Using fetal nuchal translucency to screen for major congenital cardiac defects at 10-14 weeks of gestation: population based cohort studyBMJ 1999; 318 doi: http://dx.doi.org/10.1136/bmj.318.7176.81 (Published 09 January 1999) Cite this as: BMJ 1999;318:81
- Jon Hyett, research fellowa,
- Marc Perdu, research fellowa,
- Gurleen Sharland, senior lecturerb,
- Rosalinde Snijders, lecturera,
- Kypros H Nicolaides, professor.a
- aHarris Birthright Research Centre for Fetal Medicine, Department of Obstetrics and Gynaecology, King's College Hospital Medical School, London SE5 8RX
- bFetal Cardiology Unit, Guy's Hospital, London SE1 9RT
- Correspondence to: Professor Nicolaides
Objectives: To examine the utility of measuring fetal nuchal translucency thickness in screening for major defects of the heart and great arteries at 10-14 weeks of gestation.
Design: Population based cohort study.
Subjects: 29 154 singleton pregnancies with chromosomally normal fetuses at 10-14 weeks of gestation.
Setting: Fetal medicine centre in London.
Main outcome measure: Prevalence of major defects of the heart and great arteries.
Results: Of 50 cases with major defects of the heart and great arteries (prevalence 1.7 per 1000 pregnancies) 28 (56%, 95% confidence interval 42% to 70%) were in the subgroup of 1822 pregnancies with fetal nuchal translucency thicknesses above the 95th centile of the normal range. The positive and negative predictive values for this cut off point of nuchal translucency thickness were 1.5% and 99.9% respectively.
Conclusions: Measurement of fetal nuchal translucency thickness—traditionally used to identify fetuses at high risk of aneuploidy—at 10-14 weeks of gestation can identify a large proportion of fetuses with major defects of the heart and great arteries.
The majority of congenital abnormalities of the heart and great arteries are associated with subcutaneous oedema in the neck region at 10-14 weeks of gestation
Subcutaneous oedema in the neck region can be visualised by ultrasonography as increased nuchal translucency thickness
The prevalence of major defects of the heart and great arteries increases with increasing fetal nuchal translucency thickness at 10-14 weeks of gestation
Measurement of fetal nuchal translucency thickness at 10-14 weeks of gestation is a sensitive method of screening for major defects of the heart and great arteries
Cardiac defects are the most common congenital abnormalities. The prevalence is 3-8 per 1000 pregnancies. 1 2 In the United Kingdom, studies on screening women with low risk pregnancies by ultrasonography showed a prevalence of congenital abnormalities of 1.3-3.2 per 1000 pregnancies. 3 4 About half of these defects are asymptomatic, and the other half are classed as major because they are either lethal or require surgery. Specialist fetal echocardiography at around 20 weeks of gestation can identify most of the major cardiac defects, 5 6 but the main challenge in prenatal diagnosis is to identify those pregnancies at high risk that need referral to specialist centres. Currently, screening is based on examination of the four chamber view of the heart during routine ultrasonography at 16-22 weeks of gestation, but a recent study of 7459 pregnancies showed that an abnormal four chamber view can identify only 26% of major cardiac defects.7
A more sensitive method of screening may potentially be the ultrasonographic measurement of fetal nuchal translucency thickness at 10-14 weeks of gestation.8 Extensive studies have recently shown a strong association between cardiac defects and increased nuchal translucency thickness, which is thought to be the result of subcutaneous oedema in the nuchal region.9–11 In four studies reporting on a total of 21 fetuses with major cardiac defects diagnosed at 10-14 weeks of gestation, 17 (81%) had increased nuchal translucency thickness.12–15 In addition, pathological studies of both chromosomally abnormal and normal fetuses with increased nuchal translucency thickness have shown a high prevalence of abnormalities of the heart and great arteries. 16 17 Furthermore, a study of 1389 chromosomally normal fetuses with increased nuchal translucency thickness reported that the prevalence of major cardiac defects increased exponentially with nuchal translucency thickness.18
Our study of 29 154 singleton pregnancies, presumed to be chromosomally normal, examined the utility of measuring fetal nuchal translucency thickness at 10-14 weeks of gestation as a method of screening for major abnormalities of the heart and great arteries.
Subjects and methods
We undertook a retrospective study to examine the prevalence of major cardiac defects in a population of 31 162 singleton pregnancies. The pregnancies were screened for chromosomal abnormalities by a combination of maternal age and fetal nuchal translucency thickness at 10-14 weeks of gestation.
The fetal crown-rump length and nuchal translucency thickness (fig 1) were measured as previously described by transabdominal ultrasonography, unless visualisation was poor in which case vaginal ultrasonography was carried out.8 Demographic details and ultrasound findings including number of fetuses, crown-rump length, and nuchal translucency thickness were entered into a computer database at the time of scanning. Karyotype results and details on pregnancy outcome were added as soon as these became available. Pregnancy outcome was obtained from maternity units, general practitioners, or the patients; the patients were given a questionnaire at the time of the initial scan. If no outcome data were received within two months of the expected date of delivery, we contacted the patients by telephone.
A computer search identified all singleton pregnancies with live fetuses at 10-14 weeks of gestation with a crown-rump length of 38-84 mm and an estimated date of delivery before 1 June 1997. The database was then searched for major abnormalities of the heart and great arteries, diagnosed before or after birth, in the group with a normal fetal karyotype or the birth of a baby with no dysmorphic features suggestive of a chromosomal abnormality.
The prevalence of major cardiac defects was determined and the utility of screening for nuchal translucency thickness including sensitivity, specificity, and positive and negative predictive values were calculated for the nuchal translucency thickness cut off points of the 95th and 99th centiles for crown-rump length. Nuchal translucency thickness normally increases with crown-rump length, and the 95th centile is 2.2 mm for a crown-rump length of 38 mm and 2.8 mm for a crown-rump length of 84 mm; the 99th centile does not change significantly with crown-rump length and is about 3.5 mm.9
Our study group comprised 31 162 pregnancies. We excluded from further analysis 323 pregnancies with chromosomal abnormalities, 317 pregnancies with spontaneous miscarriages, and 1368 pregnancies where the women had changed address and despite repeated attempts through their general practitioners and maternity hospitals, it was impossible to obtain information on pregnancy outcome. In the remaining 29 154 pregnancies the median gestation at the time of screening was 12 weeks (range 10-14 weeks) and the median maternal age was 34 years (range 15-48 years). The fetal nuchal translucency thickness was above the 95th centile for crown-rump length in 1822 (6.3%) pregnancies and above the 99th centile in 315 (1.1%) pregnancies.
Major defects of the heart and great arteries were identified in 50 of the pregnancies. These included 18 pregnancies that were diagnosed antenatally by ultrasonography at 16-31 weeks of gestation (table 1), 13 that were diagnosed at pathological examination after intrauterine death or termination of pregnancy for conditions other than cardiac defects (table 2), and 19 that were diagnosed in live births (table 3).
The prevalence of major cardiac defects was 1.7 per 1000 pregnancies (50/29 154), which increased with nuchal translucency thickness from 0.8 per 1000 for those pregnancies with nuchal translucency thickness below the 95th centile to 63.5 per 1000 for those pregnancies with nuchal translucency thickness above the 99th centile (table 4). There were essentially six groups of cardiac defects; tetralogy of Fallot, hypoplastic left heart, transposition of the great arteries, coarctation of the aorta and aortic stenosis or atresia, ventricular and atrioventricular septal defects, and a sixth group of other complex defects (table 5). Table 6 shows the sensitivity, specificity, and positive and negative predictive values of nuchal translucency thickness cut off points of the 95th and 99th centiles in the detection of major cardiac defects, and figure 2 shows a receiver-operator characteristic curve.
Our study shows that 55% of major abnormalities of the heart and great arteries are associated with increased fetal nuchal translucency thickness at 10-14 weeks of gestation. This method of screening compares favourably with the reported sensitivity of 26% using the four chamber view of the heart at 16-22 weeks of gestation. The clinical implication of our findings is that increased nuchal translucency thickness constitutes an indication for specialist fetal echocardiography. Certainly the overall prevalence of major cardiac defects in such a group of fetuses (about 2%) is similar to that found in pregnancies affected by maternal diabetes mellitus or with a history of a previously affected offspring, which are well accepted indications for fetal echocardiography. 19 20
At present there may be insufficient facilities for specialist fetal echocardiography to accommodate the potential increase in demand if the 95th centile of nuchal translucency thickness is used as the cut off point for referral. In contrast, a cut off point of the 99th centile would result in only a small increase in workload, and in this population the prevalence of major cardiac defects would be very high.
The consequence of combining data from terminations of pregnancy and intrauterine deaths with data from live births would be an overestimate of the true prevalence of cardiac defects in live births from a non-selected population; a proportion of cardiac defects diagnosed antenatally would have resulted in intrauterine death. However, the opportunity to determine such prevalence in “undisturbed” populations was lost in the early 1970s with the advent of prenatal diagnosis. In any case, as shown in the group of patients where the parents chose to continue with the pregnancy after the prenatal diagnosis of major cardiac defects, although the defect obviously persists throughout pregnancy, increased nuchal translucency thickness observed at 10-14 weeks of gestation often resolves by the second trimester and the babies are born alive. Similarly, information on intrauterine deaths is important both for offering a possible explanation for the death and in defining the risk of recurrence. This risk rises from 2% to 10% if there are two rather than one previously affected offspring, and the recurrence is frequently of the same severity or greater severity as the index case.20
The prevalence of major cardiac defects in our population (1.8 per 1000 pregnancies) of presumed chromosomally normal fetuses is compatible with the previously reported prevalence for all cardiac defects in live births (3-8 per 1000 pregnancies); the higher prevalence was reported in studies using echocardiography for diagnosing lesions that are asymptomatic, minor, and even without murmurs. Our study did not include detailed cardiological assessment of all neonates, and since many of the asymptomatic defects would inevitably have been missed, we used major defects as the outcome measure. In contrast to previous studies, our population was confined to those pregnancies presumed to be chromosomally normal; about 25% of affected individuals are chromosomally abnormal. Another factor that may have underestimated our birth prevalence of cardiac defects is the short interval between delivery and obtaining outcome data (in most cases within one month); previous studies have reported that only about 60% of congenital cardiac defects are diagnosed by one month after birth.4
The distribution of different types of cardiac defects in our study (table 6) is similar to that described in previous prenatal and postnatal series. 3 4 7 21 22 Although increased nuchal translucency thickness was observed with all types of major abnormalities of the heart and great arteries, there was a stronger association with left sided defects such as hypoplastic left heart and coarctation of the aorta.
Pregnancies identified by nuchal translucency scanning as being at high risk of cardiac defects need not wait until 20 weeks for specialist echocardiography. Improvements in the resolution of ultrasound machines have made it possible to undertake detailed cardiac scanning in the first trimester of pregnancy. 13 23–25 Furthermore there are several case reports or small series on the ultrasonographic diagnosis of cardiac defects at 10-14 weeks of gestation, and it is interesting that in 21 fetuses with major cardiac defects, 17 (81%) had increased nuchal translucency thickness.12–15
Measurement of fetal nuchal translucency thickness at 10-14 weeks of gestation—traditionally used to identify fetuses at high risk of aneuploidy26—also identifies the majority of pregnancies with major defects of the heart and great arteries.
Contributors: JH, MP, GS, RS, and KHN contributed to the planning and execution of the study and the writing of the paper; KHN will act as guarantor for the paper.
Funding: Fetal Medicine Foundation.
Conflict of interest: None.