Critical aortic valve stenosis

With critical aortic stenosis, there is duct-dependent systemic circulation. It is, thus, often difficult to predict whether relief of the outflow obstruction will result in a viable biventricular circulation or whether the left ventricle itself is too small to support the systemic output. To complicate issues further, the mitral valve is often stenotic or becomes stenotic over time. Left ventricular hypoplasia is associated with critical aortic stenosis likely related to cessation of growth.Reference Mocellin, Sauer and Simon^17, Reference Gundry and Behrendt¹⁸ Over the past three decades, many investigators have reported the potential risk factors for poor outcome after biventricular repair in patients with critical aortic stenosis. Smaller left ventricle size has been consistently associated with higher mortality in this cohort in multiple studies.Reference Mocellin, Sauer and Simon^17–Reference Leung, McKay and Smith²¹ A left ventricular end-diastolic volume <20 ml/m² has been considered too small for systemic viability.Reference Hammon, Lupinetti and Maples²² Other researchers have focused on specific characteristics of the left ventricle such as endocardial fibroelastosis.Reference Mocellin, Sauer and Simon^17, Reference Gundry and Behrendt¹⁸ In earlier studies, endocardial fibroelastosis was diagnosed only during autopsy. More recent studies have suggested that endocardial fibroelastosis is difficult to diagnose by echocardiography and can be more accurately diagnosed by cardiac magnetic resonance.Reference Stranzinger, Ensing and Hernandez²³ Nevertheless, even in prospective analyses, endocardial fibroelastosis can be difficult to diagnose using echocardiography or cardiac magnetic resonance.Reference Tuo, Khambadkone and Tann²⁴ Endocardial fibroelastosis portends left ventricle diastolic dysfunction and poor filling even when the aortic stenosis is relieved by surgical or transcatheter valvotomy. In the presence of endocardial fibroelastosis, the consequence of biventricular repair may be left atrial hypertension and eventual pulmonary hypertension.Reference Tuo, Khambadkone and Tann²⁴ In the long term, restrictive cardiomyopathy may develop.Reference Schwienbacher, Schweigmann and Neu^25, Reference Friedman, Margossian and Grahan²⁶

In 1991, Rhodes et al designed one of the first studies that analysed multiple variables and their impact on surgical outcomes for critical aortic stenosis.Reference Rhodes, Colan and Perry²⁷ A multivariate equation using echocardiographic measurements was developed to help predict patients who would be more suitable for a biventricular repair (Fig 1). A score of less than −0.35 correctly predicted death (retrospectively) after biventricular repair in 88% of patients in whom it was applied. Risk factors were also identified for death after biventricular repair, including left ventricle length, aortic root size, mitral valve area, and left ventricle mass. The mortality rate was 100% in patients who had two or more of these risk factors and only 8% in those with one or zero. As these criteria are from an earlier surgical era, the Ross and Konno procedure were not taken into account. Thus, the small aortic root size may not be as relevant today. Despite this limitation, the “Rhodes” score is ubiquitous and is often used to help paediatric cardiologists and surgeons determine the best surgical approach for patients with congenital aortic stenosis and left ventricle hypoplasia.

Figure 1

Multivariate equation used as Rhodes score for successful biventricular repair.Reference Rhodes, Colan and Perry²⁷

The Congenital Heart Surgeons Society also sought to determine patients with aortic stenosis and small left ventricles who would do well with a biventricular repair.Reference Lofland, McCrindle and Williams²⁸ This study was unique in that it also considered risk factors for single ventricle palliation. A multivariable analysis of factors associated with outcomes was used to predict the magnitude and the direction of the survival benefit for the optimal pathway – biventricular repair or single ventricle palliation (Fig 2). Unlike the Rhodes score, the Congenital Heart Surgeons Society incorporated endocardial fibroelastosis into the scoring system, as well as tricuspid regurgitation – a risk factor in single ventricle palliation. Of note, it did not include a measure of mitral valve size, which many consider critical to help determine left ventricle viability.

Figure 2

Congenital Heart Surgeons Society’s multiple linear regression equation for survival benefit after biventricular repair.Reference Lofland, McCrindle and Williams²⁸

More recently, Colan et al further refined the Rhodes criteria, yielding a model (Fig 3) that was similar to it, but with a reported 90% accuracy at predicting survival with biventricular circulation among neonates with critical aortic stenosis and a mitral valve area z-score of less than −2.Reference Colan, McElhinney and Crawford²⁹ This equation assumes that any patient with a mitral valve z-score worse than −2 does not require biventricular repair. It should be noted that all these scoring systems have not been validated for other types of left ventricle hypoplasia. In fact, many patients who have “failed” these criteria have had successful biventricular repair procedures.Reference Blaufox, Wyman and Lopez^30, Reference Tani, Minich and Pagotto³¹ This is especially true for patients with arch hypoplasia. Therefore, the Congenital Heart Surgeons Society and modified Rhodes criteria should be used exclusively for patients with critical aortic stenosis.

Figure 3

New scoring system for aortic stenosis and left ventricular hypoplasia.Reference Colan, McElhinney and Crawford²⁹

Once a child is born with critical aortic stenosis, the evaluation of left ventricle size becomes critical in order to determine whether biventricular repair is possible; however, much has been learnt about this disease by how it evolves before birth. Many fetuses with aortic stenosis progress to develop hypoplastic left heart syndrome in utero.Reference Allan, Sharland and Tynan^32, Reference Sharland, Chita and Fagg³³ This observation, first noted in the 1980s, has been the impetus for the field of fetal cardiac intervention. In mid-gestation, some fetuses with critical aortic stenosis will present with a left ventricle that is dilated, poorly functioning, and evident of endocardial fibroelastosis.Reference Sharland, Chita and Fagg³³ Over the course of the pregnancy, this type of left ventricle often has cessation of growth such that it is markedly hypoplastic at birth. In a fetal study assessing critical aortic stenosis, fetuses that developed hypoplastic left heart syndrome were more likely to exhibit reversal of flow in the transverse aortic arch and foramen ovale, monophasic flow across the mitral valve, and left ventricle dysfunction compared with fetuses with viable left ventricles at birth.Reference Mäkikallio, McElhinney and Levine³⁴ Others have suggested that patients who develop aortic stenosis postnatally have little or no growth of the left ventricle, aortic valve, or mitral valve during fetal life; thus, following the trajectory of growth of these structures is helpful to predict severity of disease.Reference Hornberger, Sanders and Rein^35, Reference McCaffrey and Sherman³⁶ Mid-gestation risk factors for cessation of left ventricle growth are used to justify fetal aortic balloon valvuloplasty to potentially promote forward flow through the left ventricle, and thus growth of the left-sided structures.Reference Tworetzky, Wilkins-Haug and Jennings³⁷

Aortic arch hypoplasia

Aortic arch hypoplasia has been associated with left ventricle hypoplasia as well. The aortic valve may be small but is typically not stenotic. Often, the mitral valve is normal in structure but hypoplastic. In some cases, the mitral valve is structurally abnormal, present either as a parachute or an arcade valve. Similar to critical aortic stenosis, the decision-making process regarding aortic arch hypoplasia and a small left ventricle is also challenging.

There are several issues that make aortic arch hypoplasia different from critical aortic stenosis. In critical aortic stenosis, the left ventricle is often small but distended because of endocardial fibroelastosis (Fig 4a). Thus, it is not as amenable to alterations in filling to promote growth. In contrast, the left ventricle in aortic arch hypoplasia is typically slim and not apex-forming (Fig 4b). Thus, increased flow into the left ventricle may potentially increase the left ventricle size.

Figure 4

(a) In critical aortic stenosis, the left ventricle (LV) is often small but distended because of endocardial fibroelastosis (EFE). Thus, it may not be amenable to alterations in filling to promote growth. (b) In contrast, the LV in aortic arch hypoplasia is typically slim and not apex-forming. Increased flow into the LV may potentially increase the LV size.

Early postnatal studies have demonstrated that neonates with coarctation of the aorta and a hypoplastic, but morphologically normal, left ventricle can have successful secondary growth of the left-sided structures after repair.Reference Tani, Minich and Pagotto^31, Reference Minich, Tani and Hawkins³⁸ Pulchalski et al reported that the mitral and aortic valve z-scores increased significantly after repair of coarctation of the aorta even in patients who start with mitral and aortic valve z-scores of as low as −4.Reference Puchalski, Williams and Hawkins³⁹ In their cohort, all the patients survived, but 20% of them required re-intervention, which was primarily to address the left ventricular outflow tract. A follow-up study of the same population demonstrated that the overall freedom from re-intervention was 76% at 15 years after the original arch repair.Reference Gray, Tani and Weng⁴⁰

Recently, a scoring system has been devised to help determine viability of biventricular repair in patients with borderline left ventricle, regardless of other associated lesions (Fig 5).Reference Mart and Eckhauser⁴¹ Predictors include mitral valve annulus size, aortic valve annulus size, left ventricular length, right ventricular length, main pulmonary artery diameter, and patient’s body surface area. Successful biventricular repair has been associated with a score of ⩽16.2, with 100% sensitivity and specificity. This score has not yet been prospectively validated.

Figure 5

Echocardiographic scoring system for borderline left ventricle.Reference Mart and Eckhauser⁴¹ AVPSLA, aortic valve annulus measured in cm in the parasternal long-axis view in mid systole; BSA, patient’s body surface area, in m²; LVL4C, left ventricular length measured in cm in the 4-chamber view at end diastole; MPA, main pulmonary artery diameter measured in cm either in the parasternal long-axis or short-axis view in mid systole; MV4C, mitral valve annulus measured in cm in the apical four-chamber view at end diastole; RVL4C, right ventricular length measured in cm in the 4-chamber view at end diastole.

Arch obstruction often presents in a more subtle manner in utero than critical aortic stenosis. Often, a discrepancy between left ventricle and right ventricle size is the most notable feature.Reference Brown, Durfee and Hornberger⁴² Quartermain et al reported that postnatal aortic arch obstruction was common in fetuses with LV:RV size discrepancy.Reference Quartermain, Cohen and Dominguez⁴³ A ratio of ⩽0.6 had good sensitivity for the prediction of need for neonatal arch intervention. Others have shown that measures of the ascending aorta, aortic isthmus, and pulmonary valve-to-aortic valve diameter ratio are good predictors of postnatal coarctation.Reference Jantzen, Gelehrter and Yu^44, Reference Gómez-Montes, Herraiz and Gómez-Arriaga⁴⁵ Slower rate of growth of the aortic valve is also quite sensitive and specific for predicting coarctation of the aorta.Reference Gómez-Montes, Herraiz and Gómez-Arriaga⁴⁵

Mitral valve disease

The mitral valve is one of the most important factors in determining the feasibility of biventricular repair. Mitral valve abnormalities are associated with arch hypoplasia and critical aortic stenosis and are often a component of Shone’s complex.Reference Shone, Sellers and Anderson⁴⁶ When mitral valve abnormality is the primary disease, survival after surgical intervention is quite poor compared with intervention on other valves.Reference McElhinney, Sherwood and Keane⁴⁷ Many patients with left ventricle hypoplasia have structural mitral valve disease, most typically arcade mitral valve, with shortened or no chordae tendinae, or parachute mitral valve, with the majority of chordae tendinae attached to one papillary muscle. Importantly, mitral stenosis is often not present at birth, only to progress over time after initial repair. Previous studies assessing the success of biventricular repair in patients with borderline left ventricle have reported high rates of re-intervention, particularly of the mitral valve, with mitral valvuloplasty or mitral valve replacement being common.Reference Alboliras, Mavroudis and Pahl^12–Reference Tchervenkov, Tahta, Jutras and Béland¹⁴ In a recent study of 49 infants and children who underwent the Ross/Konno procedure, mitral valve disease was highly associated with mortality.Reference Vergnat, Roubertie and Lambert¹⁶ Morbidity included multiple re-operations on the mitral valve and pulmonary hypertension. Others have used selection criteria that focus on mitral valve disease in determining whether a biventricular repair should be attempted.Reference Avitabile, Mercer-Rosa and Ravishankar⁴⁸ Using these factors, biventricular repair can be successful with the mitral valve having little impact on the outcome; however, re-operation rates for other left-sided issues remain quite high.Reference Avitabile, Mercer-Rosa and Ravishankar⁴⁸ Among all considerations for biventricular repair, the mitral valve must remain an important focus of the evaluation.

Fetal intervention

With the observation that fetal critical aortic stenosis can evolve into hypoplastic left heart syndrome, fetal intervention was explored as early as 25 years ago.Reference Allan, Sharland and Tynan^32, Reference Sharland, Chita and Fagg³³ Transcatheter fetal aortic valve dilation has been performed successfully in over one hundred fetuses over the last decade, much to the credit of the team at Boston Children’s Hospital. Although it has been technically successful, the number of patients who have had a successful and sustained biventricular repair after birth has remained limited.Reference Tworetzky, Wilkins-Haug and Jennings^37, Reference Freund, McElhinney and Marshall⁴⁹ Moreover, multiple re-interventions over the first few years of life are common.Reference Freund, McElhinney and Marshall⁴⁹ Fetal aortic valvuloplasty can result in aortic valve and mitral valve growth.Reference Maxwell, Allan and Tynan^50, Reference McElhinney, Marshall and Wilkins-Haug⁵¹ It has also been shown to improve left ventricle function in some patients.Reference Freund, McElhinney and Marshall^49, Reference McElhinney, Marshall and Wilkins-Haug⁵¹ Importantly, these cohorts have been studied without a control group. Thus, it is difficult to know whether fetuses without fetal intervention would have had a different outcome. Moreover, fetal intervention poses a procedural risk to the mother and a 10% mortality rate for the fetus. Nonetheless, fetal intervention has moved the field of paediatric cardiology forward and will likely continue to have a place in treatment of the borderline left ventricle.

Left ventricle rehabilitation strategy

For left ventricle hypoplasia with arch obstruction, arch repair is often the only surgical procedure necessary because the mitral and aortic valves are unobstructed and endocardial fibroelastosis is not present. On the other hand, for patients with aortic valve disease, mitral valve disease, and endocardial fibroelastosis, surgical strategy may be more complicated. Several new treatment strategies have been proposed over the last decade to augment forward flow into the left ventricle, and thus promote growth of the left-sided structures; one such method, now termed left ventricle rehabilitation, includes a combination of techniques to relieve inflow and outflow tract obstruction coupled with endocardial fibroelastosis resection. Endocardial fibroelastosis resection during a Ross/Konno procedure was first reported in 1997 and resulted in improved diastolic function and improved left ventricular stroke volume.Reference Van Son, Falk and Mohr⁵² Emani et al recently evaluated nine patients who underwent left ventricular rehabilitation surgery; there was no operative mortality and no deaths from cardiac causes after a median of 25 months of follow-up,Reference Emani, Bacha and McElhinney⁵³ but two patients did require re-operation. There was a decrease in left atrial pressure and left ventricular end-diastolic pressure with improved ejection fraction post-operatively; however, pulmonary artery pressure and pulmonary vascular resistance were not reported. These patients require close follow-up over a more extended period of time, with follow-up haemodynamic measurements to determine what happens to the pulmonary vascular resistance and pulmonary artery pressures over time.

The same group has also proposed another newer surgical strategy – that is, performing left ventricle recruitment in a staged manner. An initial Norwood procedure is performed, but the atrial septum is left restrictive on purpose to promote forward flow through the mitral valve. After a variable period of time of growth, the left ventricle is “rehabilitated” with endocardial fibroelastosis resection, mitral valve chordal elongation, and repair of the left ventricular outflow tract. Of the 34 patients in the staged recruitment group, 13 of them underwent biventricular conversion.Reference Emani, McElhinney and Tworetzky⁵⁴ At the median follow-up of 2.9 years (range 1–6 years), there has been no mortality. Re-operation after biventricular conversion has been performed in four patients. In about half of the patients, the right ventricle pressure has been reported as half the systemic arterial pressure. In the other half, the right ventricle pressure has not been measured or is more than half of the systemic pressure. Of note, pulmonary vascular resistance, left atrial pressures, and pulmonary artery pressures have not been reported in this cohort. It also remains to be seen whether the left ventricle growth that is seen is actual growth or pathologic dilation. Further follow-up studies need to be performed to validate these findings. Hammel et al published a report of four patients who had critical aortic stenosis and severe left ventricular dilation and dysfunction.Reference Hammel, Duncan and Danford⁵⁵ They also proposed a two-stage surgical approach. The first stage consisted of surgical aortic valvotomy, bilateral pulmonary artery banding, and atrial septectomy. The second stage included patch closure of the atrial septal defect, ligation of the ductus arteriosus, and removal of the pulmonary artery bands. Prostaglandin infusion continued between the stages to maintain right ventricle contribution to systemic perfusion. The results were mixed, with two patients dying before the second stage and two patients surviving with a normal left ventricle ejection fraction. These reports demonstrate the extreme variability of outcomes in patients with borderline left ventricles, regardless of treatment.