VSD or bulboventricular foramen for systemic output because this anatomic substrate can lead to systemic outflow obstruction. The postoperative course after banding a patient with this anatomy can involve significant myocardial hypertrophy and obstruction in as many as 70% of patients. These patients would be better served with a DKS or a modified Norwood procedure with pulmonary blood flow provided by a systemic to PA shunt to provide reliable systemic outflow.
Hybrid palliation consisting of bilateral PA banding and ductal stenting and a modified hybrid approach with bilateral PA banding and continuation of PGE 1 are strategies that have been used for HLHS and other complex congenital cardiac defects until second-stage palliation can be performed. This has been shown to be an effective method of resuscitation for high-risk single-ventricle neonates. The second surgery can include conventional Norwood (rapid stage) or primary transplantation. Caution should be used with ductal stenting in the presence of a diminutive ascending aorta secondary to the increased risk of compromised coronary perfusion.
Systemic to Pulmonary Artery Shunt
Shunts connecting the systemic to pulmonary circulation trace back to the early history of congenital heart surgery. Originally these shunts provided lifesaving pulmonary blood flow to patients with cyanotic heart defects at a time when there simply was no technology to allow for complete repair of the underlying defect. There are many who believe that the field of cardiac surgery was launched with the creation of the first shunt by Alfred Blalock that connected the divided subclavian artery to the side of the right PA in an infant with severe tetralogy of Fallot. The success of this procedure (a Blalock-Taussig [BT] shunt), an outcome from research in Dr. Blalock’s animal laboratory with help from surgical technician Vivien Thomas and innovative encouragement from his cardiology colleague, Helen Taussig, opened the possibility for surgeons to treat a spectrum of congenital heart defects characterized by inadequate pulmonary blood flow. Soon other types of shunts were created, including the Waterston shunt and the Potts shunt. Although the Waterston and Potts shunts are of only historical interest, the classic BT shunt, which is a direct end-to-side anastomosis of the transected subclavian artery to the PA, is still used in some centers. The Waterston shunt involves an aortopulmonary connection with anastomosis between the posterior aorta and the anterior right PA. The Potts shunt involves a connection between the descending thoracic aorta and the left PA ( Fig. 62.3 ).
