Article Text

Download PDFPDF

Original research article
Retrospective UK multicentre study of the pregnancy outcomes of women with a Fontan repair
  1. Matthew Cauldwell1,
  2. Philip J Steer1,
  3. Samantha Bonner2,
  4. Omar Asghar3,
  5. Lorna Swan4,
  6. Kenneth Hodson5,
  7. Catherine E G Head6,
  8. Adam Daniel Jakes7,
  9. Nicola Walker8,
  10. Margaret Simpson8,
  11. Aidan P Bolger9,
  12. Farah Siddiqui10,
  13. Katherine M English11,
  14. Lucy Maudlin12,
  15. Dilip Abraham13,
  16. Andrew J Sands14,
  17. Aarthi R Mohan15,
  18. Stephanie L Curtis16,
  19. Louise Coats17,18,
  20. Mark R Johnson1
  1. 1 Academic Department of Obstetrics and Gynaecology, Chelsea and Westminster Hospital, London, UK
  2. 2 Department of Obstetrics, St Mary’s Hospital Manchester, Manchester, UK
  3. 3 Department of Adult Congenital Heart Disease, Manchester Heart Centre, St Mary’s Hospital, Manchester, UK
  4. 4 Department of Adult Congenital Heart Disease, Royal Brompton and Harefield NHS Foundation Trust, London, UK
  5. 5 Department of Obstetrics, Royal Victoria Infirmary, The Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle upon Tyne, UK
  6. 6 Adult Congenital Heart Disease Service, St Thomas’ Hospital, London, UK
  7. 7 Department of Obstetrics, St Thomas’ Hospital, London, UK
  8. 8 Scottish Adult Congenital Cardiac Service, Golden Jubilee National Hospital, Glasgow, Scotland, UK
  9. 9 Department of Adult Congenital Heart Disease, Leicester University Hospital, Leicester, UK
  10. 10 Department of Obstetrics, Leicester University Hospital NHS Trust, Leicester, UK
  11. 11 Department of Adult Congenital Heart Disease, Leeds Teaching Hospitals NHS Trust, Leeds, UK
  12. 12 Department of Obstetrics, Norfolk and Norwich University Hospital, Norwich, UK
  13. 13 Department of Adult Congenital Heart Disease, Norfolk and Norwich University Hospital, Norwich, UK
  14. 14 Department of Adult Congenital Heart Disease, Royal Belfast Hospital for Sick Children, Belfast, UK
  15. 15 Department of Obstetrics, University Hospitals Bristol NHS Foundation Trust, Bristol, UK
  16. 16 Department of Adult Congenital Heart Disease, University Hospitals Bristol NHS Foundation Trust, Bristol, UK
  17. 17 Department of Adult Congenital Heart Disease, Freeman Hospital, The Newcastle upon Tyne Hospitals NHS Foundation Trust, Newcastle upon Tyne, UK
  18. 18 Adult Congenital Heart Disease, Institute of Genetic Medicine, Newcastle University, Newcastle upon Tyne, UK
  1. Correspondence to Professor Mark R Johnson, Academic Department of Obstetrics and Gynaecology, Chelsea and Westminster Hospital, 369 Fulham Road London, UK; mark.johnson{at}


Background The population of women of childbearing age palliated with a Fontan repair is increasing. The aim of this study was to describe the progress of pregnancy and its outcome in a cohort of patients with a Fontan circulation in the UK.

Methods A retrospective study of women with a Fontan circulation delivering between January 2005 and November 2016 in 10 specialist adult congenital heart disease centres in the UK.

Results 50 women had 124 pregnancies, resulting in 68 (54.8%) miscarriages, 2 terminations of pregnancy, 1 intrauterine death (at 30 weeks), 53 (42.7%) live births and 4 neonatal deaths. Cardiac complications in pregnancies with a live birth included heart failure (n=7, 13.5%), arrhythmia (n=6, 11.3%) and pulmonary embolism (n=1, 1.9%). Very low baseline maternal oxygen saturations at first obstetric review were associated with miscarriage. All eight women with saturations of less than 85% miscarried, compared with 60 of 116 (51.7%) who had baseline saturations of ≥85% (p=0.008). Obstetric and neonatal complications were common: preterm delivery (n=39, 72.2%), small for gestational age (<10th percentile, n=30, 55.6%; <5th centile, n=19, 35.2%) and postpartum haemorrhage (n=23, 42.6%). There were no maternal deaths in the study period.

Conclusion Women with a Fontan circulation have a high rate of miscarriage and, even if pregnancy progresses to a viable gestational age, a high rate of obstetric and neonatal complications.

  • Fontan procedure
  • pregnancy
  • perinatal mortality

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The Fontan circulation was devised in 1971 to separate pulmonary and systemic circulations in patients with complex congenital conditions characterised by the presence of a single functional ventricle. Its success in altering the natural history of these conditions means that the already substantial number of women with this condition in the UK adult population is expected to increase by around 60% over the next decade1; many of them will wish to embark upon a pregnancy. It is clear that women with a Fontan circulation find it difficult to support a pregnancy, probably because the systemic ventricle’s ability to increase its cardiac output is compromised by restricted preload and chronically elevated systemic venous pressures.2 Pregnancy also predisposes to the development of atrial arrhythmias.3 4 Prophylactic anticoagulation during pregnancy is advocated by some, but may exacerbate any bleeding tendency and increase the risk of postpartum haemorrhage (PPH).4 5 Those with fenestrated circuits or collateral vessels may have cyanosis, which is associated with increased rates of miscarriage and fetal complications including preterm delivery and fetal growth restriction.6 7

Given these complications, women need comprehensive preconception counselling (PCC) from appropriately experienced multidisciplinary teams (MDT). However, there are limited data to inform these discussions.4 The aim of this study was to define pregnancy outcomes from a UK cohort of patients with single ventricle physiology palliated with Fontan repair.


Study population

All UK specialist centres providing a joint obstetric and adult congenital heart disease (CHD) service were emailed in September 2016 and invited to contribute data on women with a Fontan circulation who had been pregnant. Pseudoanonymised data were collected through review of medical and obstetric case notes. The study period was January 2005 to November 2016.

Data collection

Demographic and baseline maternal data included underlying cardiac anatomy, surgical repair, presence of fenestration, age at Fontan completion, history of arrhythmia or thromboembolic events, New York Heart Association (NYHA) class, oxygen saturations, anticoagulation regime and qualitative echocardiographic assessment of ventricular function and atrioventricular valve regurgitation. Cardiac Disease in Pregnancy Risk (CARPREG) scores were calculated for all women.8

Complications were classified as maternal cardiac, obstetric or neonatal. Maternal cardiac complications included arrhythmia, heart failure requiring treatment, systemic thromboembolism, development of new cyanosis (resting maternal saturations <92%) and protein losing enteropathy. Obstetric complications included pregnancy-induced hypertension (PIH, >140 mm Hg systolic or >90 mm Hg diastolic after 20 weeks’ gestation); pre-eclampsia (PIH with proteinuria >0.3 g/24 hours urine collection); HELLP syndrome (haemolysis, elevated liver enzymes and low platelets); preterm delivery (delivery <37 weeks’ gestation); PPH (blood loss ≥500 mL at vaginal delivery and ≥1000 mL at caesarean section). Neonatal complications included preterm birth (<37 weeks’ gestation); small for gestational age (SGA; corrected fetal birth weight <10th centile and <5th centile)9 10; stillbirth (fetal demise ≥24 weeks’ gestation); neonatal death (death within first month of birth); occurrence of CHD.


Data were analysed using SPSS V.23 for Windows. Continuous variables were expressed as medians, range and IQR. Correlations were calculated using Pearson’s product moment if variables were continuous and Spearman’s rank-order correlation if either of the variables was ordinal. Differences between continuous variables were assessed with the Mann–Whitney U test as they were not normally distributed. Categorical data were expressed as absolute numbers and percentages of the denominator and compared using Fisher’s exact test or the χ2 test if cell numbers were too large. All tests were two tailed and p<0.05 was considered statistically significant.


Collection of identifiable data previously recorded as part of routine clinical care was restricted to members of the clinical team. Meta-analysis was limited to data provided without patient identifiers (pseudoanonymisation) and therefore this study did not require ethical approval (UK National Research Ethics Committee recommendation).


Study population

Of 14 centres contacted, 12 agreed to participate but 2 subsequently withdrew due to inability to retrieve necessary clinical information. Fifty women had 124 pregnancies. Seventeen women (34%) had one pregnancy, 11 (22%) had two, 15 (30%) had three, two (4%) had four, three (6%) had five, one (2%) had eight and one (2%) had nine. Thirteen (26%) women had no viable pregnancies, 26 (52%) had one, seven (14%) had two, three (6%) had three and one (2%) had four. There were 53 (42.7%) live births in 37 women. In addition, there were 68 (54.8%) miscarriages (63 (92.6%) of these were in the first trimester and 5 (7.3%) second trimester), 2 terminations, 1 intrauterine death at 30 weeks’ gestation and 4 (3.2%) neonatal deaths. Median age at first pregnancy was 27 years (range 19–36 years, IQR 24–29 years). Underlying congenital heart lesions and type of Fontan repair are shown in table 1. Median age at Fontan repair was 7 years (range 1–26 years, IQR 4–12 years). Maternal baseline characteristics are shown in table 2. Forty-two women (84%) were white European, eight (16%) were South Asian. Median body mass index at first pregnancy was 25 (range 18–39, IQR 23–28). Two women smoked pre-pregnancy.

Table 1

Underlying cardiac anatomy and type of Fontan repair

Table 2

Baseline maternal characteristics at first pregnancy

Preconception counselling

Ninety-six (77.4%) of all pregnancies were planned and 41 (82%) of women had received PPC prior to their first pregnancy. Of the 12 women whose first pregnancy was unplanned, only four (33%) had had PCC. Both women who chose a first trimester termination had had PCC.


Prior to pregnancy

Table 3 shows cardiac medications prescribed prior to pregnancy in all 53 pregnancies resulting in live births.

Table 3

Cardiac medication prescribed prior to pregnancy

Anticoagulation and antiplatelet use during pregnancy

Anticoagulant and antiplatelet regimens were individualised and involved multidisciplinary decision. In 86 (70.5%) of the 122 pregnancies for which we have medication data, women were taking warfarin pre-pregnancy and one continued to take warfarin until the third trimester (late presentation). All others stopped warfarin on becoming pregnant. In the 52 of 53 cases where women had a live birth and anticoagulation details were available, 1 (1.9%) took no medication (booked late at 29 weeks), 2 (3.8%) took aspirin alone, 2 (3.8%) took a prophylactic dose of low molecular weight heparin (LMWH) (clexane 40 mg, dalteparin 5000 IU or tinzaparin 4500 IU—based on booking weight11), 12 (23.1%) took prophylactic LMWH and aspirin, 24 (46.2%) took treatment dose LMWH (clexane twice daily or dalteparin twice daily, based on booking weight, with poor compliance in two cases) and 11 (21.2%) took treatment dose LMWH and aspirin. Women taking treatment dose LMWH had factor Xa levels performed in accordance with their local guidance. One patient was converted back to warfarin following development of a subchorionic haematoma and recurrent vaginal bleeding felt to be exacerbated by treatment dose LMWH; she ruptured her membranes at 24 weeks and delivered at 26 weeks with suspected chorioamnionitis.


Three women underwent in vitro fertilisation for unexplained infertility (all had prior miscarriage). None conceived following in vitro fertilisation; one went on to have a successful pregnancy following spontaneous conception.


Sixty-eight (54.8%) of pregnancies ended in miscarriage. Maternal baseline saturations at the first obstetric visit (typically 8–12 weeks) were associated with miscarriage; all eight women with baseline saturations less than 85% miscarried, compared with 60 of 116 women (51.7%) with saturations greater than or equal to 85% (Fisher’s exact test, p=0.008). No significant correlation with miscarriage frequency was seen at saturations greater than or equal to 85% (p=0.595, Spearman’s rank correlation test). The five second trimester (≥14 weeks) miscarriages were all associated with recurrent early pregnancy bleeding; subsequently developing a subchorionic haematoma. All women took LMWH (three prophylactic dose and two treatment dose). It is notable that the median initial haemoglobin concentration in pregnancies ending in miscarriage was 14.4 g/dL (range 9.9–19.6 g/dL, IQR 13.0–16.0 g/dL) whereas in pregnancies with a live birth it was 13.8 gd/L (range 10.9–16.1 g/dL, IQR 13.0–15.0 g/dL) (Spearman’s correlation, p=0.05). This is substantially higher than in normal pregnancies (mean value 12.0 SD 7.0 at 12 weeks of pregnancy), the high values likely being a consequence of maternal cyanosis.12 We did not find any correlation between oxygen saturation or haemoglobin concentration and birth weight centile (Pearson’s correlation coefficients 0.030 and 0.059, p values 0.831 and 0.672, respectively). In women with a documented fenestration prior to pregnancy there were 13 potentially viable births from 22 pregnancies (59%) compared with 41 of 102 (40.2%) women without fenestrations (p=0.145).

Cardiovascular complications in pregnancies progressing beyond 24 weeks

Analysing both the CARPREG score and the NYHA class at the beginning of each pregnancy, neither was useful in predicting miscarriage, or birth weight and gestational age in pregnancies with a successful outcome (Spearman’s correlation, p=0.453 and 0.536; p=0.707 and 0.746 and p=0.444 and 0.854, respectively) (table 4).

Table 4

Maternal cardiac complications during pregnancy

Maternal heart failure was the most common cardiovascular complication encountered, occurring in seven women, all required diuretics. The majority (six out of seven) were asymptomatic pre-pregnancy with only one of these requiring early delivery (33 weeks by caesarean section) because of worsening symptoms, three required inpatient care. One patient in NYHA class II at the onset of pregnancy deteriorated to NYHA class IV by 19 weeks’ gestation and required inpatient care until delivery.

Six pregnancies were complicated by arrhythmia. Arrhythmia during pregnancy was more common if there was a history (5 out of 6 with history vs 1 out of 48 with no history, p<0.0001 by Fisher’s exact test). Only one of the five women who had a documented prior arrhythmia had a previous ablation. One patient had atrial fibrillation at 15 weeks and following delivery; both episodes were treated with cardioversion. One patient was treated with amiodarone for supraventricular tachycardia prior to delivery. Ten women had at least one documented episode of hypoxaemia that developed for the first time during pregnancy.

Postnatally one woman developed multiple pulmonary emboli in the postpartum period, associated with poor compliance with LMWH. There were no cases of protein losing enteropathy prior to or during pregnancy.

Obstetric complications

Obstetric and neonatal complications are listed in table 5. Median gestation at live birth was 33.5 weeks (range 25–40 weeks, IQR 30–37 weeks). Delivery <37 weeks occurred in 38 of 53 (71.7%) cases. In 14/38 (30.2%) cases this was due to spontaneous preterm labour. Thirteen of 38 (24.5%) babies were delivered preterm for fetal reasons and 11 of 38 (20.8%) for combined fetal and maternal indications. Twelve out of 53 births (22.6%) were spontaneous vaginal delivery, 10 women (18.9%) had an assisted vaginal delivery and 31 (58.5%) were delivered by caesarean section (12 elective, 19 emergency).

Table 5

Obstetric and neonatal complications

Median blood loss in the 53 live births was 750 mL (range 160–3000 mL, IQR 500–1200 mL). It exceeded 500 mL at vaginal delivery and 1000mL at caesarean (classified as a PPH) in 23 (43.4%) cases. Blood loss was lowest with spontaneous vaginal delivery (n=13, median 600 mL, range 200–2500 mL, IQR 300–1450 mL) and elective caesarean section (n=12, median 600 mL, range 300–1000 mL, IQR 500–788 mL). It was higher following assisted vaginal delivery (n=10, median 875 mL, range 200–2500 mL, IQR 525–1975 mL) and emergency caesarean section (n=19, median 800 mL, range 160–3000 mL, IQR 600–1800 mL). Major obstetric haemorrhage (blood loss >2000 mL) occurred in seven women. There was no relationship between the use of anticoagulation regimes and postpartum blood loss.

Two pregnancies were complicated by pre-eclampsia, with cases of PIH or HELLP.

Neonatal complications

Thirty-nine (72.2%) babies were born before 37 weeks, 11 of 39 (28.2%) following prelabour rupture of membranes. Thirty-two of 39 (82.1%) required admission to a neonatal unit secondary to immaturity. Thirty out of 54 (55.6%) babies were SGA (<10th centile) with 19 out of 53 (35.2%) <5th centile. The median birth weight centile was 9 (n=54, range 1–74, IQR 4–25). There was one stillbirth at 30 weeks’ gestation (thought secondary to severe fetal growth restriction). There were four neonatal deaths of babies born at: 26 weeks (birth weight centile 4), 27 weeks (birth weight centile 15), 31 weeks (birth weight centile 2) and 31 weeks (birth weight centile 16). All had respiratory distress syndrome and two had necrotising enterocolitis. No babies had CHD.

Cardiac complications at follow-up

Four women defaulted from follow-up after pregnancy. Follow-up data were available for 39 pregnancies in 33 women (median follow-up was 6 years, range 4 months to 10.4 years). There was no observed decline in ventricular function or atrioventricular valve regurgitation, and no maternal deaths at last follow-up. Two women required cardioversion after pregnancy for recurrent atrial arrhythmias (at 4 and 6 years respectively) with one woman undergoing revision of her Fontan with a modified right atrial MAZE.


In this cohort of women with a Fontan circulation, the majority of those who conceived and went on to have a live birth did not encounter serious maternal cardiac complications. This is likely to be because women who fail to conceive, or who conceived and subsequently miscarried, have a more advanced physiology with lower oxygen saturations. It is also almost certainly indicative of the successful provision of a specialised multidisciplinary approach to the management of this complex group.

Although the majority of women in our cohort were NYHA class I or had a CARPREG score of 0 or 1 (86%) pre-pregnancy, we observed a high rate of miscarriage, obstetric and neonatal complications. The background rate of miscarriage in women with a clinically apparent pregnancy, but without heart disease is 20%.13 The rate of miscarriage in our study was 56%. This is double the rate of miscarriage in women with most other forms of CHD14 and similar to the high rates of miscarriage (>40%) reported in large series of women who had cyanotic heart disease.15 Presbitero et al demonstrated that the high rate of miscarriage in women with cyanotic heart disease relates to lower maternal oxygen saturations and higher baseline maternal haemoglobin concentrations.15 It may be that the high rate of pregnancy complications is related to an adverse effect of the Fontan circulation on the venous circulation in the pelvis. Interestingly, within our cohort 24% (13 of 54) of women had a fenestrated Fontan of which the majority had saturations of <92%. While cyanosis has a demonstrated association with miscarriage, which we also found in this cohort, the impact of a fenestration in the Fontan circulation is less clear and from our results does not appear to adversely affect the likelihood of achieving a viable pregnancy. The relationship between cyanosis and fenestration is complex as there are other causes for cyanosis in the Fontan circulation such as collateral formation.

Reassuringly most women had PCC although the nature of this will vary between centres. This is higher than a recent study that examined rates of PCC in women with heart disease.16 While maternal events were low in our cohort, consideration should be given at the time of PCC as to whether elective ablation in the context of prior arrhythmia or, more controversially, strategies to optimise saturations should be considered for women with Fontan circulations embarking on pregnancy.

Obstetric events were common, with 72% of women delivering preterm with 44% of these births due to spontaneous preterm labour. Spontaneous preterm labour is more common in women with heart disease, but this is poorly understood.17 18 Median birth weight centile was 9 suggesting that most babies were growth restricted as well as being born preterm. SGA babies are more common in women with CHD, but the degree of growth restriction in our series is greater than reported in other forms of heart disease.19 Kampman et al linked cardiac disease with abnormal placentation, demonstrating abnormal uteroplacental Doppler flow studies in women with pre-pregnancy cardiac dysfunction.20 21 Serial growth scans are indicated from 24 weeks’ gestation due to the significantly increased risk of SGA.

PPH was more common than expected with an overall rate of 43%. Previously, we have shown that women with a Fontan circulation have a greater risk of PPH than women with other cardiac lesions.22 In 13% of deliveries (7 of 54) women had a major obstetric haemorrhage (blood loss >2000 mL). Blood loss at delivery was lowest with spontaneous vaginal delivery, which is consistent with data from the Registry of Pregnancy and Cardiac Disease.23 However, blood loss was greatest in women who underwent an assisted vaginal delivery, largely as a result of perineal trauma.24 Since many women with a Fontan circulation may be advised to have a shortened second stage,25 the rate of assisted vaginal delivery may contribute to the higher rate of PPH. Interestingly, blood loss did not correlate with the anticoagulation regimen.

The preferred strategy for anticoagulation is controversial; the concern of thromboembolic complications secondary to the prothrombotic combination of the Fontan circulation and pregnancy. Currently, there is no consensus on which anticoagulation regime to give.26 Using aspirin and warfarin outside of pregnancy reduces the incidence of thromboembolic events although it remains unclear which strategy is optimal.27 Since both a Fontan circulation and pregnancy are prothrombotic conditions, it would seem reasonable to use anticoagulants to reduce this risk; however, the evidence is consensus opinion. Finally, there remains significant uncertainty and unease with regard to the long-term impact of pregnancy upon the Fontan circulation.


Our study was retrospective so is susceptible to missing data. Under-reporting of miscarriage and termination of pregnancy is likely. Our cohort is limited to women who were NYHA class I or II, with no women with protein losing enteropathy. This may be due to these women being counselled against pregnancy, or because they fail to conceive. Our study was multicentre and there was no standardised practice for managing this group of women in pregnancy. We are therefore unable to draw conclusions with regard to the long-term impact of pregnancy upon women with the Fontan circulations. We have been able to comment on associations found to be significant but there may be some relevant associations that were not apparent in our study simply by chance because of the relatively small numbers of patients on whom we had data.


Women who conceive with a Fontan circulation and particularly those who have low oxygen saturations have high rates of miscarriage. Those who do not miscarry are more likely to encounter obstetric and neonatal complications; in particular higher rates of preterm delivery, fetal growth restriction and PPH. Adverse cardiac events are relatively uncommon when women are managed in a specialised multidisciplinary setting.

Key messages

What is already known on this subject?

Pregnancies in women with a Fontan repair are relatively uncommon and characterised by high rates of cardiac, obstetric and neonatal complications.

What might this study add?

This is the largest multicentre study to date of pregnancy outcomes in women with a Fontan repair. We show that maternal cardiac complications are relatively uncommon in contrast to neonatal and obstetric complications, which occur in the majority of pregnancies and most frequently in those with lower baseline maternal oxygen saturations. No association was found between anticoagulation strategy and obstetric outcomes, specifically, miscarriage and postpartum blood loss.

How might this impact on clinical practice?

This study will provide clinicians with detailed and contemporary information upon which they can base preconception advice for women with a Fontan repair. Furthermore, our study underlines the importance of managing these women in specialist centres, where they will receive appropriate care from the multidisciplinary team.



  • Contributors MC conceived the idea for the manuscript and contacted all other authors who assisted in providing data. PJS and MC analysed the data. MC, PJS and MRJ wrote the first draft and this was edited and approved by all authors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.