Screening for heart defects in the first trimester of pregnancy: food for thought

I would like to start this Opinion by exercising the rights of a chairperson and going beyond the Editor-in-Chief's request for me to comment on two papers in this issue of the Journal that deal with the fetal heart in the first trimester of pregnancy1, 2. I do so in order to acknowledge Professor Yves Ville's immense support for me to perform early fetal echocardiography when we worked together many years ago3, as well as to recognize his early vision about the importance of performing (transabdominal) first-trimester cardiac scans in high-risk pregnancies, at a time when this was not common practice, but innovative. To stress his enthusiasm in this important area of fetal medicine is for me a ‘must do’ in this Opinion, but it is also a pleasure to write about first-trimester cardiac scans in Yves's last issue as Editor-in-Chief of Ultrasound in Obstetrics & Gynecology. I should also acknowledge that highly-skilled obstetricians have been performing (transvaginal) first-trimester fetal echo since the beginning of the 1990s4-15, before cardiologists became interested in the fetal heart in early pregnancy. However, use of the transabdominal route for early scans3 and the ever improving ultrasound resolution over the years has not only made it possible for cardiologists to explore the small first-trimester fetal heart but has also paved the way for sonographers, radiographers and other professionals to incorporate basic cardiac views into the routine 11 to 13 + 6-week scan. This has obviously shifted interest in early fetal echo from accurate diagnosis to screening low-risk pregnancies. ‘Business as usual’— I must now return to my task and comment on the two papers in this issue of the Journal. The studies of Sinkovskaya et al.1 and Timmerman et al.2 are both concerned with screening for major congenital heart disease (CHD) in early pregnancy. Yet they explore different aspects of screening. Sinkovskaya and colleagues1 measured the cardiac axis on the four-chamber view in a prospective study of 100 consecutive women scanned at 11 + 0 to 14 + 6 weeks of gestation. Additionally, the outflow tracts were imaged and targeted fetal echocardiography was performed later in the second trimester. In early pregnancy, the mean value for the cardiac axis, based on 94 fetuses with no cardiac abnormalities, was approximately 47° with limits of normality set between 35° and 60°. The four-chamber view was seen in all cases and there was good interobserver reproducibility. An abnormal axis was seen in four of the six cases with CHD (of which three showed a structurally abnormal four-chamber view): one with hypoplastic left heart syndrome, two with atrioventricular septal defects (one unbalanced, one associated with isomerism) and one with tetralogy of Fallot. The conclusion that followed was an obvious one: it is possible to measure the cardiac axis in early pregnancy and this may help to identify pregnancies at risk of CHD. But, in the context of screening, is it really that simple or do we have some ‘food for thought’ here? Whilst the authors report that the four-chamber view was imaged in all cases, it is of note that women with a body mass index (BMI) ≥ 30 were excluded from the study and nearly one in five cases (19%) required a combined transabdominal–transvaginal approach. Thus, for screening purposes, it may be somewhat premature to extrapolate the findings of this study to a large low-risk population that includes women with BMI ≥ 30 and in an environment where it may be less practical to perform transvaginal scans in nearly a fifth of cases. Previous investigators have shown high BMI and small fetal size to have a negative impact on success rates of first-trimester scans16, 17. Timmerman and colleagues2, on the other hand, aimed at refining risk assessment to improve prediction of CHD over and above that associated with an increased nuchal translucency thickness (NT) in chromosomally normal fetuses. In a retrospective study of nearly 800 fetuses, they explored the added predictive value of an abnormal ductus venosus pulsatility index (DV-PIV, above the 95th centile) and abnormal a-wave (consistently absent or reversed). Cardiac defects were present in 35 fetuses, 26 with major forms of CHD. An abnormal DV-PIV in the context of increased NT and normal karyotype conferred a three-fold increase in risk whereas an abnormal a-wave did not add to the prediction following correction for DV-PIV. The sensitivity and specificity of DV-PIV for major CHD were 71% and 61%, respectively. While this approach is important to streamline referrals so that the limited resources (in this case, availability of diagnostic early fetal echocardiography) can be allocated to those families who are at the highest risk, obtaining a technically good DV Doppler signal may not be so straightforward18, bringing into question its widespread use for screening low-risk pregnancies. Factors such as the learning curve19, observer variability and reproducibility of the signal20-22 may have a negative impact on its potential utility as a method of screening for CHD worldwide, as highlighted by the authors. This study also offers more food for thought in that, of all cases of CHD which had fetal cardiac scans and required postnatal intervention, half were false-negative cases. Excluding those with a patent arterial duct and a secundum atrial septal defect, 12 children had cardiac intervention and six were thought to have ‘no anomalies’ on fetal scans2. The reasons behind this finding are unclear. While it is well known that certain cardiac defects evolve during pregnancy and may not be amenable to diagnosis prenatally23, an abnormality of the atrioventricular connection such as tricuspid atresia can (and should?) indeed be diagnosed if a cardiac scan is performed. In one additional case, the diagnosis of atrioventricular and ventriculoarterial discordance was made postnatally. The ISUOG consensus statement regarding ‘What constitutes a fetal echocardiogram’24 reflects the need for a multidisciplinary approach to fetal cardiac scans, which should involve cardiologists and obstetricians alike. Of relevance to both these studies and of utmost importance to early screening for CHD in general is: how soon after identifying markers in the general population, prior to 14 weeks of gestation, are we able to refer the pregnant woman for diagnostic fetal echocardiography? This question goes beyond the studies of Sinkovskaya et al.1 and Timmerman et al.2 as it applies to all potential early markers of CHD (abnormal cardiac axis, increased NT, abnormal DV, tricuspid regurgitation and even aberrant subclavian artery25, 26). Following a suspected cardiac abnormality in mid-gestation, the accepted recommendation in the UK is that the pregnant woman be offered an appointment as soon as possible, but preferably within a week. An unexpected abnormal ultrasound finding leads to parental anxiety, vacillation between emotional confusion and sense of reality. Parents adapt but they need additional information about diagnosis and treatment without delay27. The magnitude of such anxiety may be difficult to measure but is potentially devastating for some families. When the abnormality at stake in the first trimester is a chromosomal defect, resolving the issue may be relatively easy (albeit with a small risk of miscarriage through invasive procedures), but definition of the cardiac anatomy accurately in the first or early second trimester when major CHD is suspected, or following identification of a high-risk pregnancy, is certainly not widely available. For many women, therefore, identifying the risk early but having to wait until the 20-week scan may be associated with a period of excessive stress. Rosenberg et al.28 found, after controlling for race and maternal age, that referral for fetal echocardiography was an independent predictor of maternal ‘state anxiety’, i.e. how patients feel at the time of the scan. Revealing to parents that the fetal heart shows an abnormality is likely to raise maternal and family anxiety levels. Thus, it is important that we think further about the implications of incorporating new markers for CHD in early pregnancy. Identifying the high-risk group but being unable to offer the appropriate means to clarify whether there is an abnormality at an early gestational age may not be appropriate. Furthermore, definition of an abnormality must come hand in hand with diagnostic accuracy. To conclude, identifying cases at risk of CHD in the first and early second trimesters of pregnancy poses two important questions: how soon can early fetal echocardiography be offered to families so that the diagnosis of normality or abnormality can be made, and how accurately can this be achieved?