Down's syndrome, cardiac anomalies, and nuchal translucency

Papers p 81

Since the first report on the ability of nuchal translucency measurement to detect pregnancies affected by Down's syndrome by Nicolaides et al in 19941 over 20studies have been published on the issue. Despite all these efforts, however, the exact performance of nuchal translucency measurement in detecting Down's syndrome is still unknown. Recent large studies in low risk populations have evaluated the performance of nuchal translucency measurement in detecting Down's syndrome, but the result of the nuchal translucency measurement had already been used in the risk assessment (by identifying cases of Down's syndrome that would never have reached term). In these studies fetuses affected by Down's syndrome which have an increased nuchal translucency are thus more likely to be detected than those affected fetuses with a normal nuchal translucency‐and this may inflate the reported detection rate of nuchal translucency measurement.2^–4 Even the two largest studies reported detection rates as different as 72%5 and 54%.6 As a consequence, the choice between nuchal translucency measurement and serum screening for Down's syndrome remains subject to debate. 78

The paper by Hyett et al in this week's issue reports on the association between nuchal translucency and major defects of the heart and the great arteries (p 81).9 Among 29154 chromosomally normal pregnancies 28out of 50cases with major cardiac defects were detected using the 95th percentile of nuchal translucency as a cut off point. These findings support the hypothesis that increased nuchal translucency may be due to failure of the fetal heart. 10 11 But what are the practical consequences of these findings?

A nuchal translucency above the 95th percentile in a population in which the prevalence of cardiac anomalies is 1-2per 1000 implies a probability of a cardiac anomaly of about 1.5per 1000. Referral of women with nuchal translucency measurements above the 95th percentile to a fetal cardiology unit would imply that 5% of all pregnant women have to be subjected to specialist fetal echocardiography for a chance of 1in 66(28/1850) of finding a cardiac anomaly. It is important to realise that 13of the 50 cardiac anomalies reported by Hyett et al were detected at postmortem examination after intrauterine death or termination of the pregnancy for non-cardiac defects. In these 13cases the only clinical consequence is an increased risk of a cardiac anomaly in a future pregnancy. If we exclude them, 5% of all pregnant women would be subject to specialist fetal echocardiography for a chance of 1in 123 (15/1850) of finding a cardiac anomaly. Increasing the cut off value of nuchal translucency to 3.5mm would reduce the number of referrals to just over 1%, for a chance of having a cardiac anomaly of 1in 17.This may be far more acceptable, although the sensitivity drops under 50%.

When assessing the value of nuchal translucency measurement in detecting cardiac anomalies it is important to realise that at present there is no obstetric intervention for fetuses with cardiac anomalies diagnosed antenatally, other then referring these women to a centre with cardiosurgical facilities for delivery or termination of pregnancy. Moreover, we need to be aware of the anxiety induced in parents by referring them for additional fetal echocardiography. After the detection of increased nuchal translucency in the first trimester most women will initially be offered fetal karyotyping based on their increased risk of Down's syndrome. Once the fetal karyotype is found to be normal, additional echocardiography implies that the parents will be confronted with the possibility of cardiac anomalies.

Apart from the potential clinical implications, the association between nuchal translucency and cardiac anomalies is of interest because a substantial number of live births with Down's syndrome have cardiac anomalies. The results published today suggest that fetal heart failure might play an important part in the association between nuchal translucency and Down's syndrome. If this is so, nuchal translucency measurement might mainly detect those fetuses affected by Down's syndrome that also have cardiac anomalies. One could therefore hypothesise that the fetal loss rate in fetuses affected by Down's syndrome with an increased nuchal translucency is higher than that in fetuses affected by Down's syndrome that have a normal nuchal translucency.

Unfortunately, in virtually all the studies pregnancies affected by Down's syndrome with an increased nuchal translucency were terminated. Nevertheless, the only study that did report on intrauterine lethality in fetuses affected by Down's syndrome in relation to nuchal translucency showed an almost doubled risk of fetal loss in those with an increased nuchal translucency.10 This would have limited consequences if nuchal translucency measurement detected almost all pregnancies affected by Down's syndrome. But if the detection rate of nuchal translucency measurement for Down's syndrome is lower, its potential to detect those fetuses affected by Down's syndrome that would result in the live birth of children with Down's syndrome would not be as good as we believe it is.

In view of this problem it is important to compare the observed number of pregnancies affected by Down's syndrome in a population of women exposed to a screening programme with the number of affected pregnancies that would be expected in the absence of screening, based on the age distribution of the included women. In doing so, Haddow recently showed that the detection rate of nuchal translucency measurement in the UK mulitcentre project on the assessment of risk of trisomy 21was 60% instead of the reported 82%. 57 Extrapolation of the 60% detection rate at 12weeks to a similar rate at term presumes that the risk of fetal loss of a Down's syndrome pregnancy is independent of the nuchal translucency. However, the association between nuchal translucency and cardiac anomalies, as suggested by this week's report, make this presumption questionable.