Year: 2005
Type: review
Abstract: Consideration of HLHS as a candidate for in-utero intervention is not new to the fetal surgical literature. Work in animal models has concentrated on open surgical approaches, perfecting fetal cardiac bypass and fetoscopic insertion of transesophageal ultrasound transducers3-7. In humans, dilation of a stenotic aortic valve by direct puncture of the left ventricle under ultrasound guidance was first successfully accomplished in a fetus of 33 weeks' gestational age in 19918. By 2000, 12 cases had been compiled from centers around the world. Clinical success, however, as defined by a two-ventricle circulation at birth, was confined to one child9. Our reconsideration of HLHS as a viable candidate for fetal intervention involved a multidisciplinary approach that focused on several areas. A collaborative team of cardiologists, perinatologists, radiologists, fetal surgeons and anesthesiologists jointly reviewed: (1) the current understanding of the in-utero natural history of HLHS; (2) whether a physiologic basis for the proposed intervention existed; and (3) the successes and limitations of contemporary postnatal treatment. Additionally, maternal safety and the experience gained over the past decade with fetal imaging and intervention were incorporated. Our review here of these topics will provide the reader with a background for evaluating the place of in-utero intervention for HLHS. Over the past decade, both the improved resolution and increased frequency of second-trimester fetal ultrasound examinations have expanded our knowledge of HLHS. Even among women at low risk for congenital heart disease, HLHS remains one of the most reliably diagnosed cardiac malformations10. Information gained from these prenatal ultrasound studies has reinforced the concept that neonatal HLHS is an endpoint which can result from a spectrum of left-heart cardiac abnormalities. 'Classic' HLHS as described by Simpson in an editorial in this journal in 2000, refers to the heart with both aortic valve atresia and either atresia or stenosis of the mitral valve11. The left side of the heart is reduced in size and unable to support systemic circulation. In such fetuses, a poorly developed left side of the heart is present from early gestation. Second-trimester ultrasound images are similar to those at term and are characterized by markedly diminished capacity and contractility of the left heart. However, some newborns at birth who fulfill the criteria for HLHS increasingly are recognized as having a normal or even dilated left ventricle at mid-gestation. In this group of fetuses primary left outflow tract obstruction results in progressive ventricle compromise throughout gestation. Second-trimester ultrasonography reveals not a small, dysfunctional left ventricle but critical aortic stenosis and/or severe coarctation of the aorta. Real time images, rather than static four-chamber cardiac views, will delineate the obstructed outflow and dysfunctional contractility of the left ventricle wall. Followed through gestation, these fetuses can develop increased cardiac wall echogenicity reflecting endocardial fibroelastosis, and the left ventricle diminishes in size relative to the normally growing right ventricle12, 13 (Figure 1). At birth, such infants typically have inadequate left ventricle function to support systemic circulation and become candidates for the postnatal surgical options for HLHS, staged repair or transplantation14, 15. HLHS due to critical aortic stenosis is the current subgroup of cardiac malformations in which in-utero aortic valve dilation is proposed. Four-chamber cardiac image in a fetus (a) at 22 weeks' gestation, showing that the left ventricle is dilated in the setting of aortic stenosis; (b) at 33 weeks' gestation, with development of hypoplastic left heart syndrome. An assurance that if left untreated, the infant with critical aortic stenosis at 20 weeks' gestation has a high likelihood of HLHS at birth is an essential component of the decision to proceed with in-utero valve dilation. Criteria originally derived to assess the likelihood of two-ventricle circulation in the newborn with critical aortic stenosis had some, albeit limited, applicability to the fetus16, 17. More importantly, in the late 1990s longitudinal though retrospective case series of in-utero aortic stenosis provided insight into the frequency of progression to HLHS and the predictive ability of ultrasonography. Characteristics of the group developing HLHS included little or no growth of the aortic and mitral valves, second-trimester left ventricle dilation and arrested growth of the left ventricle as measured in the short axis. The degree of ventricle-wall echogenicity was not a differentiating feature18. A similar series from the Children's Hospital in Boston had the advantage of earlier gestational age at detection and found similar predictors of postnatal disease severity12. A compilation of no or little left sided growth, reversed flow across the foramen ovale and retrograde flow in the aortic arch was highly correlated with subsequent left ventricle dimensions inadequate to support systemic circulation. These features formed our basis for selecting those second-trimester fetuses with critical aortic stenosis with a high likelihood of developing HLHS. Aortic banding in the fetal lamb model is followed initially by a period of ventricle dilation. Among fetal lamb survivors of the initial dilation phase, changes comparable to those seen in HLHS in humans are apparent. The left ventricle becomes thickened and the chamber diminishes in size. These findings reinforce the knowledge that myocardial expansion in the fetus occurs by ventricular myocyte hyperplasia, the mitotic division of new cells. Postnatally, left ventricle expansion occurs primarily by cell hypertrophy and is thus, at least to some degree, limited19. Based on these animal studies, relief of the obstruction postnatally or even prenatally once the left ventricle is thickened is unlikely to restore left ventricle growth. Removal of the obstruction is ideal when the heart is dilated, although very recent studies in humans suggest that removal of obstruction in small left ventricles will promote catch-up growth20. Whether due to initial critical aortic stenosis or underlying atresia of the left side of the heart, the newborn with HLHS faces limited options for survival. Since the early 1980s, surgical interventions have continued to advance, utilizing either a staged repair following Norwood's description in 1981 or cardiac transplantation. Which surgical approach yields the highest survival and lowest long-term morbidity remains controversial and is often center-specific. The three-staged Norwood palliation converts the cardiac system to a functional single ventricle with systemic circulation supported by the reconstructed right ventricle, the 'Fontan' circulation21. However even during the 1990s, for the first surgery, operative mortality rates of 24–54% were reported, while lower surgical mortalities (typically quoted as 5%) accompanied the second and third repairs22, 23. These outcomes have improved with time and experience. In 2001, the Children's Hospital in Boston had survival rates of 85% after the first stage and 75% at completion of all three stages. However, the long-term cardiac prognosis for children with 'Fontan' single-ventricle circulation remains relatively unknown, but decreases with time are problematic24. Cardiac transplantation for the newborn with HLHS remains the first-choice surgical intervention at some centers. While often limited by donor availability, transplantation has the theoretical advantage of restoring the physiologic two-ventricle system. While the survival rates at 1 year may be comparable with the contemporaneous Norwood operative survival rates (approximately 80%), long-term morbidities are associated with a lifelong need for immunosuppression25, 26. In addition to cardiac morbidity, concern for long-term neurologic compromise among survivors of surgical interventions for HLHS has been raised. Current studies are limited to relatively small numbers of children but in three different centers, neurologic morbidity and developmental abnormalities have been noted with a composite rate of 22%27, 28. While some contributors to this neurologic morbidity may be inherent to HLHS and due to coexisting central nervous system malformations, a portion is probably attributable to pre-operative and operative conditions. While available, and with improving surgical success, staged cardiac repairs and transplantation remain high-risk and potentially morbid procedures. Long-term follow-up for these postnatal surgical interventions is limited, and diminished cardiac function and neurologic morbidity remain a concern. Attempts at in-utero intervention for HLHS should consider both attempts to decrease neonatal surgical intervention as well as neurologic morbidity as potential clinical endpoints. Following the initial two cases described by Maxwell et al. in 19918, 10 additional cases from 1989 to 1997 were compiled9. The mean gestational age at intervention was 29.2 weeks, the earliest at 27 weeks. Technical successes were noted in seven fetuses, although complications were high with only two survivors among the 12 fetuses—one child had a successful dilation at 33 weeks' gestation and another child had an unsuccessful attempt at 30 weeks' gestation followed by emergency delivery for sustained bradycardia. Both latter infants had left-ventricle dilation late into the third trimester, perhaps consistent more with critical aortic stenosis without progression to HLHS. Four of the 12 infants died either perioperatively or within 24 h owing to bleeding, sustained bradycardia or prematurity following emergency delivery. Balloon rupture complicated 50% of the interventions and bradycardia occurred in 75% of cases. Maternal morbidity included the need for delivery by emergency Cesarean section in 25% of cases owing to fetal compromise. While the outcomes noted above represented formidable challenges, the recognized evolution of ultrasound resolution, modifications in neonatal cardiac cauterization and the expanded knowledge of fetal surgery were viewed as advantages not available to previous teams. Early in the history of in-utero blood transfusion for fetal anemia, ultrasound imaging quickly replaced fetoscopy for guiding fetal umbilical cord transfusions, and resulted in fewer fetal losses29. Ultrasound-guided fetal umbilical vein sampling and transfusion, once performed with the mother and fetus heavily sedated in an operative suite, is now an unsedated, outpatient procedure performed as early as 18 weeks' gestation by perinatologists. Experience with ultrasound-guided intracardiac placement of needles for cardiocentesis was also gained from the work of Antsaklis et al. on hemoglobinopathies30. In over 100 second-trimester fetuses, intraventricle placement of a 22-gauge needle resulted in bradycardia in 9%, hemopericardium in < 1% and fetal loss within 24 h of 3.8%. Limited work in cardiocentesis for blood transfusions suggested similar though higher rates for bradycardia, hemopericardium and loss within 24 h, of 16%, 4% and 8%, respectively31. This cumulative experience with ultrasound guidance for needle placement and the recognized lower complication rate of needle procedures contributed significantly to the decision to forgo direct fetoscopic visualization in favor of an ultrasound-guided intervention. Additionally, during the past decade important work on the resuscitation of the fetus in utero has been gained from the University of San Francisco fetal surgery team. Studies have provided guidance for incorporating inotropic medications, correction of acidosis and, when needed, fetal transfusion32. Researchers in this area have greatly enhanced our knowledge of the stress response in the fetus and the modalities applicable to inhibit pain and fetal catecholamine response during fetal interventions33. Work evaluating the cortisol responses, hemodynamic changes and noradrenaline responses during fetal needling procedures emerged during the 1990s, leading to the advocation of the administration of fetal fentanyl prior to intervention procedures34-37. Both fetal premedication to blunt the stress response and a predetermined protocol for fetal resuscitation are new contributions to the field that we have incorporated into our current fetal interventions. Throughout the history of fetal surgery, attempts at palliation or repair of congenital malformations in utero have been undertaken in the setting of incurring risks to the mother. Despite technically sound and well-researched approaches to surgery on the human fetus, success has been hampered by high rates of preterm rupture of membranes, premature delivery and maternal complications. Hysterotomies performed for early repairs of congenital diaphragmatic hernia, or more recently for closure of neural tube defects, have been associated with high rates of preterm rupture of membranes and preterm delivery38. Even with the introduction of laparoscopic techniques, such as utilized for the studies of tracheal plugging in congenital diaphragmatic hernia, preterm delivery is essentially inevitable39. Additionally, maternal pulmonary edema and respiratory compromise, often encountered while attempting to arrest preterm delivery, have been reported. While percutaneous needle placements are associated with the lowest risk of preterm labor, abruption and premature rupture of the membranes remain potential complications. In order to accomplish a percutaneous technique under ultrasound imaging, fetal manipulation is typically needed. While often performed in the third-trimester fetus as external version to assure a vertex presentation, abruption is a present though rare risk. Whenever the amniotic cavity is entered, premature rupture of the membranes is a risk and, based on similar percutaneous fetal needle procedures, is likely to occur in 2–7% of cases. In 2000 a collaborative team of perinatolgists, cardiologists, radiologists, fetal surgeons and anesthesiologists was assembled to establish a protocol for the in-utero balloon dilation of critical aortic stenosis. One focus was accurate identification and careful case selection of the fetus earlier in gestation than previously reported (< 30 weeks). Maternal risks were to be minimized through the use of ultrasound guidance without hysterotomy or fetoscopy. Additionally, the knowledge gained from two decades of fetal surgery was incorporated into the fetal and maternal anesthesia plans. A primary endpoint of a physiologic two-ventricle circulation at 1 month of age was sought. In 2004 we published our experiences to date40 (Figures 2 and 3). Twenty-four women were offered intervention following individual case review by the Institutional Review Boards at two hospitals (Brigham and Women's Hospital and Children's Hospital, Boston, MA). Gestational age at diagnosis ranged from 19 to 26 weeks and the mean gestational age at intervention was 23.9 ± 2.3 weeks. Twenty women consented to proceed with attempted aortic stenosis dilation; the intervention was declined by four of them. Technical success was defined by ultrasound visualization of balloon insufflation across the aortic annulus and increase in postdilation aortic valve flow. In 14 of the 20 fetuses, technical success was achieved. Among those with technically successful procedures, ongoing left-heart growth was documented by multidimensional measurements of the left ventricle. Of these 14 fetuses with technically successful interventions, three had two-ventricle physiologies at 1 month of age, and two additional infants underwent surgical repair but have maintained two-ventricle circulation. Two fetuses were lost postprocedure and the remaining seven fetuses developed HLHS. Of the six fetuses with technically unsuccessful interventions, and in those who declined, serial echocardiograms noted arrest of left ventricle growth and development of HLHS. Schematic diagram of ideal approach to fetal left ventricle. Reproduced from Tworetzky W et al. Balloon dilation of severe aortic stenosis in the fetus: potential for prevention of hypoplastic left heart syndrome: candidate selection, technique, and results of successful intervention. Circulation 2004; 110(15): 2125–213140. Sequential images demonstrating aortic stenosis dilation using balloon insufflation in a fetus at 19 weeks' gestation. (a) Preprocedure with dilated left ventricle and diminished flow across the aortic valve; (b) 18-gauge styleted needle advance percutaneously, across amniotic fluid and into left ventricle; (c) following removal of stylet, premeasured soft tip wire advanced through needle and across aortic valve; (d) coronary balloon threaded over wire across valve and insufflated; (e) postprocedure increased aortic jet across dilated valve. In summary, viewed as an intention to treat, five of the 20 second-trimester fetuses with a high likelihood of HLHS delivered with two-ventricle circulation following a mid-trimester attempt at aortic valve balloon dilation. This initial series provides valuable information for further investigations. The option for laparotomy, as opposed to solely a percutaneous approach, as determined by the operative team to enhance imaging resolution and fetal positioning, significantly increased technical success. Early identification and referral of candidates are to be encouraged as progression from ventricle dilation to constriction was seen to occur quickly in the mid-second trimester. Furthermore, the criteria applied for consideration of fetal intervention were highly predictive of the evolution to HLHS, as all untreated (declined or unsuccessful procedures) progressed to HLHS. Previously described technical difficulties with balloon rupture were decreased by refinements adapted from newborn cardiac catheterizations. Imaging remains problematic, with three- and four-dimensional ultrasonography being explored. Direct intracardiac imaging through a miniaturized scope has been proposed by other investigators41. While challenges remain in the selection of fetuses, and in imaging and fetal positioning refinements in order to avoid the additional morbidity of maternal laparotomy, this initial experience reopens fetal cardiac disease as a potential candidate for minimally invasive, in-utero treatment.
DOI: 10.1002/uog.2595
Institutions: Brigham and Women's Hospital, Boston Children's Hospital
Language: English
Subfield: Epidemiology
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2063491021