The fetal heart or the lymphatic system or …? The quest for the etiology of increased nuchal translucency

Today's ease of access to Internet bibliographic searches should lead to ease of literature review as researchers shouldn't need to spend hours going through the library shelves to find the one (or many) important article(s) that need(s) to be read before writing a review article. However, electronic Internet searches are also a ‘catch-22’ as ease of access to information can mean access to excessive information. A careful choice of keywords may not be sufficient to avoid being flooded with hundreds of articles, many of which may not be relevant to the particular topic being studied. If the topic involves nuchal translucency (NT), one can expect that the risk of this happening is high and its 95% confidence limits narrow. In this Opinion, two original articles in this issue of the Journal1, 2 will be addressed, each one dealing with the pathophysiology of increased NT. The relationship between increased NT and structural cardiac abnormalities will also be reviewed and therefore the strategy used for a ‘Pubmed’ search included a combination of the following keywords: nuchal translucency, heart, cardiac, lymphatic, (a)etiology and pathophysiology. Within a few seconds, a staggering almost 500 articles were listed on the computer screen. However, whilst many of these publications addressed various aspects of the association between congenital heart disease (CHD) and increased NT, few considered the etiology of increased NT as their main research question. For didactic reasons, some of the important and historical issues related to the association between CHD and increased NT and the growing awareness of its clinical implication will be presented first. The possible mechanisms of increased NT will then be discussed with emphasis on the papers by Haak et al.2 and Bekker et al.1 who provide us in this issue with a double bill on two of the possible etiological factors linked to increased NT, namely the heart and the lymphatic system. The realization that increased NT is strongly associated with congenital cardiac malformations has been one of the most puzzling discoveries for fetal and perinatal cardiologists in the last decade. This finding, as often happens, was a byproduct of other studies primarily concerned with screening for chromosomal abnormalities. While risks for chromosomal abnormalities are adjusted for maternal age and serum biochemistry, risks for CHD appear to be solely dependent on the degree of NT itself. The initial reports demonstrating the association between NT and CHD described heart defects in chromosomally abnormal fetuses presenting with increased nuchal thickness at 11–14 weeks3-5. This was perhaps not unexpected as trisomic fetuses have a high incidence of such defects. In euploid fetuses with increased NT, however, abnormalities of the heart and great arteries was first alluded to in 1996 when Hyett et al. published a small pathological series in which 19/21 fetuses showed cardiac abnormalities, the commonest being narrowing of the aortic isthmus6. The same group had previously shown, however, that relative isthmic narrowing is a physiological feature of early gestation as demonstrated by an increase in the ratios of isthmus to aortic valve and isthmus to ductus arteriosus from gestational age 9 to 18 weeks7. At present, it is unclear if these early pathological observations have any postnatal clinical significance. Subsequently, however, data from a larger study of nearly 1500 fetuses strongly suggested that increased NT could be a useful marker for major cardiac defects8. This information came from a selected population as 9/24 (38%) fetuses with major CHD had NT ≥ 6 mm. Although not a reflection of the role of NT in screening low-risk pregnancies, these initial observations were compelling evidence of a link that was worth pursuing. This study also suggested that the incidence of CHD dramatically increased with increasing NT. Other investigators have confirmed these initial observations and provided further evidence to support this association9-16. In chromosomally normal fetuses in which NT ≥ 99th centile (i.e. NT ≥ 3.5 mm) the risk of a major cardiac abnormality is around 5–10%10, 11, 15, 17. This is unquestionably higher than the risk observed in many families traditionally referred for fetal echocardiography. Additionally, knowing the NT measurement not only identifies pregnancies at risk but it also provides an assessment of risk stratification according to the absolute measurement of nuchal thickness. Table 1 provides a ‘rule of thumb’ that can be used as practical guidance to such risks expressed in a similar format to risks for chromosomal abnormality. This information can be extremely useful in the counseling process that takes place prior to performing specialized fetal echocardiography for families whose indication is increased NT with normal karyotype. Having passed the first hurdle, namely the exclusion of a chromosomal defect, these families remain at risk of having a fetus with a major CHD. Such risk is by necessity, and at all degrees of increased NT, higher than what is accepted to be ‘a high-risk pregnancy for chromosomal abnormality (i.e. > 1:270)’. This is so because the underlying background risk for major CHD in the general population irrespective of maternal age is around 1 : 250 (i.e. 4/1000 live births). For NT that varies between the 95th and 99th centiles there may also be an increase in risk compared to background risk but evidence for this is less strong. In earlier studies, NT ≥ 95th centile meant NT ≥ 2.5 mm, and in more recent studies the 95th centile is adjusted by crown–rump length. The incidence of major CHD in this narrow range of increased NT was low (∼ 5/1000 pregnancies) in the studies by Hyett et al.8 and Mavrides et al.10 and similar to the background risk for CHD, whereas Ghi et al.11 reported a higher incidence of 25/1000. A recent review that summarizes data from five datasets shows an incidence of 17/1000 for NT ≥ 2.5 mm17. Thus, it appears that the risk in this narrow range of increased NT (2.5 mm ≤ NT < 3.5 mm) does also justify referral for fetal echocardiography. A different question relates to the resource implications and availability of specialists to perform detailed fetal echocardiography in all fetuses with NT ≥ 95th centile (∼ 5% of the population). Having identified families at risk, a natural follow-on question is: if increased NT is a clue for the presence of CHD, can the type of defect be predicted, if there is one? While initial reports suggested that narrowing of the aortic isthmus and a narrowed aorta were commonly seen in fetuses with increased NT6, 18, these initial observations were not backed up by data published subsequently. Many reports have shown that a wide spectrum of abnormalities can be seen in association with increased NT10, 11, 19-22. Using NT screening has certainly increased the ability to identify pregnancies ‘at risk’ from an apparently low-risk population. This is likely to have contributed to better detection rates of CHD in recent years, although no formal comparative studies are available. As the effectiveness of sonographers and obstetricians in examining the four-chamber and great vessel views has continuously improved, it is always difficult to know if cases with increased NT would have been detected anyway at the time of the 20-week scan. However, timing is important. Not only can families at risk be identified by their NT measurement but this process takes place early in the pregnancy—prior to 14 weeks, and much before the conventional time for routine fetal malformation screening—which may vary between 18 and 23 weeks. Early identification of fetuses at risk, coupled with awareness that fetal echocardiography can be performed transabdominally from around 12–13 weeks of gestation23 as well as transvaginally, has contributed to a surge of studies addressing detection of cardiac abnormalities in the first and early second trimester9, 22, 24-30. In performing early fetal echocardiography, however, it is important to remember that a large number of euploid fetuses with increased NT will have a normal heart and that a proportion of these will have significant extracardiac abnormalities. Thus, fetal cardiac assessment should preferably be performed in partnership with early fetal assessment30. A different aspect to be considered relates to the relative contribution of NT to the overall detection rate of CHD in the general population. By identifying high-risk families as those with NT ≥ 95th centile, it must follow that the great majority of fetuses in any unselected population will not present with increased NT (i.e. those with NT < 95th centile). Thus, there is by definition a much larger population of lower-risk families (NT < 95th centile) compared to higher-risk ones (≥ 95th centile) and many fetuses with major CHD will still be seen among those with normal NT. In 1999, data from a large cohort suggested that NT ≥ 95th centile could potentially identify the majority of fetuses with major CHD as 56% (95% CI, 42–70%) of affected fetuses had previously shown increased NT31. This initially reported high detection rate could not be reproduced in unselected populations10, 32. Data from Mavrides et al. showed that the sensitivity of using NT ≥ 95th centile to detect major CHD in unselected pregnancies was low (15%; 95% CI, 4–48%) despite a high incidence of major CHD in fetuses with increased NT10. Lower sensitivities than initially reported have also been shown by other groups12, 14. Yet, one other study reports sensitivity > 50% in low-risk pregnancies16. This information needs to be interpreted in view of the fact that 11/36 diagnoses considered to be major CHD were isolated patent arterial ducts or atrial septal defects, neither of which can be diagnosed prenatally and should not be included in antenatal detection rates. The role of the sonographer still remains the single most important determinant for specialized fetal echocardiography, particularly if views of the great vessels are used33. In summary, the finding of an increased NT in the absence of a chromosomal abnormality is a clear ‘evidence-based’ indication for fetal echocardiography. The evidence is stronger for NT ≥ 3.5 mm (risk of 5–10%) than for NT ≥ 95th centile but < 3.5 mm (risk of 1.7%). Recognition of families at risk offers the possibility of early fetal echocardiography, which in experienced hands can be performed from the end of first trimester with a great degree of accuracy. The incidence of CHD in fetuses with increased NT and normal karyotype varies with NT and approximately one third of fetuses with major CHD can potentially be identified by NT screening17. Increased NT does not predict the type of cardiac abnormality that may be encountered. The longer the list of fetal abnormalities (other than chromosomal defects) associated with increased NT, the wider the spectrum of diagnostic possibilities, and the more difficult it becomes to find a single etiological explanation for increased NT. Not surprisingly, with the CHD–NT link as outlined above, a ‘cardiac etiology’ had to be considered. In addition to a cardiac cause, various other hypotheses have been put forward including abnormalities of the extracellular matrix34, 35, mediastinal compression with venous congestion36 and abnormalities of the lymphatic system. In a recent review on the pathophysiology of increased NT, Van Vugt and Haak discussed in great depth various pertinent issues that need not be duplicated here37. From this point onwards specific consideration will be given to the heart and the lymphatic system together with the data presented by Haak et al.2 and Bekker et al.1. In 1995, shortly after the first publication on increased NT and chromosomal abnormalities38, Jackson et al. were the first to consider a pathological correlation between an 11-week fetus with trisomy 18 and ultrasonographically detected increased NT. They concluded that ‘although the etiology of increased NT remains unclear, it does not appear to be lymphatic or cardiac in origin’39. Yet, as Yves Ville indicates in his Opinion ‘Ten years on and still a pain in the neck?’40, cardiac or lymphatic origin, both or neither, the quest for the etiology of increased NT continues. When an association between any two entities is discovered, it cannot necessarily be assumed that there is a cause–effect relationship. One must first establish that there is biological plausibility and that a temporal relationship also exists (i.e. the presumed cause precedes the effect). It is quite possible that having found a strong association between two entities, both may be related to another, as yet unknown, third factor which may ultimately turn out to be the sought after explanation for the association. Structural heart defects have been initially considered as the possible cause of increased NT. Morphological studies in chromosomally abnormal18 and subsequently in euploid fetuses6 have initially suggested that ‘narrowing of the aortic isthmus’ (below the 5th centile) may lead to increased NT as a narrow isthmus was identified in a significant proportion of cases with increased NT. This was associated with a larger aorta in trisomy 21 fetuses and with a narrow aorta in euploid fetuses. Possible explanations were discussed involving overperfusion of the head and potential development of coarctation of the aorta. Clinical studies have not demonstrated coarctation to be the main cardiac lesion in fetuses with increased NT nor has there been any study linking any specific cardiac abnormality to increased NT. There remains, therefore, no evidence to support the theory that increased NT is a direct consequence of a structural cardiac abnormality. Instead, a variety of cardiac lesions leading to different hemodynamic patterns can be encountered in fetuses with increased NT irrespective of the size of the aorta or aortic isthmus. It remains difficult, therefore, to explain why NT should be increased as a result of CHD. Additionally, most fetuses with increased NT will ultimately have a structurally normal heart. An alternative explanation for the growing evidence of the association between NT and CHD is heart failure. This has been proposed as a possible mechanism because of a potential strain that cardiac abnormalities impose on the heart during early gestation6, 8. Most of the subsequent evidence for heart failure is indirect and comes from studies of the ductus venosus Doppler signal. The Doppler pattern of the ductus venosus waveforms with low, absent or reversed end-diastolic velocities in chromosomally normal and abnormal fetuses41-46 have been attributed to cardiac failure, possibly as a result of associated structural cardiac defects. It still remains difficult to explain why a heart abnormality that is normally well balanced in utero should lead to cardiac decompensation in early gestation. It is conceivable that in some instances ‘heart failure’ may reflect impaired mechanisms of fetal adaptation to CHD such as a restrictive foramen ovale in the setting of tricuspid atresia or a restrictive ductus arteriosus in the presence of aortic atresia. If that were the case, any manifestation of heart failure would be likely to progress (as opposed to regress). These cases may be the ones deemed to die in utero. The work by Haak and colleagues2 is the most recent of a small series of studies in which attempts have been made directly to assess cardiac function in fetuses with increased NT47-49. A total of 85 fetuses with normal NT and another 45 with NT > 95th centile were studied by pulsed-wave Doppler at 11–14 weeks of gestation. The population was mixed and included fetuses with normal and abnormal karyotype and fetuses with normal hearts as well as with CHD. Mitral and tricuspid valve absolute velocities during early and late diastole were measured and E/A ratios calculated. No significant differences were encountered in velocities and velocity ratios between fetuses with normal and fetuses with increased NT. Regarding absolute velocities, Haak et al. showed that aneuploid fetuses had statistically significantly lower e- and a-wave velocities and these could not be accounted for by the presence of CHD as no difference was found between fetuses with and without CHD. Based on their findings, Haak and her colleagues concluded that this study contradicts the theory that impairment of cardiac function or atrial contraction is the cause of increased NT. Being more cautious, I would rather conclude that this study provides no evidence to support a cardiac origin for increased nuchal thickness as a number of considerations have to be taken into account. Assessment of cardiac function postnatally or in fetal life is not straightforward. Non-invasive methods used in the fetus usually include M-mode echocardiography (e.g. for calculation of shortening fraction) and Doppler techniques. In general, reproducibility of both these methods in the fetus is poor47. for the tricuspid and et al. have also observed wide limits of when and of the E/A there are two physiological observations that have to be considered when Doppler velocities the These further the presence of a patent foramen ovale in the fetus of Doppler velocities the For a valve may no significant increase in velocities because and atrial may as a consequence of a wide patent foramen Doppler velocities not to of velocities in early and late diastole (i.e. normal and may not be associated with normal through the or tricuspid This is better by the the on the Doppler signal. This is in and 9 of the by Haak et Doppler to and vessels but information on is before one can If Doppler velocities are used as for cardiac the must be in Haak and are in with those of et al. who also to differences in the E/A between fetuses with normal and those with increased NT when measured at 11–14 Additionally, assessment of cardiac function using the not support being present in fetuses with increased other papers provide data on cardiac function in fetuses with increased In function was in the second trimester of at a time when NT had and colleagues found a lower E/A and lower of to time velocity for tricuspid and in euploid fetuses with increased NT compared to a There were no differences in of and studied cardiac function in fetuses with septal defects and heart by and in fetuses that previously had normal or increased NT. No difference was observed between the two the data to is In that data (as used for fraction) are the information in the pathological observations to out both cardiac and lymphatic for increased While the remains there appears to be growing evidence for the The work by Bekker and in this issue of the is a study to by the presence and of lymphatic in fetuses with normal and increased NT. This study the same earlier work on the of the in normal and trisomy as well as pathological from two fetuses with increased In that a was found in the nuchal together with both in abnormal to trisomy and in fetuses with increased NT. The were also seen by in a large proportion of fetuses with increased nuchal thickness. were reported by et al. who the in a normal fetus with normal NT thickness They suggested that the seen on to two lymphatic seen as a single considered a possible in the development of the lymphatic vessels in the to be a possible cause of increased NT. Bekker and work shows compelling evidence to support this data from fetuses at 11–14 weeks with increased NT were compared with data from fetuses with normal NT The of the was then studied by and of the were calculated. They found a significantly higher of in fetuses with increased NT compared with those in the group This strong association between increased NT and was almost to fetuses with abnormal NT, irrespective of the presence or absence of a chromosomal abnormality. The association with the lymphatic system appears to be much stronger than that reported with the heart. A temporal relationship has been demonstrated in as reported by the same but it remains to be shown in fetuses. according to the this theory the and of increased NT and may provide the link between different pathological It is that the increased of in fetuses with increased NT is to a in the of the lymphatic system. The are the first of the lymphatic system to and a in such development would lead to in the As the process is only the is when the to the venous system. In however, there remains one question the of increased if of in the in the is primarily to of from a development of the why there a need for such a mechanism in the in the in a similar to that in the The also that a similar process of and impaired development could provide the link between increased NT and cardiac abnormalities as pathological may a role in the development of structural cardiac abnormalities. This is a development has also been to by and as a explanation for the association between increased NT in the first trimester and higher risk of cardiac in the second If the reported by Bekker et al.1 can be reproduced on a larger and to the theory which involves not only the lymphatic but also the then a may when an cause for increased NT will all diagnostic this will include those with abnormalities and may an explanation for increased NT in those who ultimately are normal at the end of