The aims of this study were to ascertain the effect of a regional training program in fetal echocardiography for obstetric radiographers on the antenatal detection of major congenital heart disease (CHD) and to document short-term outcomes for major CHD.
All 87 obstetric radiographers in Northern Ireland were invited to attend 2.5 days of training during a 1-year period. Data were collected before and after the training, over a 5-year study period, to assess the effect of training on the antenatal detection of CHD in the population.
The antenatal detection of major CHD rose significantly, from 28% (72/262) pretraining to 43% (36/84) in the year of training (P = 0.008). Antenatal diagnosis of four-chamber-view defects rose significantly (from 38% to 54%; P = 0.04), as did detection of outflow-tract-view defects (from 8% to 21%; P = 0.05). Twelve per cent (13/108) of cases died spontaneously in utero and 8% (9/108) were terminated. Only 78% (67/86) of live-born cases in which CHD had been diagnosed antenatally survived the neonatal period, compared to 93% (221/238) with a postnatal diagnosis of CHD (P < 0.001).
Congenital heart disease (CHD) is the commonest congenital malformation1. It affects eight per 1000 live births and is commoner antenatally2–5. Antenatal diagnosis reduces neonatal mortality for certain CHD conditions6–8. Identification of duct-dependent CHD facilitates early therapy to prevent decompensation. Other benefits include opportunities to counsel parents, to screen for co-existing abnormalities and to plan peripartum management. In severe cases, some patients opt for termination of pregnancy. Rarely does intrauterine intervention improve outcome9.
Around 60% of fetal congenital heart defects are apparent on the four-chamber view10. Over 90% of major heart defects could be diagnosed if outflow-tract views were included. CHD is poorly diagnosed compared with other anomalies11, 12. Around 25% of babies with critical CHD are discharged without a diagnosis and some die before their CHD is detected13, 14.
Low-risk cardiac screening in the UK takes place during the anomaly scan performed by a radiographer. Pregnancies with risk factors for CHD, or suspected abnormality on screening, receive echocardiography. Approximately 90% of affected pregnancies have no historical risk factors, and therefore antenatal diagnosis is extremely dependent on obstetric screening15. Increased nuchal translucency is associated with CHD. In areas with nuchal screening programs, up to 30% of pregnancies affected by major CHD may have identifiable risk factors.
The most recent study of antenatal detection of serious CHD in the UK reported the average antenatal detection rate to be 23%16. Two multicenter European studies found detection rates of 25% and 28%, respectively11, 17. Detection is much better in some areas18. Few studies have examined the effect on the antenatal detection of CHD, of training sonographers to examine outflow-tract views19–22, and the training that was used in most of these studies was very time consuming. A survey of ultrasound screening in England identified that major staffing deficiencies limit attendance at training23.
Northern Ireland has 23 000 live births per year and a population of 1.7 million. There is a single center for fetal and pediatric cardiology and for pediatric postmortems. Termination of pregnancy for fetal abnormality is illegal (although a small proportion of patients travel to other countries for this procedure) and therefore the termination rate is low. These factors make Northern Ireland ideal for a population-based study of CHD.
We hypothesized that delivery of a low-impact regional training program in fetal echocardiography (including outflow-tract views) for obstetric radiographers would significantly improve the antenatal detection rate of major CHD.
A training program in fetal echocardiography involving 2.5 days of formal training over a period of 1 year was devised (Figure 1). All 87 obstetric radiographers in Northern Ireland were invited to participate in training. Each participant attended one foundation day in the tertiary center, where lectures and small-group ‘hands on’ training were provided. A total of seven foundation days took place. Each participant also attended two half-day sessions in the Regional Fetal Cardiology Clinic, during which fetal echocardiograms were observed and counseling of parents was witnessed. The training team visited each participating hospital on two occasions for a half-day training session. During the visits, radiographers received ‘hands on’ training using their departments' ultrasound equipment. A refresher afternoon took place 6 months after the start of the training program. At this session, a selection of cases of fetal CHD detected by radiographers during the training period was reviewed and a lecture on fetal echocardiography was given. All components of the course included teaching on the four-chamber and outflow-tract views.
Study design, and inclusion and exclusion criteria
A before-and-after study design was used. Training was delivered from September 2006 until September 2007. ‘Before’ data were gathered for the 4-year period before the start of training (pretraining period). Data for the ‘after’ group were collected during the year of training. Ethical approval was obtained from the Northern Ireland Research Ethics Committee.
Anomaly scans are performed at 22 weeks of gestation, so women scanned at the start of the training period had an estimated delivery date in early January 2007. Therefore, babies with dates of birth between 1 January 2007 and 31 December 2007, and fetuses who died in utero with estimated dates of delivery between 1 January 2007 and 31 December 2007, were included in the ‘after’ group. The ‘before group’ included children born between 1 January 2003 and 31 December 2006, or who died in utero with an estimated date of delivery between 1 January 2003 and 31 December 2006.
Major CHD was defined as CHD requiring surgery or intervention, or resulting in death, within the first year of life. The same definition was used in the last major study of antenatal detection of CHD in the UK. Fetuses that died in utero after 22 completed weeks of gestation, with an antenatal diagnosis of major CHD, were included. Conditions that could not reasonably be expected to receive an antenatal diagnosis were excluded (patent arterial duct, isolated secundum atrial septal defect and coronary artery abnormalities). Conditions that usually develop late in pregnancy and therefore may not be apparent at the time of anomaly scanning were also excluded, such as cardiac tumors and cardiomyopathies, unless diagnosed antenatally. Neonatal death was defined as the death of a live-born infant within 28 days of birth.
The Regional Fetal and Paediatric Cardiology database was used to identify cases. Pediatric postmortem records were used to identify live-born cases diagnosed with major CHD after death. The following outcome data were documented: live birth, intrauterine death, termination of pregnancy and neonatal death or survival. The presence or absence of karyotype abnormalities was recorded for each case where possible.
Cases were grouped according to whether they could have been diagnosed based on findings in the four-chamber-view alone (‘four-chamber-view defects’), or whether images of the outflow tracts would have been required for detection (‘outflow-tract-view defects’). The researchers A.M. and A.S. independently allocated patients into these groups. Where there was disagreement, cases were discussed and agreement reached regarding the most appropriate classification (Table 1). Cases were also classified according to the presence or absence of a functionally univentricular circulation (based on cardiac anatomy and treatment provided).
Table 1. Classification of congenital heart disease according to the view required to detect it antenatally
Four-chamber view alone
Mitral atresia (including hypoplastic left heart syndrome)
Pulmonary atresia intact ventricular septum
Double-inlet left ventricle
Atrioventricular septal defect
Coarctation of aorta
Large isolated ventricular septal defect
Anomalous pulmonary venous connection
Outflow-tract views required
Transposition of the great arteries
Pulmonary atresia with ventricular septal defect
Tetralogy of Fallot
Double-outlet right ventricle
Aortic stenosis (unless associated with endocardial fibroelastosis or hypoplastic left heart syndrome)
Pulmonary stenosis (unless associated with hypoplastic right ventricle)
Data were analyzed using SPSS version 14.0 (SPSS Inc., Chicago, IL, USA). Categorical data were summarized as percentages and compared using the chi-square test for 2 × 2 contingency tables (or Fisher's exact probability test where indicated) to determine statistical significance. The chi-square test for trend was used for ordered categories.
In total, 90% (78/87) of obstetric radiographers in Northern Ireland participated in training. There were 346 cases of major CHD identified during the 5-year study period of which 108 (31%) were diagnosed antenatally. There were 324 live-born cases among 114 017 live births during this 5-year study period, giving an incidence of 2.8 cases of major CHD per 1000 live births. Of the live-born cases, 72% (234/324) were diagnosed postnatally, 27% (86/324) were diagnosed antenatally and 1% (4/324) were diagnosed at postmortem.
Comparison of antenatal detection rates before and during training
There was a significant increase in the antenatal diagnosis of major CHD, from 28% (72/262) in the pretraining period to 43% (36/84) in the training period (P = 0.008) (Figure 2). The chi-square test for trend indicated that there was no temporal trend towards increased antenatal detection by year during the pretraining period (χ2 = 0.39, P = 0.53).
Referral indications for antenatally diagnosed cases
There were 150 referral indications for the 108 cases of major CHD diagnosed antenatally (Figure 3). Of these cases, 86% (93/108) were thought to have cardiac abnormalities (i.e. an abnormal heart) on obstetric scanning at the time of referral. The second commonest referral indication was the presence of a fetal extracardiac abnormality (40/108). Other high-risk factors were poorly represented among those with an antenatal diagnosis of CHD.
In the entire cohort of antenatally diagnosed cases, there was a clear predominance of four-chamber-view defects (Figure 4), with only 12% (13/108) of antenatally diagnosed cases having outflow-tract-view defects. By contrast, 45% (107/238) of postnatally diagnosed cases had outflow-tract-view defects. There was a statistically significant improvement in the antenatal diagnosis of four-chamber-view defects between the pretraining (38%, 65/170) and training (54%, 30/56) periods (P = 0.04). There was also a significant improvement in the antenatal detection of outflow-tract-view defects between the pretraining (8%, 7/92) and training (21%, 6/28) periods (P = 0.05).
Almost half of those with an antenatal diagnosis (46%, 50/108) of CHD had functionally univentricular physiology (Table 2). This accounts for 71% (50/70) of total cases with functionally univentricular physiology. There was no significant difference in the proportion with known karyotype abnormalities between the antenatally diagnosed group (25%, 27/108) and the postnatally diagnosed group (23%, 54/238) (P = 0.64).
Table 2. Antenatal detection rates of congenital heart disease (CHD) during the pretraining and training periods, with cases subcategorized according to defect type
Antenatal detection rate (% (n))
Type of CHD
Defects associated with univentricular physiology
Mitral atresia (including hypoplastic left heart syndrome)
Conditions with hypoplastic right ventricle
Double-inlet left ventricle
Unbalanced atrioventricular septal defect
Other complex single ventricle
Other defects usually associated with an abnormal four-chamber view
Balanced atrioventricular septal defect
Ventricular septal defect
Anomalous pulmonary venous connection
Defects usually associated with abnormal outflow-tract views
The overall termination rate for antenatally diagnosed major CHD was 8% (9/108). Of the cases with an antenatal diagnosis, 12% (13/108) died spontaneously in utero. Overall survival until the end of the neonatal period for antenatally diagnosed cases was 62% (67/108). Survival for live-born cases with an antenatal diagnosis was 78% (67/86), which is significantly lower than the survival for live-born cases with a postnatal diagnosis (93%, 221/238) (P < 0.001). There was no significant difference in the rate of intrauterine death, termination of pregnancy or neonatal survival between the pretraining and training groups.
As a result of training, the antenatal diagnosis rate for major CHD in Northern Ireland is now well in excess of the published averages for the UK and Europe11, 16, 17. This results from significant improvements in screening of both the four-chamber view and the outflow-tract views. The most marked improvement was in the detection of outflow-tract abnormalities. However, the detection rate remains lower than those reported by the highest achieving areas in Europe17, 18, 21, 22. Further training will be required in order to attain these standards.
Other studies have also shown an improvement in the antenatal detection of CHD following implementation of a training program for obstetric sonographers focusing on the assessment of outflow tracts. Carvalho et al. performed a single-center study and found that following training the antenatal detection rate for major CHD was 75%19. Training relied on close links between the fetal cardiology and obstetric screening departments. Of the pregnancies screened, 9.6% received a fetal echocardiogram, which was attended by the referring sonographer where possible. Whilst the results are impressive, this would be difficult to implement on a regional basis.
Recently published Italian data showed that training, including assessment of the outflow-tract view, resulted in an antenatal detection rate of CHD of 66%21. The training required sonographers to attend a 2-week course at the referral center and 1 day of further training every 4 months. It is unlikely that this intensive training model could be implemented by many other health services. Pezard et al. also showed that the antenatal diagnosis of CHD can be improved by weekly attendance at regional multidisciplinary prenatal diagnosis meetings22.
By contrast, our training program required radiographers to attend only 2.5 days of formal training during a 1-year period. This training model may therefore be more easily implemented in countries with limited healthcare resources available for training. This is corroborated by the study of Hunter et al., in which a similar training program was delivered and the antenatal detection of CHD rose from 17% to 36%20.
Another technique to improve the antenatal detection of major CHD is the use of internet resources to deliver fetal echocardiography teaching. Online training has been produced by several bodies, including the International Society of Ultrasound in Obstetrics and Gynecology and The Fetal Medicine Foundation. Online resources have the advantage of convenience and relatively low cost; however, they lack the important interactive aspect of face-to-face training.
During our training program, national guidelines for obstetric screening suggested that, as a minimum standard, the four-chamber view of the heart should be assessed in all pregnancies. Since the completion of our research, these guidelines have been updated and now state that the outflow tracts should also be routinely assessed during the anomaly scan. New guidelines will be published this year, which also recommend inclusion of the three-vessel view in routine screening. This move demonstrates the increasing recognition of the importance of antenatal identification of outflow-tract defects. However, widespread training of obstetric radiographers will be necessary to successfully implement this change.
The current study appears to have good case ascertainment. Bull et al. used the same definition of major CHD and reported a live birth rate of 1.7 per 100016. In that study, complex balanced CHD and postmortem cases were not included, which may explain their slightly lower incidence.
Despite a relatively low pregnancy-termination rate, our study concurs with others regarding the relatively poor survival rates for antenatally diagnosed cases with major CHD24–27. This is probably because the antenatally diagnosed population often have more complex CHD (as demonstrated by the high proportion with functionally univentricular physiology) and a higher frequency of extracardiac abnormalities. There was a higher rate of spontaneous intrauterine death in this cohort compared with other studies; this may be related to higher pregnancy-termination rates in other series, which may have otherwise resulted in spontaneous intrauterine death17, 25, 28.
The risk factors for CHD present in the antenatal population diagnosed with CHD reinforce the importance of low-risk screening in the antenatal detection of major CHD. The vast majority of cases were suspected to have CHD on obstetric screening. The low rate of other risk factors suggests that, without an obstetric suspicion of cardiac abnormality, the majority of cases would have been missed.
A weakness of this study was the inability to include cases of CHD that were not diagnosed antenatally and died in utero. This is a weakness intrinsic to the study of antenatal detection of fetal CHD. Even those studies that include fetal cases of CHD diagnosed at postmortem do not accurately reflect the total number of undiagnosed fetuses with CHD because they exclude cases which did not receive a postmortem.
An antenatal diagnosis of certain outflow-tract defects has been demonstrated to improve neonatal mortality, and therefore detection of these defects is particularly important6. Despite significant improvement in the detection of outflow-tract abnormalities during this study, there remains a high proportion without an antenatal diagnosis. Undoubtedly the limited training time available was partly responsible for this. The concept of outflow-tract scanning was relatively new to the radiographers and therefore further training will be required in order to strengthen the rates of detection of these key defects.
This study targeted obstetric radiographers for training. This group perform the vast majority of anomaly scans in Northern Ireland; however, this is not the case in other countries. Ideally, other groups of obstetric sonographers should be included in fetal echocardiography training in order to maximize the antenatal diagnosis of major CHD.
Obstetric screening identifies the majority of cases of major CHD which receive an antenatal diagnosis. With a relatively simple training programme, significant improvements can be made in the antenatal diagnosis rate for major CHD. Obstetric sonographers are capable of successfully assessing the outflow tracts when given adequate training. However, training must be an ongoing process in order to maintain and improve on current standards, and this will require the investment of resources.
This work was funded by Heartbeat (the parent support group for families affected by congenital heart disease). The statistics were overseen by Dr Christopher Patterson. Many thanks to Anne Honkannen Graham, who contributed to data collection. We acknowledge Fiona Alderdice's contribution to the original study design.