To compare the immediate postinterventional and long-term outcomes of children with a prenatal and those with a postnatal diagnosis of isolated congenital heart defects.
To compare the immediate postinterventional and long-term outcomes of children with a prenatal and those with a postnatal diagnosis of isolated congenital heart defects.
This was a retrospective study of 257 children admitted over a 10-year period to our pediatric cardiology unit with one of four different cardiac lesions: transposition of the great arteries, atrioventricular canal defect, tetralogy of Fallot and pulmonary atresia; 208 were diagnosed postnatally and 49 prenatally. Management was identical in the two patient groups.
The median age at admission was 22 days in the postnatal group and 10 days in the prenatal group. In the prenatal group there was a higher median preoperative O2 saturation level (P = 0.07), fewer cases of preoperative cardiac failure (P = 0.03), fewer cases of preoperative closure of the duct (P = 0.04), a shorter median duration of postoperative mechanical ventilation (P = 0.03), less need for resurgery (P = 0.02) and a shorter median duration of stay in the intensive care unit (P = 0.05). Postoperative survival was 96% in the prenatal group and 90% in the postnatal group. Assessment of long-term survival revealed a longer catheter intervention-free interval in the prenatal group (P = 0.03). At the 1-year follow-up, residual impaired cardiac function was less frequent in the prenatal than in the postnatal group (P = 0.04). Overall survival at maximum follow-up was 92% in the prenatal and 84% in the postnatal group.
Prenatal diagnosis of isolated congenital heart defects allows admission for surgery in a more stable condition and is associated with lower short-term and long-term morbidity and mortality. Copyright © 2007 ISUOG. Published by John Wiley & Sons, Ltd.
Fetal echocardiography has become widespread in recent years1, 2. In expert hands, the majority of cyanotic and non-cyanotic congenital heart defects can now be diagnosed prenatally with high accuracy3, 4. The knowledge of a cardiac anomaly before birth allows optimal early postnatal and neonatal management, such as delivery in a high-risk tertiary center, immediate administration of prostaglandins in duct-dependent congenital heart defects, or minimal invasive intervention, if necessary. It has been hypothesized that such risk-adapted management should improve the outcome of affected children by reducing hypoxemic complications and stabilizing the clinical condition before surgery.
However, so far only a few studies have been able to demonstrate a prognostic benefit for the short-term outcome resulting from the prenatal diagnosis of isolated congenital heart defect5, 6, and the influence of prenatal diagnosis on the long-term outcome has not yet been analyzed at all. This study compared the short and long-term outcomes of children with prenatal and those with postnatal identification of isolated congenital heart defects.
We focused on four isolated cardiac lesions: transposition of the great arteries (TGA), complete atrioventricular septal defect, tetralogy of Fallot and pulmonary atresia with or without ventricular septal defect. The prenatal group consisted of all children with one of the target cardiac malformations diagnosed prenatally during the study period at the Charité Campus Virchow tertiary center in Berlin; these were all referred on to the German Heart Centre Berlin. The postnatal group consisted of all patients with postnatal identification of the same lesions who had been referred to the German Heart Centre from any of the non-tertiary birth centers in north-east Germany. This meant that for each malformation the postnatal control group was larger than was the prenatal group. To make the two groups as comparable as possible, preoperative cardiac function evaluated by echocardiograms was compared between the two groups to exclude differences in the severity of the cardiac lesion. Moreover, children with one of the target lesions but with associated intracardiac lesions potentially influencing the prognosis (e.g. coarctation of the aorta, aortic stenosis, anomalous pulmonary venous return or complex defects) were excluded, as were children with extracardiac malformations. Trisomy 21 and 22q11 microdeletion were not considered exclusion criteria because several reports suggest that there is no difference in the outcome of children with cardiac defects due to these chromosomal aberrations7–9. Children with all other chromosomal aberrations were excluded. There were 257 children who presented at the German Heart Centre Berlin between 1994 and 2004 and who met the inclusion criteria: 49 with a prenatal diagnosis and 208 with a postnatal diagnosis.
In the group diagnosed prenatally, diagnosis had been achieved at a median gestational age of 24 (range, 17–38) weeks. A multidisciplinary team consisting of an obstetrician, neonatologists and pediatric cardiologists had counseled the mother following diagnosis and they were notified of her arrival at the delivery unit. Immediately after birth, administration of intravenous prostaglandin was initiated in all neonates with duct-dependent pulmonary or systemic circulation. All neonates were transferred to the pediatric cardiac intensive care unit for observation, and early intervention if necessary. Postnatal echocardiography confirmed the diagnosis. All patients were scheduled for surgery at the German Heart Centre Berlin, the timing of which depended on the type of cardiac lesion and clinical status.
In the group diagnosed postnatally, diagnosis had been achieved at a median age of 22 (range, 1–405) days after delivery. Clinical symptoms leading to the diagnosis were acidosis and cardiac dysfunction in 50% of cases and heart murmur in 35%; in 15%, other findings, including clinical suspicion of Down syndrome, were the reason for further cardiac diagnostic procedures. Twenty-seven percent of these children had been discharged from their birth center before the heart defect was diagnosed. Once admitted to the pediatric cardiology department, the patients of the postnatal group received the same diagnostic and therapeutic management as did the prenatal group.
The following clinical parameters were considered as short-term outcome measures: preoperative clinical status (O2 saturation, cyanosis, clinical cardiac dysfunction); first operation and postoperative course (rate of complete repair versus palliation, duration of mechanical ventilation, need of resurgery, duration of intensive care, need for prolonged in-patient care in extern hospital after primary recovery); mortality within first hospital stay.
For long-term follow-up, 1-, 2- and 5-year examinations were evaluated. The most recent examination for each child was termed ‘maximum follow-up’. Long-term outcome measures included: residual cardiac defects after corrective surgery, defined as a residual septal defect or moderate to severe atrioventricular valve insufficiency or aortic/pulmonary stenosis with a gradient on echocardiography across the valve of ≥30 mmHg; dependency on cardiac medication; long-term survival free from catheter intervention; long-term survival free from resurgery; overall long-term survival. The maximum follow-up achieved was 130 (median, 37.3) months.
The four different cardiac malformations were analyzed individually using SPSS for Windows version 13.0 (SPSS Inc., Chicago, IL, USA). If the individual analyses showed similar trends in outcome, a combined statistical evaluation was performed to increase the statistical power. Continuous and ordinate variables were compared between groups by independent-samples t-test or Mann–Whitney U-test. Dichotomous and categorical variables were compared using χ2 analysis or Fisher's exact test. Survival times were assessed according to Kaplan–Meier combined with log rank tests. The level of clinical significance was set at 0.05.
Neonatal characteristics were similar in both patient groups and these are listed in Table 1.
|Gestational age at delivery (weeks, median (range))||39 (35–42)||39 (31–42)||0.17|
|Mean birth weight (g, mean ± SD)||3160 ± 540||3140 ± 650||0.81|
|Mean birth length (cm, mean ± SD)||49.5 ± 3.1||49.9 ± 3.6||0.51|
|Hypotrophy* (n (%))||7 (14)||19 (9)||0.3|
|APGAR ≤7 at 10 min (n (%))||5 (10)||13 (6.3)||0.48|
|Arterial umbilical cord pH (mean ± SD)||7.27 ( ± 0.07)||7.26 ( ± 0.05)||0.33|
|Male gender (n (%))||27 (55)||120 (58)||0.75|
There were 63 cases of trisomy 21: 54/208 (26%) children in the postnatal group and 9/49 (18%) in the prenatal group. There were four cases of 22q11 microdeletion in the postnatal TGA group. There was no statistical difference between the groups regarding the rate of chromosomal aberrations (P = 0.21). Preoperative echocardiography showed a higher rate of tricuspid insufficiency in the prenatal group compared with the postnatal group with pulmonary atresia (75% vs. 30%, P < 0.007). There were no differences between the groups regarding all other echocardiographic parameters.
The median age at admission for surgery was not significantly different between the groups: 10 (range, 1–1050) days in the prenatal group and 22 (range, 1–979) days in the postnatal group. Preoperatively, the prenatal group had a higher median O2 saturation, although the difference was not significant (91% vs. 88%; P = 0.07) and there were fewer cases of cardiac insufficiency (22% vs. 40%, P = 0.03). Pretherapeutic closure of the duct occurred in none of the prenatal group treated with prostaglandins but in 15 children (18%) in the postnatal group (P = 0.04). Preoperative cyanosis was less frequent in the prenatal group (61% vs. 73%), although the difference did not reach statistical significance (P = 0.12).
|Clinical characteristic||Heart defect||Group diagnosed||P|
|Age at admission||Overall||10 (1–1050)||22 (1–979)||0.46|
|(days, median||TGA||2 (1–10)||2 (1–42)|
|((range))||AVSD||57 (2–1050)||111 (1–765)|
|TOF||152 (10–188)||60 (2–339)|
|PA||2 (1–5)||5 (1–979)|
|Cardiac||Overall||11 (22)||83 (40)||0.03|
|insufficiency||TGA||5 (33)||37 (51)|
|(n (%))||AVSD||3 (18)||33 (42)|
|TOF||2 (22)||3 (8)|
|PA||1 (13)||10 (48)|
|O2 saturation||Overall||91 (65–100)||88 (30–100)||0.07|
|(%, median (range))||TGA||85 (65–97)||80 (30–94)|
|AVSD||96 (77–100)||95 (70–100)|
|TOF||97 (75–100)||89 (50–100)|
|PA||89 (80–94)||80 (40–98)|
|Cyanosis (n (%))||Overall||30 (61)||152 (73)||0.12|
|TGA||14 (93)||72 (100)|
|AVSD||6 (35)||36 (46)|
|TOF||4 (44)||24 (67)|
|PA||6 (75)||20 (95)|
|Closure of the||Overall||0 (0)||15 (18)||0.04|
|ductus before||TGA||0 (0)||12 (18)|
|(n (%))||TOF||—||3 (100)|
|PA||0 (0)||0 (0)|
There was no difference in the rate of primary corrective surgery versus primary palliation between the groups (P = 0.82), although looking at the four heart defects independently in each lesion, the number of cases operated correctively in the prenatal group exceeded that in the postnatal group. The median duration of postoperative mechanical ventilation was shorter in the prenatal group than it was in the postnatal group (4 vs. 6 days, P = 0.03). None of the prenatal group but 10% of the postnatal group needed resurgery during the hospital stay (P = 0.02). Sub-analysis showed that the children requiring resurgery were admitted more frequently with cardiac insufficiency (P = 0.001) than those without the need for resurgery. The median stay at the intensive care unit was shorter in the prenatal group (6 vs. 8 days, P = 0.05). After recovery from surgery, only one child in the prenatal group needed prolonged in-patient observation in an outside institution, compared with 17% in the postnatal group (P = 0.01). Postoperative survival was slightly better in the prenatal group (96% vs. 90%, P = 0.27), but the difference was not statistically significant. Sub-analysis of the children who died within the first hospital stay revealed that these children were significantly more likely to need resurgery (P = 0.001), have low cardiac output failure (P = 0.02), have a capillary leakage after surgery (P = 0.02), and need postoperative circulatory support using extracardiac membrane oxygenation (ECMO) (P = 0.001).
|Primary correction||Overall||42 (87.5)||176 (85)||0.82|
|(n (%))||TGA||15 (100)||70 (97)|
|AVSD||16 (94)||68 (86)|
|TOF||9 (100)||33 (92)|
|PA||2 (29)||5 (25)|
|Duration of||Overall||4 (0–97)||6 (0–125)||0.03|
|postoperative||TGA||4 (2–97)||6 (1–20)|
|mechanical||AVSD||3 (1–20)||6 (1–86)|
|ventilation (days,||TOF||2 (1–14)||4 (1–45)|
|median (range))||PA||4 (0–10)||5 (0–125)|
|Resurgery during||Overall||0 (0)||21 (10)||0.02|
|hospital stay (n (%))||TGA||0 (0)||2 (3)|
|AVSD||0 (0)||14 (18)|
|TOF||0 (0)||1 (3)|
|PA||0 (0)||4 (19)|
|Duration of intensive||Overall||6 (0–92)||8 (0–86)||0.05|
|care (days, median||TGA||8 (2–92)||9 (2–53)|
|(range))||AVSD||7 (2–34)||9 (2–72)|
|TOF||3 (2–24)||6.5 (3–44)|
|PA||5 (0–17)||7 (0–86)|
|Prolonged in-patient||Overall||1 (2)||32 (17)||0.01|
|care in extern||TGA||1 (7)||14 (22)|
|hospital* (n (%))||AVSD||0 (0)||12 (17)|
|TOF||0 (0)||4 (11)|
|PA||0 (0)||2 (12)|
|Death within first||Overall||2 (4)||21 (10)||0.27|
|hospital stay (n (%))||TGA||0 (0)||8 (11)|
|AVSD||2 (12)||8 (10)|
|TOF||0 (0)||1 (3)|
|PA||0 (0)||4 (19)|
The median maximum follow-up period of the survivors was 37.8 (range, 0.5–111.2) months in the prenatal group and 48.3 (range, 0.4–130.1) months in the postnatal group. The survival period free from intervention after corrective surgery was significantly longer in the prenatal group (survival free from catheter intervention: P = 0.03, Figure 1; survival free from resurgery: P = 0.07, Figure 2). At the 1-year follow-up, residual impaired cardiac function after corrective surgery was significantly less frequent in the prenatal group (31% vs. 54%, P = 0.04). Statistical significance was lost at maximum follow-up (34% vs. 49%, P = 0.13). There were no significant differences in dependency on cardiac drugs in the prenatal compared with the postnatal group (42% vs. 52% at 1 year, 28% vs. 34% at 2 years, 20% vs. 26% at 5 years and 21% vs. 30% at maximum follow-up). The overall long-term survival was 92% in the prenatal group compared with 84% in the postnatal group; this difference was not significant (Figure 3).
Several studies have shown that survival of newborns with major congenital heart disease is influenced by the point in time at which treatment is started10, 11. A delay in diagnosis can lead to detrimental closure of the duct in duct-dependent lesions and to progressive cardiac insufficiency in duct-independent lesions, with the associated risks of hypoxia, acidosis, hemodynamic collapse, brain damage and multiorgan failure. Emergency surgical intervention for corrective or palliative procedures under such circumstances bears the risk of postoperative cardiac failure and may influence the early and long-term morbidity and the later physical and psychomotor development12–14. Prenatal diagnosis allows delivery in experienced tertiary centers with early initiation of the appropriate treatment, if indicated. Despite this potential advantage, many studies have failed to show a clear benefit of prenatal diagnosis15, 16; some even showed an adverse outcome in the prenatal series17. There are several possible reasons for these conflicting results. In cohorts in which the diagnosis of congenital heart disease has been made prenatally, children with associated extracardiac malformations or chromosomal abnormalities predominate, because a complex disease is more likely than a minor defect to be detected prenatally18–20. Moreover, many studies include populations of heterogeneous heart defects combined with additional cardiac anomalies, making their comparability limited3, 21. In this study, therefore, we focused on children with isolated congenital heart defects. The strict exclusion of patients with associated intra- and extracardiac malformations significantly reduced the number of cases, but it increased the value of comparisons between the groups and accordingly the value of this comparative study. We deemed this a priority, because we sought to investigate not only short- but long-term measures, which are at increased risk of being obscured by multifactorial influences. Moreover, we restricted our study to a single institution to exclude confounding factors such as differing preoperative management, operation technique or surgical expertise. Finally, some studies might have failed to show differences because of a small study size in rare cardiac lesions. In our study, we included four different isolated congenital cardiac defects and evaluated each separately. Only when each malformation revealed similar results in the outcome measures, did we perform a combined analysis to strengthen the statistical power.
Some other recent studies have demonstrated the prognostic benefit of prenatal diagnosis. Better preoperative conditions have been found in children with hypoplastic left heart6, left heart obstruction12, coarctation of the aorta22 and TGA5. Duct-dependent congenital heart defects such as TGA often require early intervention, and therefore these children are most likely to benefit from identification of the defect before birth. Our study confirmed a better preoperative state, with higher O2 saturation and less cardiac insufficiency, in the children with a prenatal diagnosis, and this was also evident in duct-independent cardiac lesions.
Importantly, the differences between the study groups with and without prenatal diagnosis of congenital heart disease were detectable in spite of the high standard of infrastructure for primary neonatal care provided in Berlin, which means an experienced neonatologist is present immediately in non-tertiary centers for emergency treatment and referral to our center within a very short time. This is reflected in the fact that the age at referral, especially in the duct-dependent heart defects, did not differ much between the groups. Adding to this the fact that of the postnatal group many were resuscitated and treated with prostaglandins before transferral to our institution, one might speculate that the observed differences would become more obvious in less organized settings.
The prenatal group showed a lower need of resurgery and this again was associated with less cardiac dysfunction at primary referral. The duration of postoperative ventilation as well as the duration of stay at the intensive care unit was longer in the postnatal group. These factors are known to be associated with poor neurodevelopment23, 24.
To our knowledge, this is the first study to investigate the long-term morbidity of children with and without prenatal diagnosis of a congenital heart defect. There was a longer survival period free from catheter intervention and a trend towards a longer overall survival period free from resurgery after primary repair in the prenatal group. One year after corrective surgery we found less residual cardiac dysfunction and a lesser need of cardiac medication in the prenatal group in comparison to the postnatal group.
Bonnet was the first to describe improved survival after prenatal diagnosis in children with TGA5. Lately, improved survival rates have also been detected in children with prenatal diagnosis of hypoplastic left heart syndrome6 and in infants amenable to biventricular repair23. Though the difference was not statistically significant, the children with prenatal diagnosis in this study had better postoperative survival and this was associated with a decreased risk of resurgery and related complications such as cardiac failure, capillary leakage or the need for ECMO. The risk of resurgery was again strongly associated with the clinical condition at admission. We therefore believe that the early survival advantage can be attributed to the better preoperative clinical condition of the children with diagnosis before birth. The children with prenatal diagnosis also had better long-term overall survival than did children with postnatal diagnosis, but again the differences did not reach statistical significance.
A limitation of the study is its retrospective nature. Moreover, the study did not allow us to identify those neonates without prenatal diagnosis that died before referral to our center because of circulatory shock due to spontaneous closure of the duct. For such information, all pediatric deaths in a well-defined population would have to be evaluated regarding a potential cardiac origin, with autopsy confirming the suspicion. As population-based studies have revealed, death from unrecognized congenital heart disease occurs in approximately 1.6–1.7% of all children born with a congenital heart defect25, 26. Consequently, no conclusions can be drawn regarding whether prenatal diagnosis reduces preoperative mortality. If these deaths had been recognized, it seems likely that we would have shown an even bigger survival advantage in the prenatal compared to the postnatal group. Finally, because of the strict inclusion criteria, the number of cases included in this study seemed too small to keep statistical power in some long-term measures, especially long-term survival. Larger studies are needed for certain long-term measures.
In conclusion, the prenatal diagnosis of congenital heart defects is associated with lower immediate and long-term morbidity and improved survival rates. So far, a general prenatal screening program for congenital heart defects has been rejected, because of an unfavorable cost–benefit analysis. With more studies emerging proving the advantage of prenatal rather than postnatal identification of such lesions with respect to long-term morbidity, quality of life and survival of the affected children, it may become worth reconsidering the matter of a general prenatal screening program.
We would like to thank Prof. Plagemann, Dept. Of Experimental Obstetrics, for the helpful discussions especially regarding statistical questions.