To investigate the outcome of fetuses affected by congenital heart defects (CHD), either detected or undetected at ultrasound screening, according to their complexity and severity.
To investigate the outcome of fetuses affected by congenital heart defects (CHD), either detected or undetected at ultrasound screening, according to their complexity and severity.
The study group comprised 3633 malformed fetuses entered into the Eurofetus database of which 798 had CHD. We compared the short-term outcome in cases where a CHD was detected by ultrasound screening with that in cases where a CHD was not detected. Isolated and associated CHD (ICHD and ACHD) and the degree of severity of defects were considered separately. Outcome data included termination of pregnancy (TOP), intrauterine fetal death, neonatal death (< 6 days after birth), gestational age at diagnosis and at delivery, mode of delivery and birth weight.
Of the 798 fetuses with CHD, 595 had ICHD and 203 had ACHD. The diagnosis of an anomaly was made significantly earlier in ACHD cases. TOP was chosen in 28% of cases with a prenatal diagnosis of CHD, 20% for ICHD and 37% for ACHD (P < 0.001). The survival rate of antenatally diagnosed fetuses was lower in those with ACHD than in those with ICHD (P < 0.001) and lower for fetuses with antenatal diagnosis than with postnatal diagnosis (P < 0.001); this was due to significant differences in the complexity and severity of the defect. Premature delivery (< 32 weeks) was more frequent in fetuses in which an antenatal diagnosis of CHD had been made. Severe CHD were diagnosed earlier and were associated with a higher rate of TOP and spontaneous loss.
The severity of CHD has a positive influence on the sensitivity of screening but a negative influence on the outcome. CHD that are not diagnosed antenatally include a high proportion of benign malformations. This explains the apparent paradox of a poorer outcome in fetuses in which a CHD was detected prenatally compared to those fetuses in which the defect was undetected prenatally. However, prenatal diagnosis provides parents with the option of TOP or of preventive care in specialized centers. Copyright © 2003 ISUOG. Published by John Wiley & Sons, Ltd.
Of the malformations affecting fetuses and newborn babies, malformation registries1 report that congenital heart defects (CHD) are among the most frequent (25.9%) and the most life-threatening. The poor rate of prenatal diagnosis of CHD compared to other important malformations was the motivation for our study on the potential benefit of antenatal screening for CHD2. The diagnosis and prognosis of CHD compared to other fetal malformations is unclear because CHD constitute a very heterogeneous group of anatomical lesions, differing in complexity, outcome and in their amenability to prenatal diagnosis. It is generally assumed that ease of diagnosis and long-term prognosis reflect the complexity of the anatomical lesions. This particular assumption should be clarified. Only a few reports have compared the outcome of pregnancies with fetal CHD diagnosed during and after pregnancy3–5, and few reports consider isolated and associated CHD (ICHD and ACHD) separately6. This approach prevents bias introduced by the inclusion of associated non-cardiac anomalies, which are particularly prone to having an effect on sensitivity, outcome and prognosis.
Using the data provided by Eurofetus, a large collaborative study on ultrasound screening for fetal malformations performed in unselected pregnant women2, 7–11, we investigated the different outcomes of cases of CHD. We compared the outcome of (1) antenatally diagnosed cases versus non-diagnosed cases; (2) ICHD versus CHD associated with extracardiac malformation and/or chromosome anomalies (i.e. ACHD); and (3) for ICHD, three classes of severity. By analyzing the subgroups in this way we hoped to determine how variations in the incidence of these different types of cardiac abnormality within a cohort would influence the sensitivity of diagnosis and the long-term outcome.
The data for the present study were extracted from the Eurofetus database. Eurofetus is a collaborative, hospital-based study on the effectiveness of ultrasound screening for fetal malformations. The study was conducted prospectively on a routine basis on unselected pregnant women. Sixty-one obstetric ultrasound centers from 14 European countries participated in the study. The screening program methodology and initial results have been detailed in earlier reports2, 7–11. Analyzable data were available for 3633 malformed fetuses screened from January 1990 to July 1993. Of these fetuses, 798 (22%) were affected by CHD12.
The fetuses were separated into two groups: the ‘antenatal diagnosis’ group, which included 136 fetuses with ICHD and 135 fetuses with ACHD that had been diagnosed during pregnancy screening (the true-positive cases); and the ‘postnatal diagnosis’ group, which included 459 fetuses with ICHD and 68 fetuses with ACHD in which their malformations were recognized only after birth or following death in utero (the false-negative cases).
The possible outcomes included termination of pregnancy (TOP) (which was an offered option in all of the participating centers), fetal death, postnatal death before 6 days and survival at day 6. In addition, we considered the gestational age at prenatal detection and at delivery, the mode of delivery including spontaneous labor, induced labor and Cesarean section, and the birth weight.
Antenatal ultrasound diagnosis was confirmed by postnatal or post-termination diagnosis using clinical, surgical, imaging or necropsy findings. In addition, for fetuses with CHD, findings on echocardiography performed by pediatric cardiologists were used to complete the postnatal ascertainment. The CHD were classified according to the International Classification of Diseases (ICD 9th edition)12. The fetuses with ICHD were additionally allocated into four diagnostic classes based on clinical severity. Class A comprised severe malformations only partially correctable surgically. Class B comprised severe malformations totally correctable surgically. Class C comprised moderate and mild malformations. Class D comprised unspecified malformations12 and was not included in the severity-outcome comparison. Fetuses with multiple cardiac anomalies were classified according to the most serious lesion. The prenatal detection rate of CHD according to clinical severity has been analyzed elsewhere12. The t-test was used for continuous variables and the χ2 test for categorical variables. Odds ratio (OR) and 95% confidence interval (CI) were compared to estimate the risk of preterm delivery, Cesarean section and labor induction between the antenatal and postnatal diagnosis groups. A value of P < 0.05 was considered to be statistically significant. All tests were two-tailed. Analyses were performed using SPSS statistical software version 10.0 (SPSS Inc., Chicago, IL, USA).
Of the 798 fetuses with CHD, 595 (74.6%) were ICHD, 494 (83%) with a single lesion and 101 (17%) with multiple CHD; 203 (25.4%) were ACHD of which 40.4% had a chromosomal abnormality. The overall antenatal detection rate of CHD was 34%, 23% for ICHD and 66% for ACHD (P < 0.001 by χ2 test). The diagnosis of anomalies was made significantly earlier in fetuses with ACHD (24.3 weeks) than in those with ICHD (28.3 weeks) (P < 0.001 by t-test). There was a highly significant difference in the number of TOP in the antenatal diagnosis group between those with ICHD and those with ACHD (20% vs. 37%, P < 0.001 by χ2 test) (Table 1). The listing and grading of ICHD cases into four classes of severity (from A to D) is given in Table 2.
|Malformations||Total n||Antenatal detectionn (%)||Mean GA at detection Weeks (range)||TOP n (%)|
|Single CHD||494||106 (21.5)||28.4 (11–42)||19 (17.9)|
|Multiple CHD||101||30 (29.7)||27.9 (17–41)||8 (26.7)|
|Total ICHD||595||136 (22.9)||28.3 (11–42)||27 (19.9)|
|Without chromosomal anomalies||121||85 (70.2)||23.9 (12–40)||34 (40.0)|
|With chromosomal anomalies||82||50 (61.0)||24.9 (10–37)||16 (32.0)|
|Total ACHD||203||135 (66.5)||24.3 (10–42)||50 (37.0)|
|Tricuspid atresia and stenosis||7461||4||4||1||0||9|
|Hypoplastic left heart syndrome||7467||20||17||5||1||43|
|Endocardial cushion defect||7456||11||10||13||11||45|
|Stenosis of aortic valve||7463||2||11||0||2||15|
|Transposition of great vessels||7451||10||44||4||0||58|
|Tetralogy of Fallot||7452||15||21||10||3||49|
|Other bulbus cordis anomalies||7458||2||0||0||0||2|
|Anomalies of pulmonary valve||7460||6||18||1||1||26|
|Ventricular septal defect||7454||6||227||46||27||306|
|Atrial septal defect||7455||5||38||12||12||67|
|Unspecified defect of septal closure||7459||0||3||0||0||3|
|Other anomalies of aorta||7472||0||4||1||0||5|
|Anomalies of pulmonary artery||7473||3||4||2||2||11|
|Anomalies of great veins||7474||0||4||0||0||4|
|Other specified anomalies of circulatory system||7478||1||0||0||0||1|
|Other specified minor anomalies||7468||21||5||12||2||40|
|Unspecified anomalies of heart||7469||8||10||13||4||35|
Of the pregnancies in which a malformation had been detected on antenatal ultrasound but TOP was not performed, spontaneous prenatal loss occurred in 46% and was significantly more likely in fetuses with ACHD (68.2%) than in those with ICHD (29.4%) (P < 0.001 by χ2 test). Consequently, at the time of hospital discharge, 53.6% (104/194) of the antenatal diagnosis group were alive (70.6 % for ICHD and 31.8% for ACHD). In the postnatal diagnosis group, 456/459 ICHD and 68/68 ACHD cases were available for outcome study. Death occurred in 9.9% (8.6% for ICHD and 19.1% for ACHD, P < 0.01 by χ2 test between ICHD and ACHD) of cases. The survival rate at 6 days after birth was significantly lower in the antenatal diagnosis group than in the postnatal diagnosis group for both ICHD fetuses (70.6% vs. 91%, P < 0.001) and ACHD fetuses (31.8% vs. 80.9%, P < 0.001) (Table 3).
|Outcome||AD n (%)||PD n (%)|
|Intrauterine fetal death||13 (11.9)||2 (0.4)|
|Neonatal death (< 6 days)||19 (17.5)||37 (8.2)|
|Alive at 6 days||77 (70.6)||417 (91.4)|
|Total ICHD||109 (100)||456 (100)|
|Intrauterine fetal death||30 (35.3)||0 (0)|
|Neonatal death (< 6 days)||28 (32.9)||13 (19.1)|
|Alive at 6 days||27 (31.8)||55 (80.9)|
|Total ACHD||85 (100)||68 (100)|
For the ICHD, so far as the severity of the lesion was concerned, those with a less severe defect (Class C) had a much better outcome in both the antenatally and postnatally diagnosed patients: the survival rates at 6 days were 81.3% and 97.2% for Class C, 69.0% and 92% for Class B, and only 12.1% and 48.6% for Class A, respectively. Gestational age at antenatal detection decreased from Class C to Class A: 32.4 weeks (standard deviation, SD = 3.8), 28.9 weeks (SD = 7.2) and 25.8 weeks (SD = 6.2), respectively. The rate of TOP increased with the severity of the lesion (Table 4).
|Severity (Class†)||Total (n)||TOP (%)||IUD (%)||Live birth (%)||Death < 6 days (%)||Total loss (%)||Alive at 6 days (%)|
The mean gestational age at delivery and the mean birth weight in liveborn infants are shown in Table 5. In general, gestational age and weight were lower in the antenatal diagnosis group than in the postnatal diagnosis group (36.8 vs. 38.3 weeks, 2721 vs. 3087 g, P < 0.001), although the differences were not statistically significant for ACHD. Of the cases with ICHD, prematurity (gestational age < 32 weeks) was more frequent in the antenatal diagnosis group compared to the postnatal diagnosis group (12% vs. 3%, OR = 3.9, 95% CI 1.7–9.1), whereas 33% of deliveries occurred before 37 weeks in the antenatal diagnosis group vs. 19% in the postnatal group (OR = 2.1, 95% CI 1.3–3.4).
|AD (n = 151)||PD (n = 522)||P*|
|n||Mean (SD)||n||Mean (SD)|
|GA at delivery (weeks)†|
|ICHD||94||37.2 (3.3)||434||38.4 (2.7)||< 0.001|
|ACHD||53||36.0 (3.5)||64||37.1 (3.1)||0.072|
|Total CHD||147||36.8 (3.4)||498||38.3 (2.8)||< 0.001|
|Birth weight (g)†|
|ICHD||91||2928 (729)||431||3158 (766)||0.009|
|ACHD||50||2344 (730)||66||2621 (831)||0.065|
|Total CHD||141||2721 (780)||497||3087 (796)||< 0.001|
A comparison of ACHD and ICHD showed that delivery occurred significantly earlier in fetuses with ACHD than in those with ICHD in both the antenatal (P < 0.05 by t-test) and postnatal diagnosis groups (P < 0.001 by t-test). Average birth weight was also significantly lower in ACHD than in ICHD in both the antenatal (P < 0.001 by t-test) and postnatal diagnosis groups (P < 0.001 by t-test).
As for the mode of delivery, the group undergoing spontaneous labor served as a reference for comparison between the Cesarean section and induction of labor groups. Table 6 shows that labor was significantly more frequently induced in the antenatal than in the postnatal diagnosis group for both ICHD and ACHD fetuses, the risks (OR) were 3.7 (95% CI 1.9–6.7) and 3.8 (95% CI 1.2–12.4), respectively. However, the Cesarean section rate was not significantly different for both the ICHD and ACHD fetuses.
|Mode of delivery*||AD (n = 142) n (%)||PD (n = 468) n (%)||OR (95% CI)|
|Cesarean section||18 (20.0)||71 (17.5)||1.5 (0.8–2.7)|
|Induced labor||21 (23.3)||34 (8.4)||3.7 (1.9–6.7)|
|Spontaneous labor||51 (56.7)||301 (74.1)||1|
|Cesarean section||18 (34.6)||17 (27.4)||1.8 (0.7–4.2)|
|Induced labor||11 (21.2)||5 (8.1)||3.8 (1.2–12.4)|
|Spontaneous labor||23 (44.2)||40 (64.5)||1|
This study investigated the outcome of fetuses affected by CHD up to 6 days after birth and was based on data taken from a large, routine, ultrasound screening program on unselected pregnant women. The study's originality lies in the fact that it examines the influence of the timing of the diagnosis and the type of defect on the outcome. The influence of many factors on the sensitivity13, and the relationship between complexity, severity of the malformation and the sensitivity, has been examined in previous studies12. In this study we also examined the possible influence of complexity and severity of CHD on the outcome of fetuses, including rate of spontaneous fetal loss, frequency of TOP, gestational age at diagnosis, gestational age at birth and mode of delivery.
It is important to consider the frequency of malformations in a given sample with respect to screening sensitivity and outcome. To be representative of the general population, a sample should include the studied anomaly at a frequency close to its frequency in the general population. Indeed, significant difference in frequencies might alter the results of the screening of unselected pregnancies by modifying the composition of the sample. For instance, a too high frequency might be generated by the inclusion of a significant proportion of high-risk pregnancies, and a too low frequency by the exclusion of too many anomalies, especially benign ones. In this study, the prevalence of fetuses with CHD was 22% among malformed fetuses, which is close to the figure of 25.9%1 representative of the European population. Furthermore, the balance between ICHD and ACHD should also be considered when interpreting the results. Since ICHD are less frequently diagnosed than ACHD, are diagnosed later in gestation and have a lower rate of TOP, and since ACHD not diagnosed antenatally have a worse outcome than diagnosed cases of ICHD, any excess of one of these two groups of malformations could strongly alter the results of ultrasound screening. In the present study the ratio of ICHD to ACHD (a figure rarely found in the literature) is 2.9, which is similar to the value of 2.2 reported in a recent paper6. It is to be stressed that the two studies quoted for comparison1, 6 were population-based. The sensitivity of antenatal screening for ICHD was significantly lower than for ACHD (Table 1). The relatively low impact of a large number of ICHD fetuses, i.e. low in terms of severity, was recently emphasized14. Indeed, in our study, of those fetuses with ICHD, 12% were affected by the most severe anomalies (Class A) and 54% by less severe anomalies (Class C), with a respective sensitivity of 49% and 5% (Table 4). Class C includes a large number of ventricular septum defects, 46% in our sample12, which are usually benign. So the fetuses with ACHD and fetuses with ICHD that belong to Class A had a much higher prenatal detection rate.
The expected quality of life of the child is, for parents and caregivers, the main factor in the decision to terminate a pregnancy. The TOP rate is usually associated with the incidence of extracardiac and chromosomal anomalies15, 16. Hence the complexity of the malformations contributes between 52%5, 15 and 91%17 to the TOP decision. Indeed, in this study TOP was chosen significantly more frequently in fetuses affected by ACHD (37%) than by ICHD (20%). Moreover, a much higher TOP rate (52%) was found for the fetuses with ICHD falling within Class A (Table 4).
Gestational age at diagnosis of malformation has a significant impact on the decision to interrupt pregnancy, on the one hand because many European countries have a legal upper limit for termination, and on the other hand because parents' decisions may alter with advancing gestational age. Fortunately, gestational age at diagnosis was associated with the severity and complexity of CHD in this study: ACHD were diagnosed nearly 4 weeks earlier than ICHD and gestational age at diagnosis increased from Class A to Classes B and C. The TOP rate has been shown to reach 70% in those with an early (< 19 weeks) diagnosis, and to decrease to 61% when it was made before 23 weeks18. The finding in our study was close to the latter figure with a TOP rate of 57% when the diagnosis was made before 24 weeks. Rules concerning legal termination on account of malformations differ in the 14 countries that took part in this study and this fact might have had a bearing on the mean gestational age at TOP. Currently in the UK a pregnancy may be terminated up to 28 weeks of gestation for a severe abnormality5. Two studies from the UK demonstrated that the TOP rate among the cases detected before 28 weeks' gestation were 69%4 and 55%5 in selected populations at risk for CHD. No upper gestational age limit for severe anomalies exists in France, Belgium or The Netherlands. In France, routine screening in a non-selected population resulted in a TOP rate of 47% in fetuses with CHD19. The absence of a gestational age-linked limitation may well reduce the number of TOP by allowing a more precise diagnosis that is not rushed on account of the time limit.
Certainly TOP can be considered as an important option offered by antenatal diagnosis, although the proportion of parents choosing this option might decrease with the progress of surgical and medical treatment. Most authors agree that antenatal detection is generally linked to poor outcome5, 15, 16, 18, 20. Our study confirms this view, and after excluding TOP, perinatal mortality remained significantly higher in the antenatal (46%) than in the postnatal diagnosis groups (10%) both in the ICHD (29% vs. 9%, respectively) and ACHD (68% vs. 19%, respectively) groups (Table 3). While ultrasound diagnosis could induce the TOP option, it is not responsible for the poor outcome, and only the severity of the anomalies has a major effect on poor outcome. The influence of the severity of the cardiac lesions and the proportion of severe lesions in the studied sample on the mortality rate in the antenatal diagnosis group was also investigated. The antenatal diagnosis group included 85% of infants with severe malformations (Classes A and B), a significantly higher proportion than the 36% observed in the postnatal diagnosis group (P < 0.001) (Table 4). It might be surprising that even the less severe defects (Class C) had a worse outcome in antenatally than postnatally diagnosed patients, however the classification system for anomalies might be too crude to make observations with absolute certainty. Indeed, as an example, all ventricular septum defects were categorized as Class C because the large majority of them are benign. However, some of these defects are severe and carry a poor prognosis and, again, screening will pick up the worst cases.
The better outcome obtained in false-negative cases is explained by a significant shift towards more benign malformations among this group. The associations between high mortality, antenatal diagnosis and serious malformation are linked by the significantly higher sensitivity obtained in the complex malformations (66%) and the severe malformations were predominantly included in the antenatal diagnosis group.
Our study only included cases with follow-up up to 6 days of age, which may have resulted in an artificially higher survival rate (54%) of antenatal diagnosis compared to that (17%) after a 1-year follow-up5.
Ultrasound screening lowers the prevalence of CHD at birth and particularly that of certain types of CHD seen at term18, 21. In this study, 54% (20/37) of the cases of hypoplastic left heart were detected antenatally, 50% (10/20) of them ended in TOP, and in the continuing pregnancies 90% (9/10) of newborns died before attaining 6 days of age. Conversely, for the patients in whom the diagnosis was not made antenatally, 76% (13/17) of the newborns died before attaining 6 days of age. Clearly, a higher antenatal detection of severe lesions should reduce the birth prevalence of CHD (because of TOP) unless acceptable management can be offered in the future. On the contrary, antenatal diagnosis seems to allow a higher survival rate for serious but treatable CHD, as already published for some particular anomalies22, 23. For example, the fetuses affected by a transposition of the great vessels definitely benefit from antenatal diagnosis by the almost total absence of pre-and postoperative mortality23, and future improvements in sensitivity should reinforce the impact of screening on survival. Even for severe cardiac malformations requiring surgical treatment shortly after birth, morbidity will decrease and the outcome will improve if treatment is started before the infant develops severe symptoms24, 25. Similar observations have been reported for coarctation of the aorta26. This topic has been debated recently27, and further studies on the subject are warranted to truly define the benefits offered by antenatal diagnosis.
Preterm delivery and low birth weight were significantly more frequent in the antenatal diagnosis group and this might reflect the higher frequency of labor induction in this group (Tables 5 and 6). Furthermore, independent of its effect on antenatal diagnosis rate, complexity of lesions may favor prematurity since fetuses with ACHD were born earlier than those with ICHD in both the antenatal and postnatal diagnosis groups.
Cesarean section is generally undertaken for obstetric reasons, the fetal cardiac status being rarely an indication. Our study shows that the Cesarean section rate was not significantly different between the antenatal and postnatal diagnosed groups.
In conclusion, antenatal diagnosis of CHD has, up until now, been associated with a lower survival rate at 6 days, and a lower gestational age at birth. This is principally due to the fact that anomalies are more severe in fetuses in which the defects have been diagnosed prenatally. Mothers with fetuses with a prenatally diagnosed CHD should be offered amniocentesis or a more detailed ultrasound examination. TOP is a reasonable option for severe anomalies detected early in pregnancy. However, a too-strict dependence on gestational age-limited TOP might have a negative influence on the quality of the decision. Nevertheless, prenatal diagnosis and subsequent immediate maternal transport to highly specialized care units may lead to some women with severe fetal CHD being advised to continue the pregnancy. Conversely, improving sensitivity, especially for ICHD, could also lead to a better survival rate and to reduced morbidity.
The authors are grateful to the European Union for a grant to the Eurofetus Study (Contract MR4*-0225-B), to Profs Esther Vamos and Elisabeth Wollast for their advice and comments on the original manuscript, to Dr Bernardus Tubbing for revising the manuscript, to Mrs Van Noten for her efficient assistance in collecting documentation, and to the National Foundation for Research in Pediatric Cardiology, Halle, Belgium for supporting the scientific work of the Clinique de Cardiologie Pédiatrique et Congénitale, Hôpital Universitaire des Enfants Reine Fabiola.