U L T R A S O U N DC L I N I C S

Ultrasound Clin 1 (2006) 273–291

273

Prenatal Diagnosis of CongenitalHeart Disease: Where Are We Now?Wesley Lee, MD

a,b,*, Christine H. Comstock, MDa,c

& Fetal cardiac screening & Sonographic detection of selected cardiac

& What influences the prenatal detection of

congenital heart disease?Factors influencing adequate examinationof the heart

Unrecognized abnormalitiesEvolution of cardiac abnormalitiesStandardized images may not demonstratethe anomalous heart

& Fetal cardiac screening guidelinesBasic examinationExtended basic examination

a Division of Fetal Imaging, 3601 W. Thirteen Mile Road,Wb Department of Obstetrics and Gynecology, Wayne Stac Department of Obstetrics and Gynecology, University o* Corresponding author. Department of Obstetrics anBeaumont Hospital, 3601 W. Thirteen Mile Road, Royal OE-mail address: wlee@beaumont.edu (W. Lee).

1556-858X/06/$ – see front matter © 2006 Elsevier Inc. All rightsultrasound.theclinics.com

anomaliesVentricular septal defectsAtrioventricular septal defectsHypoplastic left heart syndromeTricuspid valve abnormalitiesConotruncal abnormalitiesAortic coarctationTotal anomalous pulmonary venous return

& Summary& References

The antenatal detection of birth defects is animportant public health concern with significantclinical ramifications. In 2002, more than 28,000infants died within the first year of birth, with anoverall rate of 7.0 deaths per 1000 live births in theUnited States [1]. This mortality rate was largelyattributed to birth defects, of which congenitalheart disease has been a leading cause of relatedadverse outcomes [2]. Furthermore, data from theWorld Health Organization indicates that 42% ofinfant deaths were attributed to heart defects [3].Prenatal sonography has played an important

role for the timely detection of congenital heartdisease (CHD). Despite promising results of earlystudies, however, the efficacy of fetal cardiac screen-ing programs has been variably successful. Thisarticle provides an update regarding various factorsaffecting the prenatal identification of cardiac de-

fects and summarizes current guidelines for fetalheart screening during the second trimester of preg-nancy. The diagnostic imaging characteristics andclinical significance of selected fetal cardiac abnor-malities are also reviewed.

Fetal cardiac screening

Cardiac abnormalities occur with an estimated in-cidence of approximately 4–13 per 1000 live births[4–6]. Most of the affected children will be bornto mothers with no identifiable risk factors forCHD. Consequently, standardized approaches areneeded to screen low-risk populations for car-diac abnormalities.Prenatal cardiac screening was introduced in the

mid-1980s when the four-chamber view of theheart was incorporated into a routine obstetric scan

illiam Beaumont Hospital, Royal Oak, MI 48073, USAte University, Detroit, MI, USAf Michigan, Ann Arbor, MI, USAd Gynecology, Division of Fetal Imaging, Williamak, MI 48073.

reserved. doi:10.1016/j.cult.2006.01.005

274 Lee & Comstock

between 18 and 22 weeks, menstrual age [7].After a training period, several centers in the UnitedKingdom were able to detect 77% of all cardiacanomalies over a 2.5-year period [8]. Copel andassociates [9] also examined fetuses in 1022 preg-nancies and found 74 abnormal heart cases usingthe four-chamber view. They reported 92% sensitiv-ity and 99.7% specificity for the detection of CHD.Unfortunately, others have found a wide range ofdetection rates for the four-chamber view of theheart in unselected patient populations [10–15].Additional diagnostic benefit has been subsequentlydemonstrated for the inclusion of cardiac outflowtracts into routine screening examination of theheart [11,16,17].One of the most comprehensive studies regarding

fetal heart screening in a nonselected populationhas been recently completed. Tegnander and col-leagues [18] analyzed results of a fetal heart screen-ing program over a 10-year period in Norway. Theirstudy population included 29,460 gravidas, repre-senting 98% of their deliveries at a single institu-tion. Routine fetal examinations were performed atapproximately 18 weeks, and included the four-chamber view and outflow tracts. Beginning in1995, patients were also asked to return for anotherscan if a satisfactory four-chamber view had notbeen obtained during the initial visit. Heart defectswere retrospectively classified after delivery as criti-cal when surgery was likely to be required (eg,transposition of the great arteries, hypoplastic leftheart syndrome, atrioventricular septal defect, aor-tic coarctation, and large ventricular septal defects)[19]. This investigation identified 97 critical casesof CHD, of which 55 (57%) had been detectedbefore birth. Forty-four percent of affected fetuseshad isolated CHD and 38% had an abnormal kar-yotype. About one-half (48%) of the abnormalfetuses with ductal-dependent lesions were de-tected. Only 27% of infants born with criticalCHD were alive at 2 years after delivery.

What influences the prenatal detection ofcongenital heart disease?

Several factors affect the quality of a successful fetalcardiac screening program. These factors help toexplain why prenatal detection rates have variedso widely in the medical literature.Chaoui [20] has summarized key reasons as to

why the four-chamber view does not always pro-vide an optimal detection rate among various cen-ters. The reasons include inadequate examination,unrecognized abnormalities, evolution of cardiaclesions, and the inability for this view to detectspecific anomalies that are best identified fromother scanning planes.

Factors influencing adequate examination ofthe heart

Inadequate examination of the fetal heart can berelated to timing of scan, fetal position, imagequality, maternal wall thickness, and a history ofprior maternal surgical procedures.

Timing of second trimester fetal heart scansSatisfactory views of the fetal heart are typically ob-tained between 18 and 22 weeks, menstrual age. Aprospective and randomized study of 1206 womenfound an incomplete four-chamber view more fre-quently between 18 to 18.9 weeks (18.4%) as com-pared with scans occurring between 20 to 20.9 weeks(2.9%) (P<.001) [21]. Many heart structures canstill be satisfactorily visualized beyond this time ifthe fetal position is favorable. Many patients, how-ever, prefer to know about major cardiac defects atan earlier stage of pregnancy.

Fetal positionOptimal views of the fetal heart are often obtainedwhen the cardiac apex is pointing toward the trans-ducer [Fig. 1] [22]. Suboptimal views occur whenthe spine causes acoustic shadowing over cardiacstructures from a prone position. This situation canbe worsened in the presence of oligohydramnios.

Image qualityTechnical specifications of the ultrasound system(eg, beam former, transducer, display screen) canalso affect satisfactory visualization of the fetalheart. One should always adjust the lowest acous-tic power intensity settings (ie, thermal index andmechanical index) that provide satisfactory diag-nostic images from using output display standardsand the ALARA (As Low As Reasonably Achievable)principle [23].Image quality relies on the examiner’s efforts to

position the transducer in a manner that is mostsuitable for acquiring this information. The degreeof screen magnification, gain, and acoustic focusshould be optimized for the region of interest. Thetransducer’s field of view should be minimizedto improve frame rate (ie, temporal resolution). Acolor filter can be applied to improve contrast be-tween soft tissue borders. Frame persistence shouldnot be set too high. Images should be zoomed to fillat least one-half the display screen with the fetalheart. Harmonic imaging often improves the imagedisplay [24].Wavelength (sound velocity divided by fre-

quency) is an important concept that is used toexplain image resolution [25]. Contemporary ultra-sound systems usually produce images with axialresolution between 2 to 4 wavelengths and lateral

Fig. 1. Fetal position affects the cardiac screening examination. Satisfactory visualization of the fetal heartdepends on orientation of cardiac structures in relation to the maternal abdomen. The best views are oftenobtained with the cardiac apex pointing toward the anterior maternal abdominal wall (A–C ). Anatomicstructuresare usually not well visualized when the cardiac apex is oriented toward the posterior maternal abdominal wall(D–F ). (Adapted from Lee W. American Institute of Ultrasound in Medicine. Performance of the basic fetal cardiacultrasound examination. J Ultrasound Med 1998;17:601–7; with permission.)

275Prenatal Diagnosis of Congenital Heart Disease

resolution between 3 to 10 wavelengths. Depend-ing on the imaging system, the axial resolutionwould range between 0.6 and 1.2 mm at 5 MHz.At the same transducer frequency, lateral resolutionranges between 0.9 and 3.0 mm. Therefore, theultrasound beam resolution and angle of beaminsonation can have important ramifications foridentifying small structures such as 2-mm ventricu-lar septal defects. The highest transducer frequen-cies, typically 5 MHz and above, often provide theimage resolution necessary to resolve subtle cardiacdefects. This has to be balanced as a trade-offbetween image resolution and acoustic penetration.

Maternal wall characteristicsDeVore and colleagues [26] examined technicalfactors that influence imaging of the fetal heartduring the second trimester of pregnancy. Morethan 700 trimester pregnancies were analyzed toidentify independent risk factors that contributeto difficult cardiac screening examinations. Gesta-tional age, maternal adipose tissue thickness, andprior lower abdominal surgery were found as themost significant factors that were associated withpoor visualization of the fetal heart.

Unrecognized abnormalities

Unrecognized abnormalities are another reasonwhy the prenatal detection rates of CHD have var-ied so widely. Examiners should be familiar withdevelopment of the human heart and how to trans-late this information to clinical practice. Anatomicand molecular methods are clarifying new aspectsof cardiac development. For example, traditionalteaching has suggested that the heart initiallyforms from paired linear tubes that fuse and con-tain all major cardiac segments. More recent work

now indicates that the embryonic heart results froma modular process with initial development of aprimary cardiac crescent. Subsequent developmentof a second, more anterior heart field is responsiblefor the appearance of the right ventricular outflowtract [27]. Cook and associates [28] have nicelyreviewed cardiac development in the human fetusas it relates to the prenatal diagnosis of CHD.Education and training of health care profes-

sionals can improve the prenatal recognition ofCHD. As a minimal goal, examiners must under-stand how to acquire images from standardizedcardiac scanning planes to classify them into nor-mal or abnormal categories. Cardiac screening pro-grams can be further improved after continuoustraining of health care professionals based on feed-back, a low threshold for echocardiography refer-rals, and convenient access to fetal heart specialists[29,30]. As one example, Hunter and colleagues[31] reported a twofold increase in the detectionrate (17% to 36%) of major cardiac defects afterimplementing a targeted training program for fetalheart screening at 16 ultrasound centers in theUnited Kingdom.

Evolution of cardiac abnormalities

Evolution of cardiac lesions is another importantreason why these abnormalities are not alwaysdetected at the time of an ultrasound scan. An ob-servational investigation of 22,050 pregnantwomen (77.5% low-risk patients) was undertakenby dividing them into two groups: Group A - 6924with initial vaginal sonography at 13–16 weeks’gestation that were followed by abdominal scansat 20–22 weeks’; and Group B - 15,126 womenwho only had initial transabdominal scans at20–22 weeks’ [32]. Both groups were scanned dur-

Box 1: Common indications for fetalechocardiography

Maternal indicationsFamily history

1st degree relative of probandPre-existing metabolic diseases

diabetesphenylketonuria

Maternal infectionsparvovirus B19rubellacoxackie

Cardiac teratogen exposureretinoidsphenyltoincarbamazepinelithium carbonatevalproic acid

Maternal antibodiesanti-Ro (SSA)anti-La (SSB)

Fetal indicationsSuspected fetal heart anomalyAbnormal fetal karyotypeMajor extra-cardiac anomalyAbnormal nuchal translucency

≥3.5 mm before 14 weeks, menstrual ageAbnormal nuchal fold

≥6.0 mm: 15–20 weeks, menstrual ageAbnormal cardiac rate or rhythm

persistent bradycardiapersistent tachycardiapersistent irregular heart rhythm

276 Lee & Comstock

ing the third trimester and diagnoses were con-firmed after birth. Two experienced examiners con-ducted all ultrasound examinations.Congenital heart disease occurred in 168 infants

(Group A - 66 infants; Group B - 102 infants). Ofthe Group A fetuses, 42 (64%) heart anomalieswere detected at the first vaginal scan, and 11(17%) were subsequently identified during theabdominal study. Three additional anomalies (4%)were found during the third trimester exam and10 more (15%) were only detected after delivery.Group B fetuses had 80 (78%) cardiac malfor-mations identified at the time of their first abdomi-nal scan at 20–22 weeks’ gestation. An additional 7(7%) and 15 (15%) cases were identified duringthe third trimester and after delivery, respectively.Ten heart anomalies that were discovered duringthe third trimester included aortic stenosis (n=2),cardiac rhabdomyoma (n=2), subaortic stenosis(n=1), tetralogy of Fallot (TOF) (n=1), aortic coarc-tation (n=1), sealed foramen ovale (n=1), ven-tricular septal defect (n=1), and hypertrophiccardiomyopathy (n=1).Their results indicated that fetal heart anomalies

can vary in appearance throughout pregnancy.In this study, two experienced examiners, usingboth early vaginal and second-trimester abdominalscans, were unable to identify 20% of CHD cases.Of note, early vaginal scans of the fetal heart wereable to detect nearly two-thirds (64%) of abnormalhearts. This observation is consistent with the find-ing of others who have described early detection ofheart anomalies, especially when increased nuchaltranslucency is present [33–39]. Although secondtrimester fetal heart screening can often be com-pleted between 18 and 22 weeks’ gestation, manyanomalies can still be identified before this stageof pregnancy.

Standardized images may not demonstratethe anomalous heart

The inability of specific scanning planes for de-tecting all forms of CHD can be explained byconsidering fetal cardiac anatomy. Although thefour-chamber view across the fetal thorax can bequite informative, this two-dimensional scanningplane may not provide satisfactory views of asmall ventricular septal defect or conotruncalanomalies involving the more anterior ventricularoutflow tracts.

Fetal cardiac screening guidelines

The primary goal of cardiac screening is to identifywhich fetuses are likely to have CHD. Currentguidelines emphasize a ‘‘basic examination’’ usinga satisfactory four-chamber view of the heart. If

technically feasible, an ‘‘extended basic’’ examina-tion of the left and right ventricular outflow tracts isalso recommended [40–42]. Fetuses with suspectedanomalies should be referred for fetal echocardiog-raphy to assess the seriousness of the anomaly andthe likelihood of a ductal dependent lesion at birth.Common indications for fetal echocardiographyare summarized in Box 1 [43].

Basic examination

General considerationsThe ‘‘basic examination’’ requires specific sono-graphic criteria using an adequately visualizedfour-chamber view of the heart [Table 1; Fig. 2].This approach must not be mistaken for a simplecount of cardiac chambers. Cardiac rate (120 to160 beats/minute) and regular rhythm should beconfirmed, although mild fetal bradycardia cantransiently occur during the second trimester. Theheart normally fills no more than a third of thethoracic area at the level of the four-chamber view.A small layer of fluid (<2 mm) can appear aroundthe normal fetal heart, although this finding may

Table 1: Basic cardiac screening examination

General Normal cardiac situs, axis,and positionHeart occupies a third ofthoracic areaMajority of heart in left chestFour cardiac chambers presentNo pericardial effusion orhypertrophy

Atria Atria approximately equal in sizeForamen ovale flap in left atriumAtrial septum primum present

Ventricles Ventricles about equal in sizeNo cardiac wall hypertrophyModerator band at rightventricular apexVentricular septum intact (apexto crux)

AV Valves Both atrioventricular valves openand move freelyTricuspid valve leaflet inserts onseptum closer to the cardiac apexas compared to the mitral valve

Adapted from [22]; with permission.

277Prenatal Diagnosis of Congenital Heart Disease

be clinically significant in the presence of cardiacfailure, other structural anomalies, or hydrops[44,45].Cardiac axis and position should be normal

[Fig. 3] [46]. The cardiac axis can be shifted as anormal variant, although a detailed examinationshould be considered for the possibility of asso-ciated abnormalities. Abnormal heart deviationinto the right chest (dextroposition) may be causedby a chest mass (eg, cystic adenomatoid malforma-tion, pulmonary sequestration, or congenital lo-

Fig. 2. Four-chamber view of the heart. Key componeninterventricular septum and atrial septum primum. There(RV) ventricles. A moderator band helps to identify theatrioventricular septal valve leaflets insert into the cruxMedicine. Performance of the basic fetal cardiac ultrasowith permission.)

bar emphysema), diaphragmatic hernia, or situsanomalies [46–48].

Cardiac chambersBoth atrial chambers should appear similar in sizewith an intact atrial septum primum. The foramenovale flap should move freely toward the leftatrium. In the human fetus, more than one-fifthof combined ventricular output is directed to thelungs and eventually drained back to the heartthrough the pulmonary veins [49,50]. At least onepulmonary vein should always be seen entering theleft atrium.The ventricular chambers also appear similar in

size with an intact intervening septum. Relativelythin ventricular walls usually have no greater than2 mm of surrounding pericardial fluid. The rightventricle has a moderator band at the cardiac apexand normally resides in the anterior chest on theside opposite the fetal stomach. It usually appearsmore triangular in shape as compared with the leftventricular chamber.

Atrioventricular valvesBoth atrioventricular valves should move freely andnot appear thickened. The tricuspid valve leafletnormally inserts on the ventricular septum at aposition that is closer to the cardiac apex as com-pared with the mitral valve septal leaflet insertion[51]. This describes the normal ‘‘offsetting’’ of bothatrioventricular valves.

Extended basic examination

If technically feasible, routine views of the outflowtracts should also be included as part of an ‘‘ex-

ts of a normal four-chamber view include an intactis no disproportion between the left (LV) and rightmorphologic right ventricle. Note how the “offset”. (From Lee W. American Institute of Ultrasound inund examination. J Ultrasound Med 1998;17:601–7;

Fig. 3. Fetal cardiac axis and position The cardiac axis can be measured from a four-chamber view of the fetalheart. A line through the interventricular axis is extended to the posterior border of the heart to produce point P,the location of which can be used to define fetal cardiac position. (Adapted from Comstock CH. Normal fetal heartaxis and position. Obstet Gynecol 1987;70:255–9; with permission.)

278 Lee & Comstock

tended basic’’ cardiac examination [Figs. 4 and 5].A four-chamber view of the heart can appear asnormal in the presence of a ventricular septal defector conotruncal anomaly that would otherwise beseen from the extended basic examination.Scanning planes for both the ‘‘basic’’ and ‘‘ex-

tended basic’’ examinations are illustrated inFig. 6 [22]. Both outflow tracts are examined asthe transducer is angled from the four-chamberview toward the fetal head. Another method forevaluating the outflow tracts has also been de-scribed when the fetal interventricular cardiac sep-tum is perpendicular to the ultrasound beam [52].This approach begins with a four-chamber view ofthe heart and involves probe rotation until a leftventricular outflow tract is visualized. The probecan then be rocked cephalad until the pulmonary

Fig. 4. Left ventricular outflow tract. A left ventricular ousel can be seen exiting the left ventricle. The aortic valve(From Lee W. American Institute of Ultrasound in Mediciexamination. J Ultrasound Med 1998;17:601–7; with perm

arterial outflow tract is observed in a plane thatappears perpendicular to the aortic outflow tract.Others have reported the use of a ‘‘three-vesselview’’ to describe relative sizes and relationshipsbetween the pulmonary artery, ascending aorta,and right superior vena cava [53–55].Most second trimester cardiac screening exami-

nations will permit satisfactory visualization ofthe four-chamber view and outflow tracts. Over18,000 second trimester fetuses underwent cardiacscreening of the outflow tracts to evaluate the stan-dardized practice of incorporating a basic fetalcardiac exam into a 30-minute scan [56]. Whentechnically feasible, an extended basic evaluationof the outflow tracts was also attempted. Of thestudies that included an adequate four-chamberview, the majority (93%) of scans had satisfactory

tflow tract view (LVOT) emphasizes that a great ves-leaflets should be freely moving and not thickened.

ne. Performance of the basic fetal cardiac ultrasoundission.)

Fig. 5. Right ventricular outflow tract. Cardiac axis and position are identical to Fig. 4. A right ventricular outflowtract (RVOT) view emphasizes that a great vessel can be seen exiting the morphologic right ventricle (RV). Thebifurcation of the pulmonary artery is not always seen in this scanning plane. Note that the RVOT exits theventricle at about 90° to the aortic outflow tract. The pulmonary valve leaflets should be freely moving and notthickened. Sometimes, the right superior vena cava (SVC) can be seen. (Adapted from Lee W. American Instituteof Ultrasound in Medicine. Performance of the basic fetal cardiac ultrasound examination. J Ultrasound Med1998;17:601–7; with permission.)

279Prenatal Diagnosis of Congenital Heart Disease

views of the outflow tracts. Nonvisualization rateswere: left ventricular outflow tract (4.2%); rightventricular outflow tract (1.6%); both outflowtracts (1.3%).

Sonographic detection of selected cardiacanomalies

This section summarizes the clinical significanceand sonographic findings associated with selectedexamples of CHD that may be detected by basicand extended basic cardiac screening exam.

Fig. 6. Scanning planes for the basic and extended basic carobtained from an axial scanning plane across the fetal tho(RVOT) ventricular outflow tracts are found by angling theW. American Institute of Ultrasound in Medicine. PerformaJ Ultrasound Med 1998;17:601–7; with permission.)

Ventricular septal defects

Ventricular septal defects are the most commontype of CHD [Fig. 7]. They are caused by incom-plete closure of the ventricular septum during fetaldevelopment and can occur at various locations[Fig. 8]. Although isolated lesions may not beclinically significant, their presence should raisethe possibility of abnormal karyotype and asso-ciated structural findings.Paladini and colleagues [57] summarized their ex-

perience with isolated lesions that are detected usingprenatal sonography. Of 26 fetuses that reached

diac examinations. A four-chamber view of the heart israx. Corresponding views of the left (LVOT) and righttransducer toward the fetal head. (Adapted from Leence of the basic fetal cardiac ultrasound examination.

Fig. 7. Perimembranous ventricular septal defect. A ventricular septal defect (VSD) is circled on the heart specimenin an abortus at 20 weeks, menstrual age. This small perimembranous lesion can be compared with PresidentThomas Jefferson's nose and can be easily missed if technical factors do not favor satisfactory visualization.

280 Lee & Comstock

1 year of age, 46.1% (12 cases) of all defects closedin utero. Trisomy 21 was found in 50% of fetuseswith inlet-type (nine cases) or large (two cases)defects. Trisomy 18 occurred in 56.3% of ven-tricular septal defects (VSD) lesions (nine cases)that were associated with aortic-septal override.None of the malalignment VSDs closed after birth.By comparison, 69% of the perimembranousand 60% of the muscular defects closed within1 year of delivery. In a related pediatric study, Turner

Fig. 8. Variable locations of ventricular septal defects.The location of a ventricular septal defect (VSD) can beused to estimate the likelihood of spontaneous clo-sure and for counseling about therapeutic interven-tion after birth. An inlet VSD of the posterior septum,near either the tricuspid or mitral valves, will rarelyclose. Infants with this type of lesion are not optimalcandidates for repair using a percutaneous septalocclusion device. (From Fetal ultrasound simulator [CD-ROM], Washington, D.C: American College of Obstetri-cians and Gynecologists; 1998; with permission.)

and colleagues [58] found that the location ofthe lesion was relevant to its natural history in68 infants with VSD. Perimembranous defects ac-counted for most of the moderate and large defectsthat required surgical correction. After more than6 years, almost a third of all perimembranousand just over two-thirds of all muscular defectsclosed spontaneously.Tegnander and colleagues [18] studied a nonse-

lected population of 30,149 fetuses: an isolatedVSD occurred in 57% (188/333) of noncriticalheart defects. Of these, 162 (86%) were muscularand 26 (14%) were perimembranous. During thefirst year after birth, 77/162 (48%) of the isolatedmuscular and 3/26 (12%) of the isolated perimem-branous VSDs closed spontaneously.Prenatal detection of a small VSD <2 mm can be

particularly difficult and depends on several factorssuch as the lateral resolution of the ultrasound sys-tem, gestational age, fetal size, maternal wall thick-ness, and a history of prior abdominal or uterinesurgeries. Aside from optimizing image settings (eg,depth, gain, focus), the entire ventricular septumshould be systematically examined during both thebasic and extended basic cardiac scans.Other diagnostic tools include the use of digi-

tal cineloop technology and Doppler flow studies.Direct measurements of intracardiac pressures inhuman fetuses indicate that no measurable dif-ferences occur between the left and right ventri-cles [59]. Therefore, color Doppler ultrasonographymay not detect flow across a VSD as easily as inthe case of adult patients. However, VSD lesionscan be identified using frame-by-frame analysis ofthe ventricular septum (digital cine-loop), some-times in conjunction with Doppler flow studies aswell [Figs. 9 and 10]. Some ultrasound systemsprovide a ‘‘write priority’’ control that allows the

Fig. 9. Intramuscular ventricular septal defect. This 2.6-mm VSD can be easily visualized using gray-scale sonog-raphy. Simultaneous use of digital cineloop and color Doppler techniques allow visualization of blood flowacross this defect. This abnormality would be much easier to visualize after birth when a high-pressure gradientis present between the left and right ventricles.

281Prenatal Diagnosis of Congenital Heart Disease

user to balance the color or power Doppler displayso it doesn’t ‘‘bleed’’ over surrounding intra-cardiacstructures that are visualized with gray scale. Fi-nally, one should remember to verify suspectedVSD lesions from more than one view to avoiddiagnostic errors due to sonographic artifacts.

Atrioventricular septal defects

Complete atrioventricular septal defects (AVSD)consist of a common atrioventricular junction in-stead of separate mitral and tricuspid valve orifices[Fig. 11]. In the past, this lesion has also beenknown as an endocardial cushion defect or AVcanal defect. Milder forms of this anomaly (ie, ‘‘in-complete AVSD’’) have only a defect in the inferiorpart of the atrial septum, just above the atrioven-

Fig. 10. Perimembranous ventricular septal defect. A smraphy (left), although the addition of a color Doppler stufetus (right).

tricular valves. In this case, two separate valve ori-fices will be visualized.Allan [60] reviewed the prenatal sonographic find-

ings of 49 fetuses with AVSD; 18 cases appearedto only involve AVSD only, of which 13 were sub-sequently found to have trisomy 21. Other abnor-malities included heterotaxy syndrome (22 cases),left ventricular malformations (8 cases), and TOF(1 case). Increased nuchal translucency has alsobeen found in embryos with AVSD at 10–14 weeks,menstrual age [61].There were 22 cases of heterotaxy syndrome in-

volving isomerism of the atrial appendages. Sixteenfetuses were suspected to have right atrial isomer-ism (eg, anomalies of pulmonary venous drainage,double outlet right ventricle with pulmonary valve

all VSD may not be identified on gray-scale sonog-dy can be used demonstrate this lesion in the same

Fig. 12. Atrioventricular septal defect. Sonographicfindings for a complete atrioventricular septal defectconsist of atrial septal defect (asd), ventricular septaldefect (vsd), and lack of the normal offset atrioven-tricular valve insertion sites. The common valve ap-pears as a straight echogenic line.

Fig. 11. Atrioventricular septal defect. A complete form of atrioventricular septal defect is characterized byincomplete formation of separate mitral and tricuspid valves. A cross-section demonstrates a common valveapparatus (inset). RA, right atrium; LA, left atrium; RV, right ventricle; LV, left ventricle. (Adapted from Fetalultrasound simulator [CD-ROM]. Washington, D.C: American College of Obstetricians and Gynecologists; 1998;with permission.)

282 Lee & Comstock

atresia, bilateral right bronchi and lung lobation,and asplenia). Six fetuses had sonographic evidenceof left atrial isomerism, where typical findings mayinclude an interrupted inferior vena cava with azy-gous continuation, complete heart block, bilateralleft bronchi and lung lobation, and polysplenia.This series emphasized the high rate of intrauterinedeath for fetuses with left atrial isomerism becauseonly one of six abnormal fetuses survived to deliv-ery, but died soon after birth. Although the prog-nosis for fetuses with AVSD and right atrialisomerism typically is to survive pregnancy, thereis a high postnatal mortality rate found in infancyand early childhood [62].The sonographic detection of complete AVSD is

usually straightforward because the normal offset-ting of the atrioventricular valves is not present.Instead, the common atrioventricular valve appearsas a straight line [Fig. 12]. Occasionally, a dilatedcoronary sinus can simulate an AVSD lesion [63].Incomplete forms of AVSD also may obscure thediagnosis because two separate inlet valves are stillseen, in addition to both atrial and ventrticularseptal defects. As a final consideration, ventriculardisproportion may appear as an ‘‘unbalanced’’ formof AVSD. Complete AVSD can be typically repairedwith low mortality and good intermediate to long-term results [64].

Hypoplastic left heart syndrome

Prenatal identification of hypoplastic left heart syn-drome (HLHS), using the basic cardiac screening

examination, is extremely important because ofimproved surgical outcome in fetuses in whomthe lesion was detected prenatally [65] [Fig. 13].Severe ventricular disproportion is the hallmark ofthis abnormality where the left ventricle can appearvery small [Fig. 14].This condition refers to an underdeveloped left

ventricle from abnormal development of the mitralor aortic valve. Valve obstruction causes a shift ofblood flow back over the foramen ovale to the right

Fig. 13. Hypoplastic left heart syndrome Hypoplasticleft heart syndrome consists of a small left ventricle(LV) with abnormal mitral or aortic valve develop-ment. The LV will initially appear slightly small duringearly pregnancy, with subsequent development as arudimentary slit-like cavity as the pregnancy prog-resses. The dominant chamber is the right ventricle(RV). Blood flow across the foramen ovale is reversedfrom the right (RA) to left (LA) atrium. (Adapted fromFetal ultrasound simulator [CD-ROM].Washington, D.C:American College of Obstetricians and Gynecologists;1998; with permission.)

Fig. 14. Hypoplastic left heart syndrome. Sonographicfindings demonstrate marked underdevelopmentof the left ventricular cavity that gives it a “slit-like” appearance. This degree of ventricular dispro-portion may not be as obvious at an earlier stageof pregnancy.

283Prenatal Diagnosis of Congenital Heart Disease

atrium. The left ventricle stays small due to de-creased blood flow. A truly anatomic univentricularheart is very uncommon since a small ‘‘slit’’ canusuallly be seen to the left of the right ventricle.In severe mitral valve dysplasia, the left ventricle

is usually quite small at the time of the screeningexam. Milder involvement of the valves causes dis-parity of the size of the right and left ventricles, butboth are still visible. Another presentation is that ofan enlarged left atrium with a large noncontractingleft ventricle. Although the ventricle is large, blooddoes not flow into or out of it. The enlarged ven-tricle can shrink with advancing pregnancy. Asso-ciated findings may include aortic valve atresia,coarctation of the aorta, and echogenic thickeningof the ventricular walls due to endocardial fibro-elastosis [66].Hypoplastic left ventricle can also occur with

other defects such as atrioventricular septal defect.It is rare that chromosomes are abnormal, but ifthey are, trisomy 18 is the most frequent aneu-ploidy. Hypoplastic left ventricle can also occurin left-sided diaphragmatic hernia due to pressureon the left ventricle. The pregnancy course is usu-ally uneventful in the absence of heart failure (eg,atrioventricular valve insufficiency, pericardial effu-sion, cardiomegaly).After birth the newborn will depend on a patent

foramen ovale and ductus arteriosus to get bloodto the aorta and neck vessels. Prostaglandin infu-

sion will keep these fetal shunts open temporarily,but eventually surgery will be necessary. The con-ventional surgical repair is a three-stage procedure,the first of which is a Norwood procedure thatinvolves connecting the aorta to the proximal pul-monary artery, thus allowing the right ventricle topump blood to the body [67]. Two lower riskoperations, the hemi-Fontan (4 to 6 months of age)and Fontan (18 months to 2 years of age) aresubsequently required [68]. Although these chil-dren are now reaching young adulthood and aredoing remarkably well, recent studies suggest in-creased risk for cognitive, neuromotor, and psycho-social problems [69].

Tricuspid valve abnormalities

Tricuspid valve atresia is caused by marked dyspla-sia of the leaflets and cords where a connection failsto develop between the right atrium and ventricle.Therefore, systemic venous blood return bypassesthe right heart and traverses a patent foramen ovale(secundum atrial septal defect, ASD) into the leftatrium and ventricle. An inlet VSD typically allowssome blood flow from the left ventricle into anunderdeveloped right ventricle.Tongsong and colleagues [70] published prenatal

sonographic findings for isolated tricuspid valveatresia. The four chamber view fails to demonstratea patent tricuspid valve and this area appears echo-genic [Fig. 15]. A small right ventricle is seen, de-

Fig. 15. Tricuspid valve atresia. Key findings consistof an echogenic mass of non-mobile tissue betweenthe right ventricle (RV) and atrium (RA), small rightventricle (RV), and the presence of a ventricular septaldefect (*). Increased systemic venous return flowsfrom the RA through the foramen ovale (FO).

284 Lee & Comstock

pending on size of the VSD. Right ventricular out-flow tract obstruction may occur as subvalvular pul-monary stenosis or pulmonary valve stenosis—apotential cause of ductal dependency. Approximately20% of cases will be associated with transposi-tion of the great arteries. Aortic arch abnormalitieshave also been reported. Chromosomes are usuallynormal and it is rare to have extracardiac defects.Increased nuchal translucency, however, has beenreported in embryos with triscuspid atresia as earlyas 11 weeks, menstrual age [61].The surgical outcome of infants born with this

lesion depends on the presence of associatedanomalies and relies on use of the Fontan proce-dure. The Hospital for Sick Children recently sum-marized a total of 137 infants who underwent aFontan procedure [71]. Cohort survival was 90% atthe age of 1 month, 81% at 1 year, 70% at 10 years,and 60% at 20 years.The tricuspid valve septal leaflet is normally

not formed before 12 weeks, menstrual age [28].However, Ebstein’s malformation is another tri-cuspid valve anomaly that is caused by failure ofthe inferior atrioventricular cushion to properlydevelop this leaflet from the right side of the ven-tricular septum. This anomaly is characterized byincreased offsetting of the atrioventricular valvesand tricuspid valve insufficiency that leads to anenlarged right atrium. A nomogram of the mitral totricuspid valve distance is helpful for confirmingthis diagnosis [51]. Melendres and colleagues [72]recently reported a missed case of Ebstein’s anom-aly that was obscured by misinterpretation of anatrioventricular groove. Ebstein’s anomaly has a

poor perinatal prognosis with a mortality rate ashigh as 85% [73].

Conotruncal abnormalities

A primitive endocardial heart tube, superior tothe primitive ventricles, becomes divided in halfby a spiral septum to form great arteries early ingestation. This process may be interrupted forunknown reasons, leading to a spectrum of de-velopmental cardiac anomalies such as truncusarteriosus, transposition of the great arteries,double-outlet right ventricle, and TOF. The funda-mental lesion depends on where formation of thespiral septum is disrupted and where the greatarteries are positioned in relation to each other atthat time. Sonographic distinction between thesespecific abnormalities may be very difficult beforebirth [74,75].Some conotruncal abnormalities, such as TOF

and double outlet right ventricle, are also at in-creased risk for chromosomal abnormalities [76].In a recent population-based study of 255,849 births,43 children were found to have 22q11.2 deletionwith an overall prevalence of 1 in 5950 births(95% CI, 1 in 4417 – 1 in 8224 births) [77]. Mostaffected children (81%) had a heart defect thatincluded conotruncal anomalies (46%), interruptedaortic arch (19%), ventricular septal defects (16%),and other assorted extra-cardiac vascular anomalies(51%). Conotruncal aberrations are more frequentin diabetics and in those women who have a con-genital abnormality themselves or have had a childwith a genetic disorder.The basic cardiac examination, using a four-

chamber view alone, is notoriously unreliable fordetecting these anomalies. However, an extendedbasic examination of the outflow tracts is the keyfor their effective prenatal detection. The cardiacaxis may provide an initial sonographic indicationof a conotruncal anomaly during the basic exami-nation [78,79]. DeVore [80] has also described theuse of color Doppler sonography to visualize greatvessel relationships for second and third trimes-ter pregnancies.

Truncus arteriosusTruncus arteriosus occurs when one great arteryarises from the base of the heart and gives rise tothe coronary, pulmonary, and aortic circulations.The truncus usually overrides a VSD. The pulmo-nary arteries arise from the truncal root as a com-mon trunk (Type I, most frequent) [Fig. 16A, B],close but separate (Type II), or widely separate(Type III) [81]. Fortunately, they are not ductal-dependent lesions.Fetal echocardiography should focus on the ven-

tricular origin of the truncus, truncal valve annular

Fig. 17. Transposition of the great arteries. The main-diagnostic findings is the manner by which the greatarteries exit the ventricles in a parallel manner. Fetalechocardiography would demonstrate that the aor-tic root originates from the right ventricle (RV),whereas the pulmonary artery would exit the leftventricle (LV). A small ventricular septal defect is alsopresent (arrow).

Fig. 16. Truncus arteriosus. (A) Axial view of the fetal thorax reveals a small pulmonary artery (arrow) that directlyoriginates from a large truncal root (Tr). The fetal spine (Sp) is seen as a point of reference. (B) Oblique coronalview of the left hemithorax demonstrates a large truncal vessel, with ventricular septal defect, that exits the right(RV) and left (LV) ventricles. This truncal vessel overrides the ventricular septum and gives off a small pulmonaryvessel (arrow).

285Prenatal Diagnosis of Congenital Heart Disease

diameter, and evidence for valvular dysplasia.Doppler studies may reveal truncal valve stenosisor insufficiency [82]. In a pathology study of28 Type I and Type II cases, other cardiovascularlesions included anomalous position of the leftcoronary artery (18.5%), right aortic arch (36%),and interrupted aortic arch (11%) [83]. The distinc-tion between truncus arteriosus and pulmonaryvalve atresia with VSD can be very difficult [84].An investigation from the University of Michigan

described 46 infants with truncus arteriosus under-going early primary repair and found an actuarialsurvival rate of 81 ± 6% at 90 days and beyond [85].More recently, an observational study of 23 affectedfetuses indicates the following outcomes: termina-tion of pregnancies (34.8%); intrauterine death(8.7%); postnatal deaths (21.7%) [84]. The eightremaining neonates (34.8%) were alive and doingwell after surgery (n=6) or awaiting repair (n=2).Major cardiac surgical centers currently favor a pri-mary repair of truncus arteriosus during the neona-tal period.

Transposition of the great arteriesTransposition of the great arteries (TGA) occurswhen the aorta arises from the right ventricle andthe pulmonary artery arises from the left ventricle.The usual spiral relationship of the great arteriesis lost so that the outflow tracts are parallel toeach other [Fig. 17] [86]. In about half of casesthere is no VSD. Affected newborns without a largeVSD may not have adequate mixing of oxygenatedblood and can experience rapid hemodynamicdecompensation as the ductus arteriosus closes.This abnormality can be easily missed from the

four-chamber view unless a very large VSD is pre-

sent. In this case, an extended basic examinationof the outflow tracts is likely to identify TGA. Con-genitally corrected transposition of the great ar-teries can also rarely occur where parallel vesselsare also seen exiting the heart. In this situation, theatria connect with anatomically discordant ventri-cles and the ventricles connect with discordant andtransposed great arteries [87]. Careful attention tothe chamber morphology, presence of moderatorband, and papillary muscle relationships can pro-

Fig. 19. Tetralogy of Fallot. The initial diagnostic cluefor tetralogy of Fallot is the presence of an overridingaorta (ao) over a ventricular septal defect (*) duringthe extended basic examination of the heart. Sono-graphic evidence for a small pulmonary artery (notseen in this image) may not occur until a much latertime in pregnancy. s, ventricular septum; sp, spine.

286 Lee & Comstock

vide important clues to help the examiner accu-rately distinguish between right and left ventricles.The potential impact of early prenatal diagnosis is

illustrated by a study that examined clinical out-come in 68 affected neonates in whom a prenataldiagnosis of TGA was suspected [88]. Results werecompared with 250 affected neonates who werefirst identified with TGA after delivery. Newbornswith late diagnosis had increased delay betweenbirth and admission for special care and weremore likely to present with metabolic acidosisand multi-organ failure. Preoperative mortality was6% for the neonatal group as compared with nodeaths in the prenatal group. Postoperative mor-bidity was not different between groups, althoughthe hospital stay was slightly longer in babies diag-nosed after birth. Postoperative death was sig-nificantly higher in the neonatal group (20 of235 versus 0 of 68, P<.01. Their results suggest thatprenatal diagnosis of TGA reduces mortality andmorbidity. Timely detection of TGA before deliveryprovides adequate planning for monitoring andantepartum care. This information also facilitatesin utero transfer of fetuses to a tertiary care facilitythat can appropriately treat newborns with ductaldependent lesions.

Double-outlet right ventricleDouble-outlet right ventricle (DORV) describes acondition in which most of the aorta and pulmo-nary artery arise from the right ventricle [89].The sonographic findings may closely resembleTOF or transposition of the great arteries with VSD[Fig. 18]. The outcome of infants with DORV de-pends largely on the associated anomalies.

Fig. 18. Double-outlet right ventricle. Both great ves-sels exit the right ventricle (RV). The most lateral onerepresents the aortic arch (Ao) because neck vesselsare present (arrows). No arteries are seen exiting theleft ventricle (LV). Sp = spine.

As in TOF there is always an accompanying VSDbut, unlike TOF the relationship of the great vesselsis often abnormal; the aorta is commonly trans-posed anterior and to the right of the pulmonaryartery. Other defects are not infrequent such aspulmonary valve stenosis, right cardiac axis devia-tion, right aortic arch, atrial septal defect, and totalanomalous pulmonary venous return. DORV withtransposition is known as a Taussig-Bing defectand is commonly associated with coarctation. Thecombination of DORV and AVSD is difficult to re-pair surgically.

Tetralogy of FallotTOF is the only conotruncal anomaly in which theusual spiral relationship of the great vessels is main-tained. The aorta overrides a ventricular sepal defect[Fig. 19]. The pulmonary artery may be small dueto uneven division of the primitive truncus. In thefetus, unlike in children, the right ventricle is nothypertrophied because of shunting across the fora-men ovale and the VSD reduces the load on theright heart.Early fetal TOF may simply present as a VSD with

aortic septal override. This anomaly can be missedon a screening four-chamber view if the VSD issmall [Fig. 20]. Left cardiac axis deviation mayprovide an initial diagnostic clue for this lesion[79]. However, the extended basic cardiac examina-tion is most likely to demonstrate a VSD with aorticoverride. The aortic root itself can also be enlarged,although pulmonary valve stenosis may be become

Fig. 20. Tetralogy of Fallot with normal four-chamberview The same fetus with tetralogy of Fallot appearsin Fig. 19. This four-chamber view appears normaldespite the presence of this conotruncal anomaly. lv,left ventricle; rv, right ventricle; ra, right atrium; la,left atrium.

287Prenatal Diagnosis of Congenital Heart Disease

apparent until later pregnancy [90,91]. The aorticto pulmonary size ratio will be high despite normalvalve diameter measurements. In TOF, this ratioincreases as gestational age advances becausethere is less than the usual growth of the pulmo-nary diameter [92]. The 90% confidence intervalfor the pulmonary artery to aortic diameter (Pa/Ao)ratio ranges from 0.84 to 1.41 and remains con-stant throughout pregnancy [93]. The right ventri-cular outflow tract should be re-evaluated beforedelivery to identify affected fetuses at greatest riskfor ductal dependency.

Fig. 21. Ventricular disproportion A disparity in ventricularone should also consider that this could be caused by severestriction, aortic coarctation, or an evolving hypoplastic l

These newborns may experience hemodynamicdecompensation if the pulmonary valve diameteris small (≤ 5 mm at term) or if there is retrogradeflow across the ductus arteriosus. Immediate sur-gery may not be required for affected infants withsufficient pulmonary valve flow and absence ofductal dependency.One variation, in which the pulmonary valve is

atretic and the pulmonary artery is not visible, isknown as ‘‘pseudotruncus’’ as only a solitary largevessel is seen straddling the VSD. Another variationof TOF involves absence of the pulmonary valve sothat there is regurgitation back and forth. The pul-monary artery may enlarge to massive proportionsand hydrops may occur.

Aortic coarctation

Cardiac ventricular disproportion can occur as anormal variant, a consequence of fetal growthrestriction, or an indirect sign of aortic coarctation[Fig. 21]. Aortic coarctation is usually difficultto directly visualize, especially in the context of acardiac screening examination. However, the pre-natal diagnosis of aortic coarctation has beenreported to improve survival and reduce morbid-ity [94].Allan and colleagues [95] found that 24 fetuses

had dilatation of the right heart as compared withthe left side. In 18 of these cases, the diagnosis ofaortic coarctation or interruption was correctlyinferred from indirect sonographic findings. Horn-berger and associates [96] summarized a multicen-ter investigation for 20 infants with coarctation.They found quantitative hypoplasia of the aorticisthmus and transverse arch as the best predictors

dimensions can represent a normal variant. However,ral other conditions that include intrauterine growtheft heart syndome.

288 Lee & Comstock

for coarctation and emphasized the importance ofconducting serial studies for suspected cases. Ven-tricular disproportion has only a moderate degreeof sensitivity (62%) for the detection of coarcta-tion that occurs with a high false-positive rate after34 weeks [97]. The 90% confidence interval for theright ventricular to left ventricular diameter ratio isconstant throughout pregnancy and ranges from0.79 to 1.24 [93].

Total anomalous pulmonary venous return

The left atrium is normally positioned near thedescending aorta. In the space between these twostructures are the main left and right pulmonaryveins. They can be followed to their drainage pointin the back of the left atrium. In total anomalouspulmonary venous return (TAPVR), both main pul-monary veins drain into the right atrium via avertical vein which then drains into a coronarysinus, persistent left superior vena cava, innomi-nate vein, the hepatic veins, or even below thediaphragm (20%) into the inferior vena cava.Thus oxygenated blood from the lungs neverreaches the body and brain but rather circulatesaround and around in the lungs with the onlymixing available across a patent foramen ovale.TAPVR can lead to cardiorespiratory decompen-

sation of the newborn that is unresponsive to pros-taglandin infusion. This circulatory abnormalityshould be considered when a dilated coronary si-nus, cardiac chamber disproportion, or size dis-

Fig. 22. Total anomalous pulmonary venous return Atleast two pulmonary veins (arrows) drain into theright atrium instead of the left. A ventricular septaldefect is also present between the two markers. LV,left ventricle. Increased volume loading of the rightheart and cyanosis are the key clinical findings for thiscondition. An atrial septal defect is desirable becausethis defect will allow blood to flow from the rightatrium into the left heart.

parity between the great arteries is seen. It is par-ticularly common in fetuses with heterotaxic ab-normalities. Direct documentation of pulmonaryvenous flow into the left atrium is the most reliableway to exclude TAPVR [98] [Fig. 22].A nonlethal variation is partial anomalous

venous return in which the right vein drains tothe right atrium but the left still drains to the leftatrium. As long as some return goes to the leftatrium, oxygenated blood is available to the bodyand brain.

Summary

CHD is a leading cause of infant morbidity andmortality that results from birth defects. Diagnosti-cians who use ultrasonography to evaluate the fetalheart must be familiar with key factors that canimpact the success of their cardiac screening pro-grams. A good understanding of practice guidelinesfor the ‘‘basic’’ and ‘‘extended basic’’ cardiac exami-nation is essential. Efforts should also be made tostandardize diagnostic training of these who per-form these examinations in an ongoing manner.The primary goal is for these individuals to accu-rately identify who should be referred for a moredetailed evaluation of the fetus.

References

[1] Kochanek KD, Murphy SL, Anderson RN, Scott C.Deaths: final data for 2002. Natl Vital Stat Rep2004;53:1–115.

[2] Centers for Disease Control and Prevention.Trends in infant mortality attributable to birthdefects - United States, 1980–1995. MMWRMorbMortal Wkly Rep 2005;47:773–8.

[3] Rosano A, Botto LD, Botting B, Mastroiacovo P.Infant mortality and congenital anomalies from1950 to 1994: an international perspective. J Epi-demiol Community Health 2000;54:660–6.

[4] Ferencz C, Rubin JD, McCarter RJ, et al. Con-genital heart disease: prevalence at livebirth. TheBaltimore-Washington infant study. Am J Epide-miol 1985;121:31–6.

[5] Meberg A, Otterstad JE, Froland G, et al. Out-come of congenital heart defects–a population-based study. Acta Paediatr 2000;89:1344–51.

[6] Cuneo BF, Curran LF, Davis N, Elrad H. Trendsin prenatal diagnosis of critical cardiac defects inan integrated obstetric and pediatric cardiac imag-ing center. J Perinatol 2004;24:674–8.

[7] Allan LD, Crawford DC, Chita SK, Tynan MJ.Prenatal screening for congenital heart disease.BMJ 1986;292:717–9.

[8] Sharland GK, Allan LD. Screening for congenitalheart disease prenatally. Results of a 2 1/2 yearstudy in the South East Thames Region. Br JObstet Gynaecol 1992;99:220–5.

289Prenatal Diagnosis of Congenital Heart Disease

[9] Copel JA, Pilu G, Green J, Hobbins JC, KleinmanCS. Fetal echocardiographic screening for con-genital heart disease: the importance of the four-chamber view. Am J Obstet Gynecol 1987;157:648–55.

[10] Crane JP, LeFevre ML, Winborn RC, et al. Arandomized trial of prenatal ultrasonographicscreening: impact on the detection, management,and outcome of anomalous fetuses. The RADIUSStudy Group. Am J Obstet Gynecol 1994;171:392–9.

[11] Kirk JS, Riggs TW, Comstock CH, et al. Prenatalscreening for cardiac anomalies: the value ofroutine addition of the aortic root to the four-chamber view. Obstet Gynecol 1994;84:427–31.

[12] Ott WJ. The accuracy of antenatal fetal echocar-diography screening in high and low riskpatients. Am J Obstet Gynecol 1995;172:1741–7.

[13] Tegnander E, Eik-Nes SH, Johansen OJ, LinkerDT. Prenatal detection of heart defects at theroutine fetal examination at 18 weeks in a non-selected population. Ultrasound Obstet Gynecol1995;5:372–80.

[14] Stumpflen I, Stumpflen A, Wimmer M, Ber-naschek G. Effect of detailed fetal echocardiog-raphy as part of routine prenatal ultrasoundscreening on detection of congenital heart dis-ease. Lancet 1996;348:854–7.

[15] Buskens E, Grobbee DE, Frohn-Mulder IME.Efficacy of routine ultrasound screening forcongenital heart disease in normal pregnancy.Circulation 1996;94:67–72.

[16] Bromley B, Estroff JA, Sanders SP, et al. Fetalechocardiography: accuracy and limitations in apopulation at high and low risk for heart defects.Am J Obstet Gynecol 1992;166:1473–81.

[17] Wigton TR, Sabbagha RE, Tamura RK, et al.Sonographic diagnosis of congenital heart dis-ease: comparison between the four-chamberview and multiple cardiac views. Obstet Gynecol1993;82:219–24.

[18] Tegnander E, Williams W, Johansen OJ, et al.Prenatal detection of heart defects in a non-selected population of 30,149 fetuses—detectionrates and outcome. Ultrasound Obstet Gynecol2006;27:252–65.

[19] Mitchell SC, Korones SB, Berendes HW. Con-genital heart disease in 56,109 births. Incidenceand natural history. Circulation 1971;43:323–32.

[20] Chaoui R. The four-chamber view: four reasonswhy it seems to fail in screening for cardiac ab-normalities and suggestions to improve detectionrate. Ultrasound Obstet Gynecol 2003;22:3–10.

[21] Senat MV, Holden D, Bernard JP, et al. Feasibilityof the second-trimester fetal ultrasound exami-nation in an unselected population at 18, 20or 22 weeks of pregnancy: a randomized trial.Ultrasound Obstet Gynecol 1999;14:92–7.

[22] Lee W. American Institute of Ultrasound inMedicine. Performance of the basic fetal cardiacultrasound examination. J Ultrasound Med 1998;17:601–7.

[23] Barnett SB, Ter Haar GR, Ziskin MC, et al.International recommendations and guidelinesfor the safe use of diagnostic ultrasound in medi-cine. Ultrasound Med Biol 2000;26:355–66.

[24] Paladini D, Vassallo M, Tartaglione A, et al. Therole of tissue harmonic imaging in fetal echo-cardiography. Ultrasound Obstet Gynecol 2004;23:159–64.

[25] Kossof G. Basic physics and imaging charac-teristics of ultrasound. World J Surg 2000;24:134–42.

[26] DeVore GR, Medearis AL, Bear MB, et al. Fetalechocardiography: factors that influence imagingof the fetal heart during the second trimester ofpregnancy. J Ultrasound Med 1993;12:659–63.

[27] Moorman A, Webb S, Brown NA, et al. Devel-opment of the heart: (1) formation of the cardiacchambers and arterial trunks. Heart 2003;89:806–14.

[28] Cook AC, Yates RW, Anderson RH. Normal andabnormal fetal cardiac anatomy. Prenat Diagn2004;24:1032–48.

[29] Carvalho JS, Mavrides E, Shinebourne EA, et al.Improving the effectiveness of routine prenatalscreening for major congenital heart defects.Heart 2002;88:387–91.

[30] Cuneo BF, Curran LF, Davis N, Elrad H. Trendsin prenatal diagnosis of critical cardiac defectsin an integrated obstetric and pediatric cardiacimaging center. J Perinatol 2004;24:674–8.

[31] Hunter S, Heads A, Wyllie J, Robson S. Pre-natal diagnosis of congenital heart disease in thenorthern region of England: benefits of a trainingprogramme for obstetric ultrasonographers.Heart 2000;84:294–8.

[32] Yagel S, Weissman A, Rotstein Z, et al. Con-genital heart defects: natural course and in uterodevelopment. Circulation 1997;96:550–5.

[33] Achiron R, Rotstein Z, Lipitz S, et al. First-trimester diagnosis of fetal congenital heartdisease by transvaginal ultrasonography. ObstetGynecol 1994;84:69–72.

[34] Hyett J, Moscoso G, Papapanagiotou G, et al.Abnormalities of the heart and great arteries inchromosomally normal fetuses with increasednuchal translucency thickness at 11–13 weeks ofgestation. Ultrasound Obstet Gynecol 1996;7:245–50.

[35] Hyett JA, Perdu M, Sharland GK, et al. Increasednuchal translucency at 10–14 weeks of gestationas a marker for major cardiac defects. UltrasoundObstet Gynecol 1997;10:242–6.

[36] Rustico MA, Benettoni A, D’Ottavio G, et al. Earlyscreening for fetal cardiac anomalies by trans-vaginal echocardiography in an unselected popu-lation: the role of operator experience. UltrasoundObstet Gynecol 2000;16:614–9.

[37] Huggon IC, Ghi T, Cook AC, et al. Fetal cardiacabnormalities identified prior to 14 weeks’ ges-tation. Ultrasound Obstet Gynecol 2002;20:22–9.

[38] Carvalho JS. Fetal heart scanning in the firsttrimester. Prenat Diagn 2004;24:1060–7.

290 Lee & Comstock

[39] Carvalho JS, Moscoso G, Tekay A, et al. Clinicalimpact of first and second trimester fetal echo-cardiogaphy on high risk pregnancies. Heart2004;90:921–6.

[40] American Institute of Ultrasound in Medicine.Guidelines for the performance of the antepar-tum obstetrical ultrasound examination. J Ultra-sound Med 2003;22:1116–25.

[41] American College of Radiology. ACR practiceguideline for the performance of antepartumobstetrical ultrasound. In: Practice guidelines& technical standards. Reston, VA: ACR; 2004.p. 689–95.

[42] American College of Obstetricians and Gynecolo-gists. ACOG Practice Bulletin. Ultrasonography inpregnancy. Obstet Gynecol 2004;104:1449–58.

[43] Small M, Copel JA. Indications for fetal echo-cardiography. Pediatr Cardiol 2004;25:201–22.

[44] Di Salvo DN, Brown DL, Doubilet PM, et al.Clinical significance of isolated fetal pericardialeffusion. J Ultrasound Med 1994;13:291–3.

[45] Yoo SJ, Min JY, Lee YH. Normal pericardial fluidin the fetus: color and spectral Doppler analysis.Ultrasound Obstet Gynecol 2001;18:248–52.

[46] Comstock CH. Normal fetal heart axis and posi-tion. Obstet Gynecol 1987;70:255–9.

[47] Smith RS, Comstock CH, Kirk JS, Lee W. Ultra-sonographic left cardiac axis deviation: a markerfor fetal anomalies. Obstet Gynecol 1995;85:187–91.

[48] Comstock CH, Smith R, Lee W, Kirk JS. Rightfetal cardiac axis: clinical significance and asso-ciated findings. Obstet Gynecol 1998;91:495–9.

[49] Sutton MS, Groves A, MacNeill A, et al. Assess-ment of changes in blood flow through the lungsand foramen ovale in the normal human fetuswith gestational age: a prospective Dopplerechocardiographic study. Br Heart J 1994;71:232–7.

[50] Rasanen J, Wood DC, Weiner S, et al. Role ofthe pulmonary circulation in the distribution ofhuman fetal cardiac output during the secondhalf of pregnancy. Circulation 1996;94:1068–73.

[51] Vettraino IM, Huang R, Comstock CH. The nor-mal offset of the tricuspid septal leaflet in thefetus. J Ultrasound Med 2002;21:1099–104.

[52] DeVore G. The aortic and pulmonary outflowtract screening examination in the human fetus.J Ultrasound Med 1992;11:345–8.

[53] Yoo SJ, Lee YH, Cho KS. Abnormal three-vesselview on sonography: a clue to the diagnosis ofcongenital heart disease in the fetus. AJR Am JRoentgenol 1999;172:825–30.

[54] Vinals F, Heredia F, Giuliano A. The role of thethree vessels and trachea view (3VT) in thediagnosis of congenital heart defects. UltrasoundObstet Gynecol 2003;22:358–67.

[55] Yagel S, Arbel R, Anteby EY, et al. The threevessels and trachea view (3VT) in fetal cardiacscanning. Ultrasound Obstet Gynecol 2002;20:340–5.

[56] Vettraino IM, Lee W, Bronsteen RA, Comstock

CH. Sonographic evaluation of the ventricularcardiac outflow tracts [Letter to the Editor].J Ultrasound Med 2005;24:566.

[57] Paladini D, Palmieri S, Lamberti A, et al. Char-acterization and natural history of ventricularseptal defects in the fetus. Ultrasound ObstetGynecol 2000;16:118–22.

[58] Turner SW, Hunter S, Wyllie JP. The naturalhistory of ventricular septal defects. Arch DisChild 1999;81:413–6.

[59] Johnson P, Maxwell DJ, Tynan MJ, Allan LD.Intracardiac pressures in the human fetus. Heart2000;84:59–63.

[60] Allan LD. Atrioventricular septal defect in thefetus. Am J Obstet Gynecol 1999;181(5 Pt 1):1250–3.

[61] Galindo A, Comas C, Martinez JM, et al. Cardiacdefects in chromosomally normal fetuses withincreased nuchal translucency at 10–14 weeks ofgestation. J Matern Fetal Neonatal Med 2003;13:163–70.

[62] Gentles TL, Mayer Jr JE, Gauvreau K, et al. Fon-tan operation in five hundred consecutive pa-tients: factors influencing early and late outcome.J Thorac Cardiovasc Surg 1997;114:376–91.

[63] Park JK, Taylor DK, Skeels M, Towner DR.Dilated coronary sinus in the fetus: misinterpre-tation as an atrioventricular canal defect. Ultra-sound Obstet Gynecol 1997;10:126–9.

[64] Bando K, Turrentine MW, Sun K, et al. Surgicalmanagement of complete atrioventricular septaldefects. A twenty-year experience. J Thorac Car-diovasc Surg 1995;110:1543–52.

[65] Tworetzky W, McElhinney DB, Reddy VM, et al.Improved surgical outcome after fetal diagnosisof hypoplastic left heart syndrome. Circulation2001;103:1269–73.

[66] McCaffrey FM, Sherman FS. Prenatal diagnosis ofsevere aortic stenosis. Pediatr Cardiol 1997;18:276–81.

[67] Norwood WI, Kirkland JK, Sanders SP. Hypo-plastic left heart syndrome: experience withpalliative surgery. Am J Cardiol 1980;45:87–91.

[68] Bove EL, Ohye RG, Devaney EJ. Hypoplastic leftheart syndrome: conventional surgical manage-ment. Semin Thorac Cardiovasc Surg PediatrCard Surg Annu 2004;7:3–10.

[69] Mahle WT, Wernovsky G. Neurodevelopmentaloutcomes in hypoplastic left heart syndrome.Semin Thorac Cardiovasc Surg Pediatr Card SurgAnnu 2004;7:39–47.

[70] Tongsong T, Sittiwangkul R, Wanapirak C,Chanprapaph P. Prenatal diagnosis of isolatedtricuspid valve atresia: report of 4 cases andreview of the literature. J Ultrasound Med 2004;23:945–50.

[71] Sittiwangkul R, Azakie A, Van Arsdell GS, et al.Outcomes of tricuspid atresia in the Fontan era.Ann Thorac Surg 2004;77:889–94.

[72] Melendres G, Ormsby EL, McGahan JP, et al.Prenatal diagnosis of Ebstein anomaly: a poten-tial pitfall. J Ultrasound 2004;23:551–5.

291Prenatal Diagnosis of Congenital Heart Disease

[73] Pavlova M, Fouron JC, Drblik SP, et al. Factorsaffecting the prognosis of Ebstein’s anomalyduring fetal life. Am Heart J 1998;135:1081–5.

[74] Allan LD, Sharland GK, Milburn A, et al.Prospective diagnosis of 1,006 consecutive casesof congenital heart disease in the fetus. J Am CollCardiol 1994;23:1452–8.

[75] Tometzki AJ, Suda K, Kohl T, et al. Accuracy ofprenatal echocardiographic diagnosis and prog-nosis of fetuses with conotruncal anomalies.J Am Coll Cardiol 1999;33:1696–701.

[76] Paladini D, Rustico M, Todros T, et al. Cono-truncal anomalies in prenatal life. UltrasoundObstet Gynecol 1996;8:241–6.

[77] Botto LD, May K, Fernhoff PM, et al. Apopulation-based study of the 22q11.2 deletion:phenotype, incidence, and contribution to majorbirth defects in the population. Pediatrics 2003;112:101–7.

[78] Comstock CH, Smith RS, Lee W, Kirk JS. Rightfetal cardiac axis: clinical significance and asso-ciated findings. Obstet Gynecol 1998;85:495–9.

[79] Smith RS, Comstock CH, Kirk JS, Lee W. Ultra-sonographic left cardiac axis deviation: a markerfor fetal anomalies. Obstet Gynecol 1995;85:187–91.

[80] DeVore GR. Color Doppler examination of theoutflow tracts of the fetal heart: a technique foridentification of cardiovascular malformations.Ultrasound Obstet Gynecol 1994;4:463–71.

[81] Collett RW, Edwards JE. Persistent truncus arte-riosus: a classification according to anatomictypes. Surg Clin N Am 1949;29:1245–70.

[82] Duke C, Sharland GK, Jones AM, Simpson JM.Echocardiographic features and outcome oftruncus arteriosus diagnosed during fetal life.Am J Cardiol 2001;88:1379–84.

[83] Butto F, Lucas Jr RV, Edwards JE. Persistent trun-cus arteriosus: pathologic anatomy in 54 cases.Pediatr Cardiol 1986;7:95–101.

[84] Volpe P, Paladini D, Marasini M, et al. Commonarterial trunk in the fetus: characteristics, associa-tions, and outcome in a multicentre series of23 cases. Heart 2003;89:1437–41.

[85] Bove EL, Lupinetti FM, Pridjian AK, et al. Resultsof a policy of primary repair of truncus arteriosusin the neonate. J Thorac Cardiovasc Surg 1993;105:1057–65.

[86] Allan LD. Sonographic detection of parallel great

arteries in the fetus. AJR Am J Roentgenol 1997;168:1283–6.

[87] McEwing RL, Chaoui R. Congenitally correctedtransposition of the great arteries: clues for pre-natal diagnosis. Ultrasound Obstet Gynecol 2004;23:68–72.

[88] Bonnet D, Solti A, Butera G, et al. Detection oftransposition of the great arteries in fetusesreduces neonatal morbidity and mortality. Cir-culation 1999;99:916–8.

[89] Smith RS, Comstock CH, Kirk JS, et al. Double-outlet right ventricle: an antenatal diagnosticdilemma. Ultrasound Obstet Gynecol 1999;14:315–9.

[90] DeVore GR, Siassi B, Platt LD. Fetal echocardiog-raphy. VIII. Aortic root dilatation—a marker fortetralogy of Fallot. Am J Obstet Gynecol 1988;159:129–36.

[91] Lee W, Smith RS, Comstock CH, et al. Tetralogyof Fallot: prenatal diagnosis and postnatal sur-vival. Obstet Gynecol 1995;86:583–8.

[92] Hornberger LK, Sanders SP, Sahn DJ, et al. Inutero pulmonary artery and aortic growth andpotential for progression of pulmonary outflowtract obstruction in tetralogy of Fallot. J Am CollCardiol 1995;25:739–45.

[93] Kirk JS, Comstock CH, Lee W, et al. Fetal cardiacasymmetry: a marker for congenital heart disease.Obstet Gynecol 1999;93:189–92.

[94] Franklin O, Burch M, Manning N, et al. Prenataldiagnosis of coarctation of the aorta improvessurvival and reduces morbidity. Heart 2002;87:67–9.

[95] Allan LD, Chita SK, Anderson RH, et al. Coarcta-tion of the aorta in prenatal life: an echo-cardiographic, anatomical, and functional study.Br Heart J 1988;59:356–60.

[96] Hornberger LK, Sahn DJ, Kleinman CS, et al.Antenatal diagnosis of coarctation of the aorta:a multicenter experience. J Am Coll Cardiol1994;23:417–23.

[97] Brown DL, Durfee SM, Hornberger LK. Ventricu-lar discrepancy as a sonographic sign of co-arctation of the fetal aorta: how reliable is it?J Ultrasound Med 1997;16:95–9.

[98] Allan LD, Sharland GK. The echocardiographicdiagnosis of totally anomalous pulmonary ve-nous connection in the fetus. Heart 2001;85:433–7.