REVIEWS Prenatal screening for structural congenital heart disease Lindsey E. Hunter and John M. Simpson Abstract | Congenital heart defects can be diagnosed during fetal life using echocardiography. Prenatal diagnosis allows full investigation of affected fetuses for coexisting abnormalities, and gives time for parents to be informed about the prognosis of the fetus and treatments that might be required. In a minority of cases, where the natural history suggests an unfavourable outcome, prenatal diagnosis provides an opportunity for fetal cardiac intervention. For some cardiac lesions, notably hypoplastic left heart syndrome, transposition of the great arteries, and coarctation of the aorta, prenatal diagnosis has been shown to reduce postnatal morbidity and mortality. Some costs of care, notably the transport of critically ill infants, are reduced by prenatal diagnosis. Prenatal screening programmes typically recommend detailed assessment of fetuses judged to be at high risk of congenital heart disease. However, most cases of congenital heart disease arise in the low-risk population, and detection of affected fetuses in this setting depends on recognizing abnormalities of the heart during the midtrimester scan. Evidence supports the use of structured training interventions and feedback to those undertaking sonographic examinations, to improve the prenatal detection of congenital heart disease. Hunter, L. E. & Simpson, J. M. Nat. Rev. Cardiol. 11, 323–334 (2014); published online 25 March 2014; doi:10.1038/nrcardio.2014.34

Introduction

Fetal Cardiology Unit, Department of Congenital Heart Disease, Evelina London Children’s Hospital, London SE1 7EH, UK (L.E.H., J.M.S.). Correspondence to: J.M.S. john.simpson@ gstt.nhs.uk

Congenital heart disease (CHD) is the most-common group of malformations affecting fetuses and new-born infants. The incidence of moderate-to-severe CHD has been estimated to be 5–6 per 1,000 live-born infants, but is higher in fetal life because some affected fetuses do not survive to full term, probably as a result of the complexity of the cardiac disease or associated abnormalities.1–3 Prenatal detection of CHD is a particular challenge, because both the right and left sides of the heart support the systemic arterial circulation, atrial and arterial shunts are present, and the placental circulation provides both oxygenation and nutritional support for developing fetuses (Figure 1). Consequently, the majority of fetuses affected by CHD do not show overt signs of cardiac failure during fetal life, because one side of the heart can compensate for an abnormality on the other side. For example, in hypoplastic left heart syndrome (HLHS), systemic output can be maintained by left-to-right atrial shunting and patency of the arterial duct, coupled with high pulmonary vascular resistance allowing the right ventricle to pump blood to the systemic arterial circulation. Consequently, identification of affected fetuses depends either on detection during routine sonographic examination or by identifying factors that put a fetus at an increased risk of CHD. In this Review, we examine current screening strategies for prenatal diagnosis of CHD, including individuals judged to be at either high or low risk of cardiac Competing interests The authors declare no competing interests.

abnormalities. The ease with which various forms of CHD can be diagnosed prenatally is discussed, along with the effect of prenatal diagnosis on outcome. For any prenatal screening programme, diagnostic accuracy, potential for prenatal intervention, and universal applicability are important considerations and are also addressed in this Review.

Detecting CHD Identifying high-risk fetuses Risk factors for CHD can be subdivided into those directly affecting a fetus, for example increased nuchal translucency (NT); maternal factors, for example exposure to teratogenic medications; and ‘historical’ factors, such as a history of CHD in a first-degree relative. The relationship between CHD and maternal drug therapy can be confounded by the potential for the underlying disease state to predispose to CHD, rather than the drug itself. Modifiable risk factors for CHD are addressed in a joint position statement from the AHA and the American Academy of Pediatrics, published in 2007.4 The authors recommended periconceptual multi­vitamins (including folate) potentially to reduce the risk of CHD, and for pregnant mothers to seek medical advice before any drug ingestion. Fairly good agreement exists between professional bodies as to the risk factors that merit detailed assessment of the fetal heart (Box 1).5–9 NT (also known as the nuchal fold) is the sonographic appearance of the fluid-filled space at the back of the fetal neck, and is measured between 11 weeks and 13 weeks and 6 days. The NT measurement was

NATURE REVIEWS | CARDIOLOGY

VOLUME 11  |  JUNE 2014  |  323 © 2014 Macmillan Publishers Limited. All rights reserved

REVIEWS Key points ■■ Prenatal diagnosis of congenital heart disease (CHD) is important for investigation of affected fetuses for comorbidities, prognostication, preparation for postnatal management, and parental choice about continuation of pregnancy ■■ Historical risk factors, such as a family history of CHD, or fetal risk factors, including increased nuchal translucency, warrant investigation with detailed fetal echocardiography ■■ Most fetuses with CHD present in the ‘low-risk’ population, and prenatal detection depends on recognizing abnormalities during obstetric scans ■■ Cardiac anomalies characterized by an abnormal four-chamber view of the heart have a higher detection rate than those in which the abnormality is primarily of the outflow tracts ■■ For particular cardiac lesions, including transposition of the great arteries, coarctation of the aorta, and hypoplastic left heart syndrome, prenatal diagnosis improves outcomes ■■ Prenatal diagnosis of CHD allows the preparation of postnatal intervention in most instances; in a minority of cases (mainly critical left-heart lesions), fetal cardiac intervention can be considered

originally instigated to identify fetuses at high risk of trisomy 21 (also known as Down syndrome), but its association with fetal CHD was quickly recognized and found to be independent of the fetal karyotype.10 The relationship between NT and CHD demonstrates that a correlation exists between the NT value and the risk of a

Ductus arteriosus

cardiac disease, rather than a specific threshold or cut-off value.11,12 The initial data suggested that the majority of CHD could be detected by using the 95th percentile of NT as a threshold for detailed fetal echocardiography.10 The logistical implication of offering fetal echocardio­ graphy in all pregnancies in which the NT is >95th percentile would be substantial, owing to the sheer number of pregnancies that would subsequently be considered to be at increased risk. Even among fetuses with an NT >99th percentile (3.5 mm), the absolute incidence of CHD has been reported to be only 6–7%.10,11 Accepting that the relationship between the NT value and the risk of CHD is a continuum, the value of NT at which detailed fetal echocardiography should be prompted has been much debated, owing to the logistical and cost–benefit implications. Results of meta-analyses have not suggested as strong an association between NT and CHD as originally reported, with a detection rate of 30% when using NT as a screening tool.13 The results of one meta-analysis showed that, among chromosomally normal fetuses with CHD, 44% had an NT >95th percentile, and 20% were >99th percentile.14 Therefore, other screening markers have been proposed to refine risk stratification and increase the predictive value of the initial screening test. These markers include the presence of tricuspid valve

b

c

Pulmonary artery Aorta

Foramen ovale LA

RA

LV RV

Umbilical cord attaches to baby’s circulation To placenta

Figure 1 | Fetal and postnatal circulations. a | Fetal circulation. Oxygenated blood from the placenta returns to the RA via the umbilical vein and ductus venosus. Blood passes from the RA to the LA via the patent foramen ovale. The second fetal shunt occurs between the main pulmonary artery and the descending aortic arch in the form of the arterial duct (ductus arteriosus). Fetal PVR is high compared with in the postnatal circulation. Therefore, most blood being pumped out of the RV passes through the arterial duct into the descending aorta. b | Immediate postnatal circulation. Physiological adaptations that occur after birth result in a marked drop in PVR and an increase in the pulmonary venous return to the left atrium. This change results in functional closure of the foramen ovale. A postnatal increase in systemic vascular resistance and arterial pressure occurs, coupled with the fall in PVR. Consequently, blood flows from the aorta to the pulmonary circulation via the arterial duct until it constricts and finally closes after birth. c | Normal postnatal circulation. Deoxygenated blood passing through the right-heart structures is completely separated from the oxygenated blood passing through the left-heart structures. The foramen ovale and arterial duct have closed, preventing shunting between the atria and the great arteries, respectively. Abbreviations: LA, left atrium; LV, left ventricle; PVR, pulmonary vascular resistance; RA, right atrium; RV, right ventricle.

324  |  JUNE 2014  |  VOLUME 11



www.nature.com/nrcardio © 2014 Macmillan Publishers Limited. All rights reserved

REVIEWS Box 1 | Indications for detailed fetal echocardiography Fetal ■■ Suspicion or detection of congenital heart disease on obstetric anomaly scan ■■ Increased nuchal translucency thickness in first trimester: >99th percentile (>3.5 mm) or >95th percentile (>2.2–2.6 mm) according to crown–rump length ■■ Major extracardiac abnormality, such as congenital diaphragmatic hernia, exomphalos, duodenal atresia, or cystic hygroma ■■ Fetal hydrops ■■ Fetal arrhythmias, including fetal tachycardia, bradycardia, or ectopic beats ■■ Abnormal fetal karyotype, such as trisomy 13 (Patau syndrome), trisomy 18 (Edwards syndrome), trisomy 21 (Down syndrome), or monosomy XO (Turner syndrome) ■■ Monochorionic pregnancy, owing to risk of cardiac abnormality or twin–twin transfusion syndrome Maternal ■■ Maternal drugs: use of prostaglandin synthetase inhibitors, such as ibuprofen; risk of teratogenicity, such as from use of lithium or anticonvulsants ■■ Diabetes mellitus or other metabolic conditions, such as phenylketonuria ■■ Maternal infection, such as with Parvovirus ■■ Maternal antibody status or connective tissue disease, such as positive anti‑Ro, anti‑La antibodies Historical ■■ History of congenital heart disease in a first-degree relative Other ■■ Assisted conception or in vitro fertilization ■■ Increased risk of fetal heart failure, such as absent ductus venosus, fetal anaemia, or presence of fetal tumours with large vascular supply

regurgitation, and absence or reversal of the a‑wave flow in the ductus venosus.15,16 Studies have shown that an altered ductus venosus flow pattern is associated with a threefold increase in the incidence of CHD in chromo­ somally normal fetuses.17 These refinements have been introduced with the aim to reduce the number of false– positive screening tests, and also to minimize any adverse effect on the sensitivity of the test. Importantly, the addition of ductus venosus flow patterns and tricuspid valve regurgitation to refine screening has currently been adopted at a limited number of specialist centres, and is not universally implemented. First-trimester NT screening has been used to identify fetuses at risk of CHD, even though the primary intention was to use this technique to identify fetuses with genetic abnormalities. In the past 3 years, non­ invasive screening techniques for trisomy 21, 13 (Patau syndrome), and 18 (Edwards syndrome) have become available by analysing cell-free DNA in maternal blood, and have increasingly been used to examine not only major trisomies, but also sex-chromosome abnormalities and microdeletion syndromes, such as 22q11 deletion. Although promising, these genetic techniques are fairly new, and current guidelines advise that they should be adjuncts to routine screening with ultrasonography and blood tests.18–20 The introduction of these techniques has substantial implications for screening for CHD, particularly early fetal echocardiography. If this form of testing proves to be both highly specific and sensitive, then the individuals presenting to fetal cardiologists will be moreeffectively risk stratified for major chromosomal abnormalities than is currently the case.21 Furthermore, fetal free-DNA techniques might mean that NT screening

is not undertaken at all. The incremental value of NT screening to identify fetuses at risk of isolated CHD would consequently be lost. A family history of CHD is one of the most-common reasons for referral for detailed fetal echocardiography. Most physicians accept that such a history in a firstdegree relative is a valid indication for specialist fetal cardiac examination. Screening of more-distant relatives is not usually undertaken, because the added risk beyond that in the general population is negligible, and would add substantially to the number of pregnancies that would need to be assessed at specialist centres. In the absence of a known syndrome or genetic association, the risk of CHD in such pregnancies has been reported to be 2–3%,22 and some studies have shown an increased risk if the mother is affected.23 At our own unit (Evelina London Children’s Hospital, UK), threefold more fetuses are referred because of maternal CHD than because of paternal CHD,22 despite a recurrence risk with either sex that is as high or higher than if a sibling has CHD.24 The precise recurrence risk should be adjusted according to the type of CHD, because studies have shown an increased recurrence risk for left-heart lesions and d­isorders of laterality.25–27

Screening in low-risk fetuses Despite the identification of risk factors for CHD, most cases occur in the population deemed to be at low risk. Diagnosis of CHD in this group depends on a sono­ grapher recognizing that the appearance of the heart deviates from normal. Debate exists as to which views of the fetal heart should be obtained to screen for CHD. The most-established screening view is the four-chamber view of the fetal heart. This view has the advantage of having external fetal reference points, because the plane of the fetal ribs matches that of the four-chamber view during fetal life (Figure 2b). Incorporating this view into obstetric anomaly scanning is supported by all the major international obstetric and paediatric cardiology societies and organizations. 5–9 Many cardiac lesions, for example HLHS, tricuspid atresia, and atrioventricular septal defects, have manifest abnormalities on the fourchamber view. Conversely, other major cardiac lesions have a normal or near-normal four-chamber view. Detection of these lesions, for example transposition of the great arteries, tetralogy of Fallot, and common arterial trunk (truncus arteriosus), require the cardiac outflow tracts to be accurately visualized (Box 2).28 In a major population-based series, the detection rate of lesions affecting only the cardiac outflow tracts was significantly lower than that of lesions detected by an abnormal four-chamber view.29 Prioritizing the training of sonographers to recognize abnormalities of the outflow tracts has enhanced the detection rate, but it remains suboptimal.30,31 Given that the cardiac outflow tracts pose the greatest difficulty for a screening sono­ grapher to recognize an abnormality, the challenge has been to develop a systematic approach to screening views that can be obtained in the vast majority of fetuses, that are achievable within the time constraints of the obstetric

NATURE REVIEWS | CARDIOLOGY

VOLUME 11  |  JUNE 2014  |  325 © 2014 Macmillan Publishers Limited. All rights reserved

REVIEWS a

b L

A L

P

R

A R

P IVC

Left atrium

Left ventricle

A L

R

Right ventricle

P

Spine

Aortic valve Right ventricle Right atrium

Descending aorta

Stomach

c

Left ventricle

Spine

LVOT

Spine Right atrium

e

d A

R

A

R

L

P

L

P

Aorta

Right ventricle

Aortic arch

SVC Main PA Duct

SVC Trachea

Right PA

Spine

Pulmonary valve Arterial duct

Spine

Figure 2 | Sonographic screening for congenital heart disease. This series of images shows the near-transverse cuts of the fetal thorax that can be performed as a sonographic, inferior-to-superior sweep, which allows the rapid identification of the major cardiac chambers and vessels. a | View of cardiac situs just below the four-chamber view, showing the fetal stomach and aorta to the left and the IVC anterior and to the right. b | Four-chamber view of the fetal heart, showing a single rib around the fetal thorax. The left-sided and right-sided chambers are balanced in size. c | View of the LVOT and aortic valve leaving the left ventricle. d | View of the PA passing in an anteroposterior direction. The right PA can be seen to arise from the main PA. The aorta and SVC are seen in short axis to the right of the PA. e | Three-vessel view of the upper mediastinum, showing the PA leading into the arterial duct alongside the transverse aortic arch. The SVC is to the right, and the trachea is to the right of the V shape formed by the aortic arch and arterial duct. Abbreviations: IVC, inferior vena cava; LVOT, left ventricular outflow tract; PA, pulmonary artery; SVC, superior vena cava.

anomaly scan, and that can be used to detect most of the major abnormalities. Current recommendations propose an approach based on transverse or near-transvers­e sonographic cuts of the fetal thorax (Figure 2).5–9 This approach produces images of the fetal heart that are similar to those obtained using CT or MRI. An advantage of this method is that a sonographer can acquire the necessary anatomical cuts by using the four-chamber projection as a reference point, and subsequent images are obtained by cranial or caudal angulation of the probe. Despite the introduction of this approach, inter­ national population-based series have shown that the accurate prenatal detection of CHD remains sub­optimal. The reasons for this shortcoming are complex. In some pregnancies, the fetal position or maternal habitus might make high-quality imaging of the fetal heart very difficult, and technical factors, such as the quality of the imaging equipment, can also influence accuracy. To add to the difficulties associated with body habitus, evidence shows an increased incidence of CHD in pregnancies associated with an elevated maternal BMI.32 The training and clinical experience of screening sonographers is central to the accurate prenatal recognition of CHD.33–35 Therefore, sonographers must be encouraged and supported by physician colleagues to continue to enhance 326  |  JUNE 2014  |  VOLUME 11

their skills and professional development.36,37 Failure to detect a cardiac abnormality antenatally can have important implications for postnatal management, particularly for duct-dependent lesions, because clinical deterioration can occur during the transition to the postnatal circulation (Figure 1b). If a cardiac abnormality is not detected, associated anomalies, such as chromosomal abnormalities, can remain undiagnosed before birth. These factors have important clinical and medicolegal implications. Data have also shown that prenatal diagnosis of CHD is cost-effective in avoiding the costs of transporting sick new-born infants who are diagnosed after birth with severe CHD to a tertiary cardiac centre.38 An important consideration for a CHD screening programme in the general population is a central record of cases of CHD. Such a registry allows the prenatal diagnostic rates of CHD to be calculated in a region or country and, to some extent, in screening centres or districts. Unfortunately, most countries fall short of this ideal. In the UK, for example, a central agency, the National Institute for Cardiovascular Outcome Research (NICOR), records all infants who undergo surgery in infancy. However, this registry excludes cases where the pregnancy does not continue to delivery, where pre­ operative death occurs, and where an infant is not



www.nature.com/nrcardio © 2014 Macmillan Publishers Limited. All rights reserved

REVIEWS deemed a surgical candidate owing to clinically significant comorbidities. Consequently, major problems of selection bias in the recording of cases exist. Investigators in one study painstakingly linked all data from maternity records and NICOR results to gauge the performance of screening centres in the detection of CHD.39 Obtaining such information prospectively in a systematic manner has the potential to identify units where specific training initiatives could be undertaken.39

Detailed fetal echocardiography

Referral to fetal cardiology or fetal medicine sub­ specialists for further assessment is warranted when a fetus is considered to be at high risk of developing CHD, or when a cardiac abnormality is suspected during a routine anomaly scan. Recommended standards for detailed fetal echocardiography have been published by both European and US professional bodies.6,8 Detailed fetal echocardiographic assessment involves not only comprehensive imaging of the cardiac connections by a sequential segmental approach, but further anatomical assessment in the form of longitudinal views of the ductal and aortic arch, bicaval views, and short-axis views of the heart. Functional assessment in the form of pulsed-wave and colour-flow Doppler imaging is also undertaken. In specialist units, a high degree of diagnostic accuracy can be expected, and large series confirm that virtually every form of CHD has been reported during fetal life.40,41 Particular cardiac lesions, however, remain a challenge to diagnose antenatally, even in specialist units, and these lesions merit special consideration.

Coarctation of the aorta Coarctation of the aorta is most-commonly suspected prenatally because of an asymmetry between the size of the right ventricle and pulmonary artery compared with the left ventricle and aorta, with dominance of right-heart structures. During fetal life, the duct is patent, which means that high Doppler-derived pressure drops are not observed in the aortic arch, in contrast to the postnatal situation. A coarctation ‘shelf ’ (a posterior indentation of the aortic arch in the region where the arterial duct inserts into the aortic arch) is observed in the longitudinal view of the aortic arch in some, but not all, fetuses. One series suggested additional value of Doppler investi­gation of the aortic isthmus to check for continuous flow during systole and diastole.42 A high suspicion of coarctation is suggested antenatally by the presence of right and left ventricular asymmetry or a discrepancy in the calibre of the aortic arch and pulmonary artery.43,44 Accurate measurement of the aortic arch, in the longitudinal or three-vessel tracheal view, and of the diameter of the aortic isthmus can help with the antenatal diagnosis of coarctation of the aorta, including transformation of a measurement into a size-specific or gestation-specific Z‑score.42,45,46 Z‑scores are an expression of how many standard deviations above or below a population mean a given observation lies. This scoring system is particularly helpful given the increase in fetal size with advancing gestational age.47 Management of coarctation of the aorta

Box 2 | Detection of cardiac abnormalities Using the four-chamber view Septal defects ■■ Atrioventricular septal defect ■■ Large ventricular septal defects (inlet types) Left-heart anomalies ■■ Hypoplastic left heart syndrome ■■ Critical aortic stenosis ■■ Severe coarctation of the aorta Right-heart abnormalities ■■ Tricuspid atresia ■■ Pulmonary atresia with intact ventricular septum ■■ Ebstein anomaly of the tricuspid valve Double inlet ventricles ■■ Double inlet left or right ventricle Using extended views of the outflow tracts* ■■ Transposition of the great arteries ■■ Tetralogy of Fallot with or without pulmonary atresia ■■ Common arterial trunk ■■ Some forms of double outlet right ventricle ■■ Some forms of coarctation of the aorta *Normal or near-normal four-chamber view.

can range from limited surgery to resect discrete obstruction of the aortic arch through to a scenario in which the left-heart structures are so hypoplastic that management is similar to that for HLHS, in which a normal biventricular circulation cannot be achieved. In this scenario, management is a staged surgical palliation involving connection of the systemic veins to the pulmonary arteries, and the dominant right ve­ntricle supports the systemic arterial circulation. Despite normative fetal values and recommendations to assess fetuses suspected of having coarctation of the aorta, an overlap remains between cases with a confirmed postnatal diagnosis of coarctation of the aorta and false–positive cases. Owing to this uncertainty, a policy of scanning neonates with a prenatal suspicion of coarctation is established, to assess whether aortic arch obstruction has developed during closure of the arterial duct. Even this policy has limitations, because coarctation has been reported in infants with prenatally suspected coarctation several months after closure of the arterial duct.48

Total anomalous pulmonary venous drainage Another important and challenging antenatal diagnosis is isolated total anomalous pulmonary venous drainage (TAPVD). Pulmonary vascular resistance is higher in the fetal circulation than in the postnatal circulation and, therefore, colour-flow Doppler visualization of antegrade flow from the pulmonary veins directly into the left atrium can be more difficult to demonstrate. In prenatal TAPVD, the pulmonary venous confluence lies in close proximity to the posterior wall of the left atrium and can give a falsely reassuring impression of normality. Volume loading of the right heart might be thought to be always evident in TAPVD, but is actually a sign of this condition that appears late in fetal life.49,50 A multicentre study has shown that only a minority of cases of isolated TAPVD were accurately diagnosed, even when assessment was

NATURE REVIEWS | CARDIOLOGY

VOLUME 11  |  JUNE 2014  |  327 © 2014 Macmillan Publishers Limited. All rights reserved

REVIEWS undertaken at a specialist unit.51 Ideally, investigation of the pulmonary veins should involve visualization by 2D and colour-flow imaging with appropriate reduction of colour Doppler scales to detect low-velocity flow. This modality can be coupled with pulsed Doppler imaging to check for a normal phasic flow pattern, which is f­requently lost in TAPVD.

Valve stenosis Disease progression is another important consideration that affects particular cardiac lesions, and complicates the diagnosis of these abnormalities in the second trimester, when most anomaly scans are performed. In particular, aortic valve and pulmonary valve stenosis can progress with advancing gestational age. Aortic and pulmonary valve flow velocities measured by Doppler imaging can be within normal limits or mildly elevated in early gestation, but subsequently progress to severe or even critical stenosis.52 Therefore, explaining to parents the limitations of fetal echocardiography in the second trimester is important. In practice, if suspicion of aortic or pulmonary valve stenosis exists, sequential fetal echocardio­ graphy will be undertaken to gauge any progression with advancing gestational age, and parental counselling might change according to disease progression. Ventricular septal defects Ventricular septal defects are much more of a challenge to diagnose prenatally than postnatally, owing to both anatomical and physiological factors. During fetal life, the pressures in the left and right ventricles are similar.53 If a ventricular septal defect is revealed, colour-flow Doppler typically shows bidirectional flow at low velocity, rather than the more-obvious left-to-right shunting noted after birth as the pulmonary vascular resistance falls. Anatomically, the membranous portion of the fetal ventricular septum is very thin. In the small fetal heart, a membranous ventricular septal defect is often suspected when none is actually present and, conversely, defects in this region are easily overlooked.54 The location of the ventricular septal defect can also affect the risk of associated fetal chromosomal anomalies, with a higher incidence of anomalies associated with membranous than with muscular defects.54,55 Timing of assessment The timing of detailed fetal cardiac assessment is itself being influenced by the techniques used to identify high-risk groups. For example, if NT is increased at 10–12 weeks of gestation, parents might opt for an invasive test for fetal karyotyping. After karyotypic abnormalities, CHD is the largest group of fetal diseases asso­ciated with increased NT, which often prompts a referral for early fetal echocardiography. This referral is performed with the aim to provide either an early diagnosis of CHD, or reassurance of normality. Technological advances have made early sonographic imaging of the fetal heart, by either a transvaginal or a transabdominal approach, increasingly feasible. Several series have shown that these early scans, when performed at specialist centres, 328  |  JUNE 2014  |  VOLUME 11

accurately detect major cardiac lesions.56–60 Even if the initial early assessment is deemed normal, most units electively repeat the scan subsequently during gestation, because minor lesions can become evident with advancing gestational age. Equally, disease progression, such as that of semilunar valve stenosis or cardiomyopathy, can occur. A major issue affecting early fetal echocardio­ graphy (gestational age 11–14 weeks) is the potential for diagnostic error, because the heart is very small and at the resolution limit of current ultrasound systems. Even if the heart can be confidently diagnosed as being abnormally formed, all the necessary prognostic information might not be obtainable at this early stage. In this setting, correlation of prenatal findings with postnatal or postmortem findings assumes an even greater importance than prenatal diagnosis in the midtrimester.21,61

Consequences of prenatal CHD diagnosis Parental perception Prenatal counselling after severe CHD has been detected can be a detailed, complex, and emotional process. Often, several counselling sessions are required to ensure that parents have appropriate insight, understanding, and acceptance of the prognosis and management of the cardiac condition and the effect of any comorbidities. This process allows parents to reach a decision that is both informed and appropriate to their family beliefs, circumstances, and life experiences. Parental perceptions and experiences of a prenatal diagnosis of CHD have been studied, showing a clear preference for comprehensive information, particularly pertaining to future quality of life.62 However, a marked variation has been shown to exist in the options offered to parents after a prenatal diagnosis of severe CHD, such as HLHS.63,64 Nevertheless, parental knowledge about CHD does seem to be better when a prenatal diagnosis has been made, compared with a diagnosis only after birth.65 Internetbased studies of prenatally diagnosed CHD have shown that the mention of termination of pregnancy after parents had declined this option made them less positive about prognosis, and that the parental perception of a cardiologist’s level of compassion was inversely linked to the likelihood of them seeking a second opinion.66,67 How parents perceive the prognosis and long-term quality of life associated with a congenital condition is unique to their own life experiences and often differs from that of health-care providers.68 Finally, the manner in which a diagnosis is initially presented to a family, the information provided, and how the family interprets the information are all factors that influence parental perception and subsequent decisions.67 Fetal and postnatal outcomes For a prenatal screening programme to be viable, evidence must exist for an effective intervention or treatment after diagnosis. CHD is known to be associated with other fetal anomalies, although the strength of the association varies according to the cardiac lesion.41 Some forms of CHD, for example atrioventricular septal defects,69 and tetralogy of Fallot,70 have a very strong



www.nature.com/nrcardio © 2014 Macmillan Publishers Limited. All rights reserved

REVIEWS association with chromosomal abnormalities, including trisomy 21 or chromosome 22q11 deletions. Therefore, prenatal detection of the cardiac defect can result in the identification of an associated extracardiac abnormality, which affects the fetal and postnatal prognosis. Consequently, full investigation of fetuses with CHD can involve multiple subspecialties, with the aim to identify all comorbidities. For many fetuses, parental decisionmaking about the pregnancy is more influenced by the extent of coexisting abnormalities than by the cardiac abnormality itself. Prenatal diagnosis of CHD gives parents the benefit of time before the birth of their child to understand the nature of the cardiac lesion, discuss the available treatment options, and understand the immediate and longterm prognoses.71,72 This time and preparation can assist parental decision-making and ensure that consent is as informed as possible before postnatal surgical or catheter interventions are undertaken. Prenatal diagnosis has been proposed to improve the ability of parents to adapt to caring for an infant through cardiac surgery and into later life, but evidence confirms that parents require continued support, regardless of when the ­diagnosis is made.73 Fundamental to the justification of prenatal CHD screening is whether detection improves the outcome of affected infants. However, designing a study to address this crucial question is extremely difficult. Midtrimester anomaly scanning might have a tendency to detect fetuses with a screening view of the heart that is particularly abnormal, whereas the postnatal cohort would contain all neonates with the condition, thereby introducing a major potential for selection bias. Therefore, even for a particular form of cardiac lesion, a strict comparison between prenatal and postnatal diagnosis might not be valid. For some cardiac lesions, observational evidence indicates that prenatal diagnosis improves post­ natal outcomes. The types of CHD studied have been those in which unrecognized conditions present acutely in the early postnatal period after constriction of the arterial duct or closure of the foramen ovale. HLHS was one of the first lesions to be studied. One large series showed reduced morbidity and mortality for infants diagnosed prenatally compared with those diagnosed postnatally.74 However, results from other studies have not confirmed a survival advantage, although the clinical condition at presentation was improved.75,76 An important aspect of prenatal diagnosis that remains under-reported compared with mortality, is quality of survival and, in particular, neurological morbidity. A prenatal diagnosis of HLHS has been associated with a reduction in early neonatal neurological morbidity when directly compared with infants diagnosed postnatally.77 This outcome is not unexpected, given the severe circulatory collapse that can occur after postnatal ductal constriction in neonates with HLHS. Coarctation of the aorta is another lesion in which study data suggest that prenatal diagnosis benefits neonatal outcomes.78 An important aspect of this study was the inclusion of infants diagnosed at autopsy who had

not presented to a cardiac centre during life. Although this group might be small, such infants are often not included in studies in which the cohort is defined by presentation to a tertiary cardiac centre. The most-compelling evidence for the effect of prenatal diagnosis of CHD on outcome is the detection of transposition of the great arteries. Postnatally, adequate systemic arterial oxygen saturations are achieved if sufficient mixing of blood occurs at the atrial septum, arterial duct, or both. A landmark study showed significantly reduced preoperative and postoperative mortality in infants with a prenatal diagnosis.79 Subsequently, the same investigators described the prenatal sonographic findings associated with transposition of the great a­rteries that might predict the need for early neonatal septo­stomy, including restriction of the foramen ovale and constriction of the ductus arteriosus.80 Researchers in some studies have examined the effect of prenatal diagnosis on the duct-dependent pulmonary circulation. Although prenatal diagnosis improves oxygen saturations at presentation, no effect on shortterm morbidity or mortality was observed.81 Particular cardiac conditions, such as pulmonary obstruction, are most-easily identified in the perinatal period by the presence of cyanosis, which can be detected clinically or by pulse oximetry.82,83

Risk stratification of fetuses with CHD After a prenatal diagnosis of CHD has been made, the timing, mode, and site of management should be optimized to improve postnatal outcomes. Important considerations include the circulatory changes that occur in the transitional circulation, such as closure of the foramen ovale, constriction of the arterial duct, fall in pulmonary vascular resistance, and increase in systemic vascular resistance (Figure 1). Particular cardiac lesions, for example an isolated ventricular septal defect, would not be expected to cause haemodynamic compromise im­mediately and, therefore, management can be considered to be fairly routine. Conversely, fetuses with HLHS and a restrictive atrial septum might be critically ill immediately after delivery, owing to the oxygenated pulmonary venous blood being obstructed from leaving the left atrium and failing to reach the systemic arterial circulation. The arrangements for management are tailor­ed according to the type of cardiac lesion and expected initial postnatal problems. Examples of cardiac lesions for which urgent postnatal cardiac intervention or high‑level intensive care are anticipated are shown in Table 1. So far, we have discussed risk stratification according to the type of cardiac lesion. However, within each group of cardiac lesions, a spectrum of severity often exists, and each diagnostic group cannot be regarded as homogeneous. Increasingly, the role of fetal echocardiography is to identify both positive and negative prognostic factors within each subset of cardiac lesions. The group of lesions for which such stratification has been most-extensively studied is left-heart disease, particularly HLHS and aortic valve stenosis. In HLHS, a number of important adverse prognostic factors exist, including tricuspid valve

NATURE REVIEWS | CARDIOLOGY

VOLUME 11  |  JUNE 2014  |  329 © 2014 Macmillan Publishers Limited. All rights reserved

REVIEWS Table 1 | Cardiac lesions and situations requiring immediate perinatal attention Cardiac lesion

Potential complication

Potential interventions

Simple transposition of the great arteries (associated with a restrictive atrial septum and arterial duct)

Metabolic acidosis Profound hypoxaemia

Urgent balloon atrial septostomy

Hypoplastic left heart syndrome or critical aortic stenosis (associated with an intact or restrictive atrial septum)

Profound hypoxaemia Metabolic acidosis

Urgent balloon atrial septostomy or surgical septectomy

Obstructed total anomalous pulmonary venous drainage

Profound hypoxaemia

Early surgical intervention

Absent pulmonary valve syndrome

Airway compression Respiratory compromise or failure Air trapping

Mechanical ventilation (positive pressure) Plication of pulmonary arteries

Severe Ebstein anomaly of the tricuspid valve

Pulmonary hypoplasia Respiratory failure Severe hypoxaemia Management of hydrops

Oxygen therapy Mechanical ventilation (positive pressure) Inhaled nitric oxide Drainage of associated pleural effusions or ascites

Complete heart block with or without congenital heart disease

Cardiac failure Hydrops

Medical therapy with chronotropic agents Temporary cardiac pacing Drainage of associated pleural effusions or ascites

regurgitation, mitral valve regurgitation, and restriction of the atrial septum.84–86 A means of assessment based on the pulmonary venous Doppler waveforms has been proposed to predict the need for early postnatal intervention of the atrial septum.87 Disease progression is known to occur in fetal aortic valve stenosis, which is characterized by failure of left-heart structures to grow, and deterioration of left ventricular function.88 In a subset of fetuses, the progression is so marked that postnatal management is similar to that for HLHS, rather than a biventricular circulation. Echocardiographic markers, including monophasic mitral valve inflow, reversal of flow in the aortic arch during systole, and left-to-right atrial shunting of blood, can be used to predict dis­ ease progression.89 The use of fetal echocardiographic markers of disease severity is not restricted to left-heart disease and has been extended to other lesions, including pulmonary atresia with intact ventricular septum,90,91 Ebstein anomaly,92 and tetralogy of Fallot.70 Therefore, in current practice, effective prenatal screening for CHD not only leads to an accurate diagnosis, but also assists with perinatal management planning and allows tailored counselling that takes individual prognostic factors into account. Good communication between all the relevant services is essential and, in most centres, includes multidisciplinary meetings to discuss individual patients and formulate an appropriate management plan.

Fetal intervention An important facet of any screening programme for fetal cardiac anomalies is whether intervention can be offered after diagnosis. The strongest evidence for the benefit associated with prenatal intervention is in the treatment of fetal arrhythmias, particularly fetal tachycardias.93–95 Fetuses with other arrhythmias, notably immunemediate­d atrioventricular block, can also benefit from therapy with corticosteroids or salbutamol to reduce inflammation and accelerate fetal heart rate, but patient selection in this setting remains controversial.96–98 330  |  JUNE 2014  |  VOLUME 11

Fetal intervention for structural heart lesions has been advocated for severe aortic valve stenosis, severe pulmonary valve stenosis, and HLHS with an intact atrial septum. In a number of cases with severe fetal aortic valve stenosis, postnatal biventricular repair is precluded, owing to the subnormal growth and function of the left heart with advancing gestational age.88 Prenatal aortic valvuloplasty has been shown to be technically feasible and has been associated with improved growth of the aortic valve, ascending aorta, and mitral valve during fetal life.89,99,100 Despite these advances, and even after technically successful prenatal intervention, a substantial proportion of fetuses still have left-heart structures that are insufficiently developed to support the systemic arterial circulation and, therefore, are managed with a staged surgical palliation towards a single-ventricle c­i rculation. 101–103 Fetuses who achieve a biventricular circulation after prenatal intervention commonly require further surgery on the mitral valve, aortic valve, and resection of endocardial fibroelastosis, and continue to have evidence of diastolic dysfunction.104,105 Debate, therefore, remains as to whether aggressive pursuit of a biventricular circulation is always preferable to accepting a single-ventricle physiology.106 Intrauterine perforation of the pulmonary valve has been undertaken, mainly in fetuses with incipient fetal hydrops, where a poor outcome was anticipated without intervention.107 After initial reports of technical success, other groups have attempted prenatal intervention for severe right ventricular outflow tract obstruction, with improved growth of the right heart compared with controls.108 The number of fetuses who have undergone intervention for severe right ventricular outflow tract obstruction is far fewer than for critical aortic stenosis, most likely because the postnatal outcome of this con­ dition, even if untreated during intrauterine life, is thought to have a better prognosis than with HLHS. Fetuses with HLHS and an intact or near-intact atrial septum are an extremely high-risk group during postnatal



www.nature.com/nrcardio © 2014 Macmillan Publishers Limited. All rights reserved

REVIEWS management. Early neonatal hypoxia, metabolic acidosis, and circulatory failure can be anticipated, which prompts emergency surgery or catheter intervention. Additionally, restriction of the atrial septum is a major risk factor for achieving a viable total cavopulmonary circulation,109 possibly because of the intrauterine development of pulmonary vascular disease and lymphangiectasia secondary to obstruction of pulmonary venous return.110 Prenatal intervention on the atrial septum has been technically successful, but a proportion of septal perforations have been reported to reseal,111,112 prompting the use of stents to maintain septal patency.112,113 Fetuses who have had a technically successful intervention tend to have an improved outcome, but long-term data are lacking.114 Although the results of fetal intervention are important, the number of fetuses who are candidates for such interventions is a small minority (

Prenatal screening for structural congenital heart disease.

Congenital heart defects can be diagnosed during fetal life using echocardiography. Prenatal diagnosis allows full investigation of affected fetuses f...
2MB Sizes 4 Downloads 3 Views