Prenatally detectable congenital heart defects in fetuses with Down syndrome




To document the incidence of congenital heart defects (CHD) that are detectable echocardiographically in the fetus with trisomy 21 and the relationship with nuchal translucency, fetal sex and ethnicity.


Data on fetuses with a karyotypic diagnosis of trisomy 21 were collected between January 2002 and March 2010. The data were analyzed for the gestational age at examination, maternal age, reason for referral for fetal echocardiography, cardiac diagnosis, fetal sex, ethnicity and outcome.


Of 917 fetuses with trisomy 21, 487 had a diagnostic echocardiogram. Cardiac examination was performed before 14 weeks' gestation in 75% of cases. The main reasons for referral were increased nuchal translucency (NT) in 76% of cases, suspected cardiac abnormality in 15% and an extracardiac anomaly in 6%. Structural CHD was found in 164/487 (34%), or 98/412 (24%) if those referred for suspected CHD are removed from the analysis. The most common diagnosis was atrioventricular septal defect (AVSD) (115/487, 24%). The ratio of female to male fetuses with AVSD was 29%:18% (P = 0.003). There was no difference in the incidence of AVSD with ethnicity. The pregnancy continued in 36 cases, but three were lost to follow-up; of the known outcomes there were 10 intrauterine deaths, six of which had structural heart disease, and 23 live births, 15 of which had CHD.


Most fetuses (66–76%) with trisomy 21 have a structurally normal heart on echocardiography. The presence of structural CHD was not associated with increased NT. The increased incidence of AVSD in females was confirmed in our study, although an ethnic difference could not be confirmed. CHD does not appear to increase the chance of spontaneous intrauterine loss in ongoing pregnancies. Copyright © 2011 ISUOG. Published by John Wiley & Sons, Ltd.


It is commonly suggested in the literature that most children or fetuses with Down syndrome have congenital heart disease (CHD). Data from detailed examinations of affected children contradict that, however, with CHD found in only about 40% of cases1–6. An atrioventricular septal defect (AVSD) is well recognized as the most common form of CHD, occurring in about 20% of the total Down group or half the group with CHD. Ventricular septal defects (VSD), secundum atrial septal defects (ASD) and persistent arterial duct (PDA) make up the majority of the rest of the cases of CHD in children7–9. Some VSDs are large and therefore might be detectable prenatally but ASDs and PDAs, of course, are not detectable in the fetus. Tetralogy of Fallot and coarctation occur in small numbers in postnatal series and one would expect these to be detectable in the fetus. Confounding the data concerning the incidence of CHD postnatally is that an aberrant right subclavian artery (ARSA) is much easier to detect (if looked for) in the fetal than in the postnatal echocardiogram10–12. This abnormality therefore is likely to be under-reported in postnatal echocardiography series, although it has been reported in earlier catheter and postmortem series in Down syndrome1, 13–16.

Our aim was to document the incidence of structural CHD detectable in Down syndrome fetuses using recent knowledge and equipment, and to explore the relationship of CHD in trisomy 21 with nuchal translucency (NT), fetal sex and ethnicity.


Our database was searched for the period January 2002 to March 2010 for a karyotype result showing trisomy 21. Only those cases in which an echocardiogram was performed by one of our three specialist fetal cardiologists, and considered technically adequate for diagnosis, were included in the study. All patients with an NT over 3.5 mm were referred for detailed fetal echocardiography at the time of the NT scan—before the karyotype was known—if the cardiologist was present in the department. There is a cardiologist in attendance in office hours on 4 days out of 5 (about 60–70% of the time), so those cases not examined by fetal echocardiography were considered to be randomly distributed. Referral to the cardiologist also took place when a pregnancy with known trisomy 21 was continuing, or when an abnormality likely to be associated with Down syndrome (including CHD) was detected at a routine anomaly scan and referred to our tertiary center. The data were analyzed for reason for referral for fetal echocardiography, maternal age, gestational age at examination, cardiac diagnosis, ethnicity, fetal sex and outcome. We used chi-square test for analysis of differences in the outcome groups.


There were 917 karyotype results showing trisomy 21, of which 487 had had a fetal echocardiogram performed that was considered diagnostic. About 10% of cases studied before 14 weeks' gestation were excluded from our study because the images were not considered diagnostic. No study examined after 14 weeks' gestation was considered non-diagnostic. The reasons for referral for fetal echocardiography are shown in Table 1. An NT > 95th centile was found in 94% of the 417 Down syndrome fetuses in which the NT thickness was known and in 89.5% of those with CHD (Table 2). The cardiac examination took place at less than 14 weeks' gestation in 364/487 (74.7%). Of the 364 fetuses examined at less than 14 weeks, 345 proceeded to termination of pregnancy after karyotyping and no further information was available. Thus, 123 cases were examined in the second trimester, not including 18 continuing pregnancies restudied after the first-trimester scan (Figure 1). Of 36 continuing pregnancies after the second-trimester scan, 33 were also examined early in the third trimester (three having been lost to follow-up).

Figure 1.

Number of fetuses that underwent echocardiography according to gestational age. equation image, Structural congenital heart disease (CHD); equation image, functional CHD; equation image, no abnormalities detected.

Table 1. Indications for referral for echocardiography in 487 fetuses with Down syndrome
Indication for referraln (%)
Increased nuchal translucency368 (75.6)
Suspected congenital heart defect75 (15.4)
Extracardiac abnormality28 (5.7)
Known trisomy 216 (1.2)
Absent nasal bone4 (0.8)
Advanced maternal age3 (0.6)
High risk from quadruple test2 (0.4)
Previous pregnancy with trisomy 211 (0.2)
Total487 (100.0)
Table 2. Nuchal translucency (NT) distribution, incidence of structural heart defects and outcome of continuing pregnancies in the 417 fetuses with Down syndrome with known NT
 Continuing pregnancy (n (%))
NT (mm)nCHD (n (%))IUD with CHDLiveborn with CHD
  • Six live births of 33 continuing pregnancies were excluded because NT was not known for four of them and two babies died in early infancy.

  • *

    95th–99th centile.

  • > 99th centile. CHD, congenital heart disease; IUD, intrauterine death.

≤ 2.42413 (54.2)1 (8)3 (23)
2.5–3.4*5515 (27.3)1 (7)3 (20)
3.5–13.733896 (28.4)8 (8)11 (11)
Total417124 (29.7)10 (8)17 (14)

Structural CHD was detected in 164/487 cases (33.7%), with the diagnostic categories listed in Table 3. An ARSA was seen in 12 cases (isolated in 11), but this was only looked for in the last 3 years of the study. Normal cardiac structure but functional abnormality in the form of tricuspid or mitral regurgitation, or both, or mild ventricular disproportion was found in a total of 179 cases (36.8%) (Table 4). These functional disturbances only involved cases seen in the first trimester of pregnancy. The incidence of functional heart disease was much higher with increased NT (Table 5). Rescan took place in three of these 179 cases and a normal heart structure was confirmed in all, with resolution of the earlier functional findings. All cases of AVSD seen in the first trimester showed atrioventricular valve regurgitation, usually on both sides of the common valve. The mean NT was 5.2 and 5.3 mm in those with a normal heart and in those with a septal defect, respectively.

Table 3. Incidence of cardiovascular abnormalities in 487 fetuses with Down syndrome
Cardiovascular abnormalityn (%)
  1. ARSA, aberrant right subclavian artery; PA, pulmonary atresia.

Atrioventricular septal defect115 (23.6)
 With tetralogy of Fallot10
 With tetralogy of Fallot and PA1
 With tetralogy of Fallot and ARSA1
 With ARSA2
 With small right ventricle3
 With coarctation of aorta4
 With interrupted aortic arch1
Tetralogy of Fallot15 (3.1)
 With PA2
 With PA and right aortic arch1
ARSA12 (2.5)
 With ventricular disproportion1
Ventricular septal defect7 (1.4)
 With great artery disproportion1
Coarctation of aorta6 (1.2)
Tricuspid dysplasia5 (1.0)
 With PA4
Ebstein's anomaly2 (0.4)
Double inlet left ventricle1 (0.2)
Persistent left superior vena cava1 (0.2)
Total164 (33.7)
Table 4. Incidence of functional heart abnormalities in 487 fetuses with Down syndrome
Functional heart findingn (%)
Tricuspid regurgitation173 (35.5)
 With mitral regurgitation5
 With ventricular disproportion9
Isolated mitral regurgitation5 (1.0)
Isolated ventricular disproportion1 (0.2)
Total179 (36.8)
Table 5. Distribution of nuchal translucency (NT) thickness and functional congenital heart disease (CHD) in 179 fetuses with Down syndrome and CHD
NT thickness (mm)Fetuses with functional CHD (n (%))
≤ 2.43 (1.7)
2.5–3.423 (12.8)
3.5–13.7153 (85.5)
Total179 (100.0)

Female fetuses with an AVSD were slightly more frequent than males overall. An AVSD was found in 29.1% of female fetuses vs. 17.8% of male fetuses (Table 6), a statistically significant difference (P = 0.003). There was no statistically significant difference in the rate of AVSD between the ethnic groups Caucasian, Black, Asian, Mediterranean or other (Table 7). The mean maternal age at examination for Africans and Caucasians was 29 and 37 years, respectively.

Table 6. Sex distribution and atrioventricular septal defects (AVSD) in 487 fetuses with Down syndrome
Sexn (%)AVSD (n (%))
Female251 (51.5)73 (29.1)
Male236 (48.5)42 (17.8)
Total487 (100.0)115 (23.6)
Table 7. Ethnic distribution and atrioventricular septal defects (AVSD) in 487 Down syndrome fetuses
Ethnicityn (%)AVSD (n (%))
Caucasian371 (76.2)91 (24.5)
African31 (6.4)8 (25.8)
Asian20 (4.1)4 (20.0)
Mediterranean9 (1.8)2 (22.2)
Other2 (0.4)0 (0.0)
Unknown54 (11.1)6 (11.1)
Total487 (100.0)111 (22.8)

The outcomes of the 487 pregnancies are shown in Table 8. There were 451 terminations and three patients were lost to follow-up after a late second-trimester scan. Of the remaining 33 continuing pregnancies there were 10 intrauterine deaths and 23 live births, two of whom later died in infancy. Three of the 23 newborns had a small VSD diagnosed postnatally that was not detected prenatally. Fifty-eight percent (21/36) of the continuing pregnancies had CHD vs. 32% (143/451) of patients who underwent TOP. Five African patients continued pregnancy, three of them with CHD (60%) as compared to 24 continuing Caucasian patients, 14 of them with CHD (58%).

Table 8. Outcome for 487 fetuses with Down syndrome
Outcomen (%)
  • *

    Two of those liveborn died in infancy.

Termination of pregnancy451 (92.6)
Intrauterine death10 (2.1)
Live birth*23 (4.7)
Lost to follow-up3 (0.6)
Total487 (100.0)


In 1894, Garrod first reported the association of congenital heart defects with Down syndrome17. Since then, many cohort- and population-based studies have been published reporting a prevalence of cardiac defects in Down syndrome children of approximately 40–50%, a striking increase over the rate of approximately 1% of CHD in children with normal chromosomes1–7, 18–23. The classically associated lesion is an AVSD, which is a thousand times more frequent in trisomy 21 than in those with a normal karyotype. However, analysis of the other types of defect found in children shows that about half the cases of CHD involve lesions that would not be detectable in the fetus.

Our study represents a large series of fetal cases of trisomy 21, where the heart was examined in detail by a specialist fetal cardiologist. The rate of structural heart disease is slightly higher than that anticipated from postnatal experience, but our data are biased towards a higher rate of CHD by referral of cases with increased NT (> 99th centile) and suspected CHD or an extracardiac anomaly commonly associated with trisomy 21—an absent nasal bone, for example. If the cases referred for suspected CHD are removed from the analysis, the rate of major CHD falls to 24%, which is more consistent with the postnatal data. This contrasts with previous publications on fetal data, in which a higher rate of CHD is reported24–26. However, the study of Paladini et al.26 included a high proportion of fetuses referred for suspected CHD while that of Hyett et al.24 was a postmortem study that included small VSDs that would not have been detectable on the fetal echocardiogram.

The majority of patients were seen by us because of increased NT thickness, but as this is present in 70–80% of cases of trisomy 21, this probably does not represent a major bias. The type of CHD seen in the fetus was consistent with postnatal experience, with a predominance of AVSD and an absence of heterotaxy syndromes or transposition of the great arteries. We found a low incidence of detection of VSDs prenatally (in contrast to Paladini et al. and Hyett et al.), as many are relatively small and therefore not detectable echocardiographically, especially in the first trimester when the majority of patients were examined. It is likely that if the whole series were re-examined in the mid-trimester, more minor defects, especially VSDs, would be detected. However, in total, 141 cases were examined in the second trimester and 33 of these were also examined in the third trimester, a larger series than has been reported previously. Three of our neonates were found to have small VSDs, missed on antenatal scanning. The expected rate of VSDs in neonates would be around 5%. In addition, more cases of isolated ARSA would be expected had this vessel been looked for in the whole series. An incidence of ARSA of around 30% in trisomy 21 fetuses has been suggested10, 11. ARSA is likely to be under-reported in postnatal echocardiographic series as this is an asymptomatic lesion and quite difficult to identify postnatally, biasing the postnatal incidence of CHD downwards in comparison with fetal series.

It has been suggested that the incidence of AVSD in trisomy 21 varies with the sex of the fetus and in different ethnic groups, being higher in females and in the African–American population27, 28. We confirmed the sex difference in our data, girls being almost twice as likely as boys to have an AVSD (Table 6). The significance of this is unknown. However, there was no statistically significant difference in the rates of AVSD between any of the ethnic groups. Of interest, only 7% of our cases of trisomy 21 diagnosed prenatally were in black women, despite the fact that 20% of patients seen at our institution at the NT scan are from this ethnic group. The mean maternal age difference between Africans (29 years) and Caucasians (37 years) would have affected the prevalence of Down syndrome, but to some extent this is also likely to reflect the difference in acceptance of prenatal testing between the ethnic groups. The study of Freeman et al.27 from the National Down Syndrome Project in the USA, where a higher rate of AVSD with trisomy 21 was found in African–Americans, only looked at the ethnicity of liveborn babies with trisomy 21. The difference found is much more likely to reflect a difference in access to prenatal care, acceptance of prenatal testing and a cultural difference in decision-making after a diagnosis of trisomy 21 in this group. A mother who has access to, and seeks, routine anomaly scanning is more likely to have an AVSD detected prenatally, and to choose termination of pregnancy after diagnosis of trisomy 21. Surprisingly, in our series the presence of CHD did not seem to influence decision-making about termination of pregnancy.

These social and cultural aspects could account for the difference in incidence of AVSD between different ethnic groups. It is certainly our impression from experience of NT scanning that, even when this test presents a high risk, or where an AVSD is detected, a significant proportion of the black population refuse prenatal testing. As a result, trisomy 21 is diagnosed only after birth and therefore this group would not be part of this study.

We could not substantiate an increase in NT thickness in trisomy 21 with septal defects as previously suggested by Hyett et al.29—the mean NT thickness was 5.2 mm in those with normal hearts and 5.3 mm in those with septal defects.

There was a high rate of spontaneous loss between 20 weeks and term in the continuing pregnancies, but the presence or absence of CHD did not appear to increase the chance of intrauterine death. It has been suggested that there is a higher mortality in trisomy 21 cases detected by screening for increased NT than in those with a normal NT30, 31. In our series of 36 continuing pregnancies, 90% of fetuses that died antenatally had increased NT compared with 84% of liveborn babies, a difference that is not statistically significant.

In summary, most fetuses with trisomy 21 have a structurally normal heart on fetal echocardiography, although functional disturbance in the form of tricuspid regurgitation is common in the first trimester, and more minor defects may become evident after birth. The presence of structural CHD was not associated with increased NT thickness. There appears to be an increased incidence of AVSD in females, although an ethnic difference could not be demonstrated. CHD was as frequent in live births as in those that suffered intrauterine death.


The authors would like to thank Dr Jaroslaw Beta for providing statistical support.