Comparison of echocardiographic findings in fetuses at less than 15 weeks' gestation with later cardiac evaluation

Authors

  • V. Zidere,

    Corresponding author
    1. Harris Birthright Research Centre for Fetal Medicine, King's College Hospital, London, UK
    2. Department of Congenital Heart Disease, Evelina Children's Hospital, London, UK
    • Correspondence to: Dr V. Zidere, Harris Birthright Research Centre for Fetal Medicine, King's College Hospital, London, Denmark Hill, London SE5 9RS, UK (e-mail: vita.zidere@nhs.net)

    Search for more papers by this author
  • H. Bellsham-Revell,

    1. Department of Congenital Heart Disease, Evelina Children's Hospital, London, UK
    Search for more papers by this author
  • N. Persico,

    1. Harris Birthright Research Centre for Fetal Medicine, King's College Hospital, London, UK
    2. Department of Obstetrics and Gynecology ‘L. Mangiagalli’, Fondazione IRCCS Ca' Granda, Ospedale Maggiore Policlinico, Milan, Italy
    Search for more papers by this author
  • L. D. Allan

    1. Harris Birthright Research Centre for Fetal Medicine, King's College Hospital, London, UK
    Search for more papers by this author

ABSTRACT

Objective

To compare the echocardiographic findings in fetuses at < 15 weeks' gestation with a later follow-up scan, and determine both the accuracy of early diagnosis and the frequency of findings that change between scans.

Methods

For the period 2002 to the end of 2009, we searched our database for all patients who had had a detailed fetal echocardiogram at less than 15 completed weeks' gestation and a repeat scan at least 6 weeks later.

Results

Of 1200 patients fulfilling our selection criteria, the cardiac findings were normal at both scans in 1069. In 46 cases the same cardiac abnormality was seen at both scans. There was a false-positive diagnosis at early scan in seven cases. In 50 cases, there were mildly abnormal functional findings early in pregnancy with no abnormality found later. In 29 fetuses, there was discordance between the early and later morphological diagnosis, 15/29 being considered significant differences, with 10/15 representing true progression of findings between the early and later scans, rather than missed or incorrect diagnoses.

Conclusions

A high degree of accuracy in the identification of congenital heart disease can be achieved by early fetal echocardiography (sensitivity 84.8 (95% CI, 75.0–91.9)%, specificity 95.3 (95% CI, 93.9–96.4)%), although the identification of every case of tetralogy of Fallot and small atrioventricular septal defects presents particular diagnostic challenges at this gestational age. A small but significant group showed progression of findings during this stage of rapid fetal heart growth, particularly in obstructive lesions. Copyright © 2013 ISUOG. Published by John Wiley & Sons Ltd.

INTRODUCTION

Echocardiography has been used for the examination of the fetal heart for almost 30 years[1, 2]. This has allowed documentation of the normal features of the fetal heart, and recognition of almost all the structural malformations that are seen in postnatal life with a high degree of accuracy[3-7]. The ability to observe the fetal heart sequentially throughout gestation has shown that, although rare, some defects change or develop between the mid-trimester and term[8]. This progression occurs particularly in valve stenosis, where an apparently normal aortic or pulmonary valve at 20 weeks' gestation can become moderately or even severely stenosed by the end of pregnancy. In critical aortic stenosis with a dilated left ventricle at 20 weeks, the ventricle can even become hypoplastic with advancing gestation[9, 10]. Other conditions, such as complete heart block, cardiomyopathies and tumors, do not usually develop until the second half of pregnancy[8]. However, the vast majority of major cardiac malformations are present and detectable at the usual time for routine obstetric screening, which is most commonly scheduled between 18 and 22 weeks.

The recognition of increased nuchal translucency (NT) thickness at 11–14 weeks as a marker for chromosomal anomalies and the later recognition of this finding associated with an increased incidence of congenital heart defects has led to anatomical evaluation of the fetus, including evaluation of the fetal heart, at an early gestational age[11-13]. Despite the small size of anatomical structures at this stage in pregnancy, the combination of high-resolution transducers, color-flow mapping and operator experience has led to the identification of many cardiac malformations at this stage[14-16]. The majority of the fetuses examined echocardiographically in early gestation at our tertiary fetal medicine referral center are karyotypically abnormal; this generally results in termination of pregnancy with no postmortem confirmation, although it was possible to examine a small series of autopsy specimens early in our experience[14]. As this is a time of rapid growth of the heart (the aorta increases in diameter from approximately 1.0 mm to 3 mm between 11 and 20 weeks) and as it has been previously described[17] that some defects change in morphological terms between these gestational ages, we evaluated a series of sequential scans in order to determine both the accuracy of our early diagnosis and the frequency of change during this period.

METHODS

All fetuses with detailed echocardiography performed by a specialist fetal cardiologist between 11 and 14 + 6 weeks, and that had a follow-up scan at least 6 weeks later, were identified from our database between 2002 and 2009. Although 13 + 6 to 14 + 6 weeks is not strictly considered ‘first trimester’, we included these cases as many were examined during this time period. The majority of patients were examined transabdominally, using an Acuson Aspen Advanced (Acuson, Mountain View, CA, USA) with a 4–7-MHz sector probe, a Voluson E8 (GE Medical Systems, Zipf, Austria) with a 4–8-MHz sector probe or a Siemens/Acuson Sequoia 512 ultrasound machine (Siemens, Erlangen, Germany) with a 9-MHz linear transducer. Transvaginal imaging was also used in a small number of patients for whom the transabdominal images were considered inadequate.

The fetal heart was assessed by sequential analysis at both early and second-trimester studies, although pulmonary veins were not consistently identified in every early scan. The four-chamber view and outflow tracts were imaged using two dimensional (2D), and color Doppler in all cases, with pulsed Doppler across all four valves in the majority of cases.

The indication for referral and the NT thickness were recorded. In cases in which there was a discrepancy in diagnosis, the first scan was reviewed retrospectively.

Information regarding prenatal or postnatal outcome, including cardiac surgery and additional extracardiac anomalies, was obtained from the tertiary hospital in which postnatal cardiac assessment was performed or by direct communication with the general practitioner or parents.

Criteria for inclusion in all of the groups studied were: gestation between 11 and 14 + 6 weeks at the time of the first assessment, and between 19 and 24 weeks at the time of the second assessment and imaging considered technically adequate to allow sequential analysis of the cardiac connections on 2D and color-Doppler evaluation.

The groups were further defined as ‘Concordant’, where there were either the same normal or the same abnormal findings at both cardiac scans or ‘Discordant’ where there were different abnormal findings at both scans or abnormal findings at only one of the scans.

Statistical analysis

Sensitivity, specificity, positive predictive value and negative predictive value were calculated for both the ability of early ultrasound to detect any congenital heart disease (CHD) and for the detection of the same CHD and its severity at the second scan.

RESULTS

One thousand two hundred fetuses that met the inclusion criteria were identified. The NT thickness was known in 1193/1200 fetuses and the majority (69%) of fetuses identified were seen at this early gestational age because of increased NT. The indication for referral in the remainder included: a family history of CHD, miscellaneous extracardiac malformations and cotwins where there was an abnormality in the other twin. Figure 1 shows the results of the first and second scans.

Figure 1.

Flow-chart showing results of first and second ultrasound scans in 1200 mainly high-risk pregnancies. *One case of left atrial isomerism was counted twice as there was discordance due to an error in diagnosis and further discordance due to progression from sinus rhythm to complete heart block. CHD, congenital heart disease; MR, mitral regurgitation; TR, tricuspid regurgitation.

Concordance

Both scans normal

The fetal heart was considered normal at both scans in 1069 fetuses (mean gestational age at first and follow-up scans 12.7 and at 20.0 weeks, respectively). At later follow-up in this group there were 10 intrauterine deaths and 18 terminations of pregnancy for extracardiac reasons (with no autopsy confirmation in these cases), and five cases were lost to follow-up. Four fetuses in this group were found to have cardiac features of Noonan's syndrome later in the pregnancy. All remaining fetuses (n = 1032) had a normal heart at delivery.

Both scans same diagnosis of CHD

In 46 patients the same morphological CHD and severity were seen on both scans (Table 1) (mean gestational age at first and second scans, 12.8 and 20.0 weeks, respectively). There were nine subsequent intrauterine deaths and two terminations of pregnancy. One of the intrauterine deaths had postmortem confirmation of transposition of the great arteries. There was therefore confirmation of the initial diagnosis in 36/46 cases.

Table 1. Concordant abnormal findings at early and later cardiac ultrasound scans in 1200 mainly high-risk pregnancies (n = 46)
Cardiac abnormalityn
  1. ARSA, aberrant right subclavian artery; AVSD, atrioventricular septal defect; CHB, congenital heart block; CoA, coarctation of the aorta; DORV, double outlet right ventricle; HLHS, hypoplastic left heart syndrome; IVS, intact ventricular septum; LAI, left atrial isomerism; MR, mitral regurgitation; PA, pulmonary atresia; RAA, right aortic arch; RAI, right atrial isomerism; TGA, transposition of the great arteries; TOF, tetralogy of Fallot; TR, tricuspid regurgitation; VSD, ventricular septal defect.

CoA11
HLHS3
RAA, right arterial duct1
Idiopathic dilatation of ascending aorta, MR, TR1
LAI, AVSD, CHB1
LAI, VSD, CHB1
RAI, AVSD, DORV1
AVSD1
AVSD, TOF3
ARSA2
TR9
TGA3
TOF4
TOF, PA, RAA1
TOF, RAA2
PA, IVS1
Cardiomegaly, cardiomyopathy1

Discordance

There was discordance between the two scans in 85 patients. In 50 of these, there was mild tricuspid regurgitation (n = 44), mild mitral regurgitation (n = 3) or both (n = 3) at the early scan and no abnormal findings at the second scan.

In a further three cases, CHD was suspected at the first scan, but not confirmed at the second scan (false-positive diagnoses) and, in four cases, less complex CHD than originally suspected at the first scan was found at the second scan (Table 2).

Table 2. Discordant abnormal findings at early and later cardiac ultrasound scans in 1200 mainly high-risk pregnancies (n = 7)
First-trimester findingSecond-trimester finding
  1. a

    This case appears in both Tables 2 and 3, as there was a false-positive diagnosis of AVSD, a missed diagnosis of possible coarctation of the aorta and progression of sinus rhythm to complete heart block in this same case. AVSD, atrioventricular septal defect; CHB, congenital heart block; CoA, coarctation of the aorta; LAI, left atrial isomerism; MR, mitral regurgitation; PLSVC, persistent left superior vena cava; TR, tricuspid regurgitation; VSD, ventricular septal defect.

Possible VSD, TR, MRNormal
Possible VSD, TRNormal
Possible CoANormal
Possible AVSD, disproportion, MRVSD
Possible CoA, TRPLSVC
AVSD, possible CoAAVSD
LAI, AVSDaLAI, possible CoA, CHB

In 29 cases an abnormality was found at the second scan that had not been detected at, or that had changed in severity since, the early study (Table 3); 15 of these cases were felt to be significant, as they affected antenatal counseling. Late follow-up in this group found five intrauterine deaths, six terminations of pregnancy and one patient lost to follow-up. There was postnatal confirmation of the 20-week diagnosis in the remaining 73 patients.

Table 3. Discordant findings at early and later cardiac ultrasound scans in 1200 mainly high-risk pregnancies, in which a new abnormality was found, or there were changes in diagnosis, at the second scan (n = 29)
First-trimester findingSecond-trimesterfindingn
  1. a

    Cases in which change in diagnosis was considered significant.

  2. b

    This case appears in both Tables 2 and 3, as there was a false-positive diagnosis of AVSD, a missed diagnosis of possible coarctation of the aorta and progression of sinus rhythm to complete heart block in this same case. ARSA, aberrant right subclavian artery; ALSA, aberrant left subclavian artery; AS, aortic stenosis; AVSD, atrioventricular septal defect; CHB, congenital heart block; CHD, congenital heart disease; CoA, coarctation of the aorta; DORV, double outlet right ventricle; IVC, inferior vena cava; IVS, intact ventricular septum; LAI, left atrial isomerism; MR, mitral regurgitation; PA, pulmonary atresia; PLSVC, persistent left superior vena cava; PR, pulmonary regurgitation; PS, pulmonary stenosis; RAA, right aortic arch; TGA, transposition of the great arteries; TOF, tetralogy of Fallot; TR, tricuspid regurgitation; VSD, ventricular septal defect.

NormalARSA2
NormalPLSVC1
NormalVSD1
NormalVSD, PLSVC1
NormalaCoAa2
NormalaAVSDa2
NormalInterrupted IVC, possible VSD1
NormalaTOFa2
TRVSD1
TRRAA1
TRaCoAa1
Moderate TRaPA, IVSa1
Moderate TRaSevere Ebstein's anomaly, PAa1
Moderate TR, possible PSMild Ebstein's anomaly and PS1
RAARAA, ALSA, possible VSD1
Possible arch abnormalityCoA, VSD1
LAI, AVSDa, bLAI, possible CoA, CHBa1
AVSDaAVSD, CoAa1
TGATOF1
TOFTGA1
DORVAVSD, DORV, RAA, PLSVC1
Complex CHDTOF1
Mild PS, PRaModerate PSa1
Possible AS, TR, MRaCritical ASa1
Ectopia cordisaPentalogy of Cantrella1

Discordance due to error in interpretation

There were 15 cases in which there was discordance due to error in interpretation, five of which were considered straightforward missed diagnoses. Two cases of tetralogy of Fallot were not detected at the first scan but were seen at the later study. Retrospective review of the recorded material obtained early in the pregnancy showed that, although aortic override was not very obvious, mild hypoplasia of the pulmonary artery had been overlooked on the initial scan. It is of note that a correct diagnosis of tetralogy of Fallot was made in eight cases at the early scan, so this is a diagnosis that can be made accurately in the majority of cases. In a further two cases, an atrioventricular septal defect (AVSD) was not recognized in the early scan, but at 20 weeks small defects were seen and on retrospective review, the diagnosis might have been suspected at the initial scan. Again, this should be compared with a correct positive diagnosis of AVSD in five cases. In the last case, Ebstein's anomaly and pulmonary atresia was found at 20 weeks, but was evident on review of the earlier images.

Discordance due to morphological progression

There were 15 cases in which discordance was due to morphological progression, in 10 of which there appeared to be a significant change in the nature or severity of the cardiac malformation as pregnancy progressed. In two cases the heart had appeared normal at the early scan but there was clear disproportion of the ventricles and great arteries at 20 weeks with hypoplasia of the transverse aortic arch, indicating coarctation of the aorta. In a further three cases, coarctation was suspected as an additional diagnosis at 20 weeks after the finding at the early scan of tricuspid regurgitation (n = 1), left atrial isomerism (n = 1) and an ASVD (n = 1). The diagnostic features of coarctation were not evident on review of the early scan in these five cases. For comparison, a correct diagnosis of coarctation was made at the early scan in 11 cases, emphasising that this diagnosis is usually possible at 12 weeks' gestation, but perhaps the less severe cases are not evident until later.

Ventricular and great artery disproportion with hypoplasia of the transverse aortic arch suggest the prenatal diagnosis of coarctation of the aorta. The degree of disproportion and hypoplasia of transverse aortic arch indicates increasing security of this diagnosis. Therefore, these patients were categorized into three groups: ‘disproportion’ (often only ventricular) ‘possible coarctation’ and ‘coarctation’, where the abnormal findings were increasingly pronounced, the last group having the most definitive diagnosis.

A case with ectopia cordis, in which the heart became more contained within the thoracic cavity as pregnancy advanced and progressed to pentalogy of Cantrell, was the subject of a previous report[18]. In one case of left atrial isomerism, there were three changes in diagnosis: the heart rate was slower than normal at 120 beats per minute, with one-to-one atrioventricular conduction at the early scan, but this fetus had developed complete heart block by 20 weeks. In the same fetus, an ASVD was suspected at the initial scan along with an interrupted inferior vena cava, but the AVSD was not confirmed at 20 weeks, although disproportion and a possible coarctation lesion were found.

In two of the 10 fetuses with significant changes between scans there was holosystolic tricuspid regurgitation at the early scan. In one case, right atrial dilatation and moderately severe tricuspid regurgitation were evident at 12 weeks but there was normal size and filling of the right ventricle, biphasic tricuspid valve flow and forward flow in the pulmonary artery on color-flow mapping and pulsed-Doppler imaging. At 20 weeks, tricuspid regurgitation was still evident but the right ventricle was small and thick-walled with no forward flow across an atretic pulmonary valve but reversed flow in the arterial duct, indicating classic pulmonary atresia with an intact septum. In the other fetus with tricuspid regurgitation, the source of this was seen to be mild Ebstein's malformation of the tricuspid valve in the later scan, a morphological malformation that had not been appreciable earlier. In this same fetus, a velocity of 56 cm/s had been noted in the pulmonary artery at 12 weeks, which is just above the normal range for the arterial valves of 40 cm/s at this gestation (authors' own data). The pulmonary velocity was 149 cm/s at 20 weeks, indicating pulmonary stenosis.

In another fetus, mild tricuspid and mitral regurgitation were found at the early study and the aortic velocity was above the normal range for 12 weeks, at 72 cm/s. However, there was equal ventricular filling and equal biventricular function. At 20 weeks there was a typical picture of critical aortic stenosis (Figure 2) with a dilated poorly contracting left ventricle, little forward flow across the aortic valve and reversed flow in the aortic arch. Finally, there was one fetus with marked pulmonary regurgitation at 12 weeks with a mild increase in pulmonary velocity (60 cm/s), but this fetus had no regurgitation and moderate pulmonary stenosis (pulmonary velocity 250 cm/s) at 20 weeks (Figure 3).

Figure 2.

Ultrasound images showing progression from mild to severe aortic stenosis. (a) Four-chamber view with color Doppler applied showing equal ventricular filling at 12 weeks' gestation. (b) Pulsed Doppler image showing that velocity across aortic valve (arrow) is increased for gestational age, at 72 cm/s. Note that flow in the aorta is not aliased in the first trimester at usual cardiac color-flow settings, which were the highest possible for the transducer. (c) Four-chamber view showing dilated left ventricle that was poorly contracting on the moving image at 20 weeks. (d) Severely dilated left ventricle with little flow across the mitral valve at 20 weeks. (e) Reversed flow in transverse arch confirming severe second-trimester aortic stenosis. L, left; LA, left atrium; LV, left ventricle; PA, pulmonary artery; R, right; RA, right atrium; RV, right ventricle.

Figure 3.

Color (a,b,c) and pulsed (d) Doppler ultrasound images showing evolution of pulmonary stenosis. (a) Unaliased forward flow in pulmonary artery at 12 weeks' gestation. (b) Pulmonary regurgitation during same study. (c) Aliased flow in pulmonary artery showing pulmonary stenosis at 20 weeks. (d) Increased velocity in pulmonary artery indicative of moderate second-trimester stenosis. Ao, aorta; L, left; PA, pulmonary artery; R, right.

Note that one case of left atrial isomerism, indicated in Tables 2 and 3, has been counted twice as there was discordance due to an error in diagnosis and further discordance due to progression from sinus rhythm to complete heart block.

Diagnostic accuracy

Sensitivity, specificity, positive predictive value and negative predictive value for both detection of any CHD and detection of the same CHD and its severity at the second scan are shown in Table 4. When looking at the detection of any CHD, early ultrasound is very specific (95.3%), with a good negative predictive value (98.9%), reasonable sensitivity (84.8%) and poor positive predictive value (55.8%). For concordance in both type and severity, again early ultrasound is very specific (94.9%), with a good negative predictive value (97.4%) but with poorer sensitivity (61.3%) and low positive predictive value (44.7%).

Table 4. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) for detection of any congenital heart disease (CHD) at first ultrasound scan and detection of the same CHD and its severity at second scan in 1200 mainly high-risk pregnancies
ParameterSensitivity (95% CI) (%)Specificity (95% CI) (%)PPV (95% CI) (%)NPV (95% CI) (%)
Detection of CHD at first scan84.8 (75.0–91.9)95.3 (93.9–96.4)55.8 (46.5–64.9)98.9 (98.1–99.4)
Concordance between scans61.3 (49.4–72.4)94.9 (93.5–96.1)44.7 (34.9–54.8)97.4 (96.2–98.2)

DISCUSSION

The increasing trend towards examining the fetal heart around the time of the NT scan (between 11 and 14 weeks' gestation) means that evaluation of the sensitivity and specificity of this early scan is paramount. Early ultrasound showed high specificity with a good negative predictive value for the detection of any CHD, while early ultrasound and a repeat scan 6 weeks or more later also showed high specificity with a good negative predictive value for concordancy of the type and severity of CHD on both scans. Our findings are consistent with those of Rasiah et al.[19], who in a systematic review collated data from a similar number of patients, reported in the literature from 10 centers, and found sensitivity and specificity of 85 and 99%, respectively. Our sensitivity was reasonable for the detection of any CHD but poor for concordance of type and severity. This is due to the natural progression of some types of lesions over time, as shown in this cohort.

A correct diagnosis of normality at both scans was possible in the majority of cases and most abnormalities were correctly diagnosed at the initial scan. It was encouraging that there were few false positives and, with the exception of two suspected ASVDs that turned out to have less severe disease later (one ventricular septal defect (VSD) and one coarctation), all false positives were minor defects for which intervention in the pregnancy would not have been considered.

In 50 cases, minor abnormal findings that were considered functional, mainly mild tricuspid valve regurgitation, had disappeared by 20 weeks. The explanation for this is not clear but chronologically it coincided with the disappearance of the excess nuchal fluid. Although atrioventricular valve regurgitation is known to be more common in fetuses with increased NT, those with karyotype abnormalities and those with CHD, most commonly it is a benign, transient finding[20].

Some of the changes in diagnosis at 20 weeks were relatively minor, such as the recognition of a VSD, an aberrant right subclavian artery or persistent left superior vena cava in later pregnancy (Table 3). However, 15/29 (51.7%) represented major changes in diagnosis, or changes that would affect the information and advice given at counseling. In 10 cases, the defect had clearly progressed in nature during this time interval when there is rapid growth of the heart. Change of this kind has also been described by other observers[14]. It is noteworthy that in the three cases with suspected mild valve stenosis in early pregnancy (two pulmonary and one aortic) the color-flow map was completely normal and the abnormality was suspected only by finding that the pulsed-Doppler velocity across the arterial valve was just greater than the normal range for this gestational age. This suggests that the color-flow map in early gestation should not be relied on to exclude mild valve stenosis, unless the pulse repetition frequency is set so that aliasing will occur at velocities greater than 40 cm/s and pulsed Doppler is used as routine in both great arteries.

There was a significant change in the findings in this series in about 1/120 cases examined, which is a higher rate of change in findings than would be expected between 20 weeks and term. It is important to note, however, that almost all the cases (except cotwins) that we studied were at high risk, usually because of increased NT, and therefore this rate may not apply to a low-risk or screened population. The transvaginal approach was used when a diagnosis was in doubt. This sometimes, but not always, proved to be an additional diagnostic aid and we would recommend that this modality (with a sonographer skilled in the technique) be available to the early fetal echocardiographer.

Although it was not the focus of our study, it is of interest that six fetuses had persistent NT and therefore the heart was re-evaluated in later pregnancy for suspected Noonan's syndrome. In four of them there were abnormal cardiac findings consistent with this syndrome at 28 weeks, despite normal findings at the early and 20-week scans. All six proved to have Noonan's syndrome and some form of cardiac involvement (pulmonary stenosis or hypertrophic cardiomyopathy) after birth.

In conclusion, early fetal echocardiography between 12 and 14 completed weeks' gestation is highly specific for the detection of CHD. The identification of every case of tetralogy of Fallot and small ASVDs presents particular diagnostic challenges in early fetal echocardiography. Changes in severity appear to occur particularly in obstructive lesions at the aortic and pulmonary valves and in the aortic arch and there are limitations on accuracy with more subtle malformations.

ACKNOWLEDGMENT

We are grateful to Dr Shane M. Tibby, Evelina Children's Hospital, Guy's and St Thomas' NHS Foundation Trust, for his kind assistance with the statistical analysis.

Ancillary