Results

Fifty women with an atrial septal defect, 22 with a ventricular septal defect, 19 with tetralogy of Fallot, 22 with an atrioventricular septal defect, and 13 with complex shunt defects reported pregnancies. Mean age at last follow up of the 126 patients who had become pregnant was 45 (13) years; 84 (67%) of them underwent corrective or definitive surgery at a mean age of 24.4 (18) years. The mean shortening fraction was 40 (8)% and cyanosis was present in 2%.

Table 1 presents the characteristics of all the women who had been pregnant, in their different diagnostic groups. All women with tetralogy of Fallot had undergone corrective surgery; in that respect, this group differed significantly (p < 0.01) from those with atrial septal defect, ventricular septal defect, and complex lesions. Women with an atrial septal defect were older at the time of corrective surgery than those with tetralogy of Fallot (p = 0.0003). The left to right shunt and the shortening fraction of the systemic ventricle were comparable between the different diagnostic groups. The shortening fraction of the systemic ventricle was normal in all patients at late follow up.

Ejection fraction was determined invasively in 10 patients with complex lesions. In three of five women who had not undergone corrective surgery the value was less than 60%; in five patients who had undergone corrective surgery, the calculated ejection fraction was greater than 60%. Angiographic determination of ejection fraction was not done in three patients.

Overall, 309 pregnancies were reported by the 126 women, giving a mean (SD) pregnancy rate of 2.5 (1.6) per woman; mean pregnancy duration was 8.8 (0.8) months and the mean miscarriage rate (spontaneous fetal loss) was 0.4 (0.9) per woman (51 abortions; 17%). Table 2 gives the characteristics of these 309 pregnancies. The number of pregnancies, rate of complications, and incidence of reduction in wellbeing at the time of the last follow up visit were comparable among the groups, but there was a trend towards a higher abortion rate in patients with complex lesions. However, in women with an atrial septal defect more pregnancies occurred before corrective repair than in the other groups, because in many cases the diagnosis had not been made before these women became pregnant. This explains why they were older at the time of corrective surgery (table 1).

Pre-eclampsia and embolic events were observed in 1.3% and 0.6% of all pregnancies, respectively. Embolic events included pulmonary embolism in one woman and a cerebrovascular accident in another. The duration of pregnancy was slightly shorter in patients with complex shunt lesions, who were more often delivered by caesarean section (p = 0.005). At the time of the last follow up, patients with complex shunt lesions reported that they were more dyspnoeic than those with atrial septal defects (p = 0.005). Cardiac reasons were never cited as an indication for caesarean section.

Two thirds of the patients (n = 84; 67%) underwent corrective or definitive surgery. As expected, women who had undergone corrective/definitive surgery had significantly larger left to right shunts preoperatively (52 (12)% v 36 (15)%), and were not cyanotic (0% v 7.3%) at follow up. The number of pregnancies per woman (2.6 (1.8)v 2.2 (0.9)) was not statistically different between women with and without surgery. Mean age at last follow up (45 (12) v 45 (14) years), shortening fraction of the systemic ventricle (40 (8)% v 39 (8)%), duration of pregnancy (8.8 (0.8) v8.8 (0.9) months), number of live births per woman (2.0 (1.3)v 2.0 (0.8)), abortion rate (18%v 13%), caesarean section rate (24%v 26%), and complication rate during pregnancy (pre-eclampsia 4.7% v 0%; embolic events 1.2% v 2.4%) were comparable between women with and without corrective/definitive surgery. Birth weight (3121 (549) v 3057 (613) g) and body length (49 (3) v 48 (4) cm) were also not statistically different between these two groups.

Only one woman reported having arrhythmias during pregnancy. This patient had transposition of the great arteries and had undergone the Mustard procedure. The arrhythmias were successfully treated with β blockers. No patient developed heart failure during pregnancy or in the postpartum period.

Thirteen of the 24 women with complex lesions had 28 pregnancies (2.2 (1.7) pregnancies per woman). These 13 patients gave birth to 17 offspring, among whom there were no reported cases of congenital heart defects.

There were 241 completed pregnancies in the series. Body size and birth weight of the offspring were comparable between the different diagnostic groups (table 3). However, the babies of mothers with complex shunt lesions were slightly, but significantly, smaller than those of mothers with other defects.

The four cyanotic patients had five live births. One baby died on the second day after delivery. This baby, whose mother had pulmonary atresia and a ventricular septal defect, was delivered by caesarean section for impending asphyxia in the seventh month of gestation. The cause of death was unknown.

The risk of recurrent congenital heart defects in the 241 completed pregnancies was 2.5% (table 3). No heart defects were found in the offspring of mothers with tetralogy of Fallot or complex shunt lesions. Most defects were found in offspring of mothers with atrial septal defects, with the same lesion predominating in these babies (three with an atrial septal defect and one with a ventricular septal defect were born alive; one with a complex congenital heart defect was spontaneously aborted in the 20th week of gestation). One woman with a ventricular septal defect gave birth to a baby with a ventricular septal defect. One woman with an atrioventricular septal defect gave birth to a baby with a complex lesion.

Overall mortality in the women who had become pregnant was 1.6% (n = 2; one with an atrial septal defect and one with an atrioventricular septal defect). Both deaths occurred more than one year after delivery.

Discussion

People with congenital heart disease are increasingly reaching adulthood owing to remarkable diagnostic and therapeutic advances in cardiology and cardiac surgery.1 Thus there are now many women with congenital heart disease who are considering pregnancy, and the maternal and fetal risks in this heterogeneous group of patients must be carefully evaluated. In this retrospective study we report the experience of a single centre relating to the outcome of pregnancy in women with congenital shunt lesions, including atrial and ventricular septal defects, tetralogy of Fallot, atrioventricular septal defects, and other complex shunt lesions.

The pregnancy rate (2.5 per woman) and birth rate (2.0 children per woman) in our patients were higher than those in the general population in Switzerland, where the birth rate is 1.5 per woman.9Only 0.5 pregnancies per woman were observed before corrective surgery, though women with an atrial septal defect had more pregnancies before surgery than women in the other diagnostic groups (table 2). Many of these women did not know they had a heart defect before they became pregnant. This observation explains their greater age at the time of corrective surgery (table 1). The overall outcome of pregnancy was good in our population.

The outcome of pregnancy was not associated with the type of previous surgery. This is consistent with the findings in a prospective study of 233 women with congenital heart disease and a recently published retrospective study of a further 221 women.6 ,10 Poor maternal functional class or cyanosis were independent predictors of neonatal events.6 The abortion rate was 0.4 per woman (17%), which is higher than previously reported.11

The duration of pregnancy was shorter in women with complex lesions, who gave birth to smaller babies for equivalent gestation than the women in the other groups. However, neither the abortion rate nor the complication rate differed statistically between the diagnostic groups, though there was a trend towards a higher abortion rate in women with complex lesions. The shorter duration of pregnancy in women with complex lesions can be explained by the fact that they more often had elective caesarean sections (70% v15–30%). Although heart problems were not cited as the reason for caesarean section in any of these women, the complex anatomy may have led to the higher caesarean section rate in this population.

Shime et al reported a major benefit of corrective cardiac surgery in cyanotic patients, leading to fewer spontaneous abortions and more term births.12 Our four cyanotic patients had five live births, but one child died on the second day. The risk of pregnancy in cyanotic patients without pulmonary hypertension has been shown to be low for the mother, but the fetal outcome is poor, with high rates of abortion, intrauterine growth retardation, intraventricular haemorrhage, premature birth, and neonatal death.7 A recently published retrospective study that did not include women with Eisenmenger syndrome did not confirm an association between pulmonary hypertension and pregnancy complications.6 However, this study only included patients with moderate pulmonary hypertension, who are at lower risk than patients with Eisenmenger syndrome. Careful monitoring in the intensive care unit is recommended for the most favourable outcome in patients with cyanosis and pulmonary artery hypertension.13-16

Pregnancies in women who have had an atrial switch operation or a Fontan procedure are reported to be well tolerated in selected groups, but the incidence of abortion is high during the first trimester.17-19 A Senning procedure was performed in three of our patients and a Mustard procedure in two, but no women underwent a Fontan procedure in our population.

Pregnancy and its outcome were comparable among the different diagnostic groups, excluding those with complex shunt lesions. This observation is consistent with other reports.20 ,21 In our study, patients with complex lesions were more often symptomatic and complained of more dyspnoea than the others during late follow up (table 2). The worse functional status in these patients (reflected by their NYHA class) may be explained by their more complex anatomical abnormalities and their residual haemodynamic abnormalities.

The incidence of pre-eclampsia in our patients was comparable with that in the general population.22

Embolic events are reported to occur in 0.7% of healthy individuals during pregnancy and the risk seems to be increased in the postpartum period.23 ,24 This observation is similar to that in our population (0.6%) including only a small number of women with complex lesions and cyanosis. However, patients with cyanotic congenital heart defects are reported to be at higher risk of embolic events.25

Long term mortality was 1.6% in our patients who had reported pregnancies. However, not all deaths were related to pregnancy or delivery.

The recurrence risk of congenital heart disease was 2.5% and was not affected by the principal diagnosis. The recurrence rate was highest in women with atrial septal defects and the same lesion predominated in their offspring. The fetal risk of recurrence of congenital heart disease is reported to be 0% to 11%.26-28 No routine echocardiographic assessment was performed in the newborn infants. Thus we may have underestimated the recurrence rate in our population. However, most newborn infants are followed by a paediatrician and all children undergo regular medical examinations at school in Switzerland. Thus the number of missed diagnoses is likely to be low. The prevalence of small for dates babies in women with atrial septal defects, ventricular septal defects, and atrioventricular septal defects was comparable with that in an unselected US population (6% of babies < 2500 g at term),11 but it was much higher in our patients with tetralogy of Fallot and complex shunt lesions.