Scoring systems

Various scoring systems have also been devised to predict the chance of a cardiac complication during pregnancy. The CARPREG score is the most commonly used . The strengths of this scoring system are that it was developed in a prospective multicentre study (which was carried out in Canada). Outcomes of 599 pregnancies in 562 women initially enrolled were reported. It included women with both congenital and acquired heart disease.

The limitations of the study were that it excluded women who terminated their pregnancy or miscarried before 20 weeks’ gestation. Although the study used recent pregnancies at the time, recruiting women between October 1994 and November 1999, the data are now old, and treatment of heart disease has changed significantly in this time.

The study reported on the risk of serious cardiac events in pregnancy, the sort of events which may deter women from embarking on pregnancy; these were pulmonary oedema, arrhythmia, stroke, cardiac arrest, or cardiac death. The scoring system is simple to use with four predictors, each scoring one point.

The four predictors of primary cardiac events were as follows: (1) prior cardiac event (e.g. heart failure, transient ischaemic attack, or stroke before pregnancy) or arrhythmia; (2) baseline NYHA class > II or cyanosis; (3) left heart obstruction (mitral valve area <2 cm ² , aortic valve area <1.5 cm ² , or peak left ventricular outflow tract gradient >30 mm Hg by echocardiography); and (4) reduced systemic ventricular systolic function (ejection fraction ≤ 40%).

The total score then predicts the risk of a serious cardiac event ( Table 3 ).

Neonatal complications occurred in 20% of pregnancies, and were associated with poor functional class or cyanosis, left heart obstruction, anticoagulation, smoking, twins, and higher order multiple pregnancies.

The CARPREG score has subsequently been validated in other institutions, and seems to have a high sensitivity and high negative predictive value . Almost all women, however, were in the low or intermediate risk group.

Other groups have subsequently sought to refine the CARPREG scoring system. A retrospective study from a Belgian tertiary centre studied 1302 women with congenital heart disease who delivered between 1980 and 2007 . The scoring system they developed seemed to predict cardiac complications more accurately in women with congenital heart disease. Their proposed system, however, is more complex than CARPREG, with different scores given to each risk factor. As consent had to be obtained from each woman included, no women who died were included. Women with complex congenital heart disease were also under-represented, as it is likely that these women were advised against pregnancy, particularly in the 1980s, the earlier part of the study period.

This study reaffirmed the risk factors described in the CARPREG scoring system, but also added moderate or severe systemic or pulmonary atrioventricular valve regurgitation or a mechanical prosthetic valve, in women with congenital heart disease. Cyanosis, a mechanical valve replacement, smoking, multiple gestation, and cardiac medication were associated with poor fetal or neonatal outcome .

An issue for a practising clinician is whether a scoring system should be used at all, and whether it is useful to quote specific percentages to women, given the limitations of both studies. Complications that are specific to certain conditions would not be predicted by such scoring systems; for example, the risk of aortic dissection in Marfan’s syndrome. Perhaps the most importance message to take from both of these studies is which risk factors are of most important in predicting outcome. It is then important for the clinician to take other comorbidities or obstetric risk factors into account when trying to provide realistic counselling about the risks involved in a pregnancy.

Cardiopulmonary exercise testing

In pregnancy, cardiac output increases. It is therefore reasonable to consider whether a cardiopulmonary exercise test, which mimics this increase in cardiac output, could predict whether a woman will be able to mount an appropriate haemodynamic response in pregnancy, and whether it would predict pregnancy outcome.

Lui et al. studied 83 women with congenital heart disease (89 pregnancies), who had a cardiopulmonary exercise test either before pregnancy or during the first trimester. One or more adverse cardiac events occurred in 18% (heart failure in 14%, and sustained arrhythmia in 7%). The investigators found that failure to raise heart rate to expected levels during the test was associated with cardiac events in pregnancy and adverse fetal or neonatal outcomes. In particular, a chronotropic index ([peak heart rate-resting heart rate]/[220-age-resting heart rate]) of less than 0.8 was associated with such events. Peak oxygen consumption was not correlated with outcome in this study.

Ohuchi et al. performed a similar study, again in women with congenital heart disease. They concurred that the heart rate response was predictive of maternal cardiac events but also found that peak oxygen uptake correlated with poor fetal or neonatal outcomes. A peak heart rate of less than 150 beats per minute, a peak oxygen consumption of less than 25 ml/kg/min, or both, were predictors of maternal cardiac events, and a peak oxygen consumption of less than 26.2 ml/kg/min predicted an adverse neonatal outcome.

Aortopathy and aortic dissection

None of the above scoring systems would predict the risk of dissection in women with aortopathy. Therefore, disease-specific risk prediction is needed. In the most recent maternal mortality report in the UK covering maternal deaths in between 2003 and 2005, aortic dissection accounted for seven out of 53 cardiac deaths . Research shows that women in whom Marfan syndrome, Ehlers Danlos Type IV, a bicuspid aortic valve, who may have an associated aortopathy, Turner’s syndrome, and coarctation of the aorta are at particular risk, of dissection in pregnancy. The risk is higher in older women and women who suffer from hypertension. Family history of aortic dissection is particularly relevant in women with Marfan’s syndrome.

For women with Marfan syndrome, the risk of dissection increases with increasing aortic root size, and has been found to be 1% in women with an aortic root less than 4 cm in diameter, and 10% if the aortic root is greater than 4 cm in diameter . However, 4 cm should not be seen as an absolute cut off for defining risk, as pregnant women are at particular risk as the aorta increased in size during pregnancy, even if the aortic root measures less than 4 cm . A recent Dutch study suggested that a figure of 4.5 cm could be used, as no aortic dissections occurred in women without previous aortic dissection and an aortic root diameter less than or equal to 45 mm .

Similar figures are not available for women with non-Marfan aortopathy. In women with Ehlers Danlos syndrome the aorta may dissect without any antecedent increase in size.

Previous peripartum cardiomyopathy

All women with a previous peripartum cardiomyopathy are at high risk of recurrence in a subsequent pregnancy (30–50%), and should be referred to a specialist joint obstetric cardiac service. Those at highest risk are the women whose ejection fraction did not return to normal after the index pregnancy. A study by Elkayam et al. compared women whose ejection fraction had normalised (>50%) with those in whom it had not and showed that the women whose ejection fraction had not normalised were at higher risk in a subsequent pregnancy of developing symptoms of heart failure (44% v 21%), dying (19% v 0%) and delivering before 37 weeks (37% v 11%).

In the most recent UK maternal mortality report, deficiencies in the care of women with previous peripartum cardiomyopathy were described, and emphasis was placed on counselling to make women aware of the risks of recurrence or deterioration in subsequent pregnancies . The lack of clinicians’ awareness and knowledge of the risks was also highlighted, and the report stated that ‘it is crucial to refer women with previous cardiomyopathy to specialist units with joint clinics where both the obstetricians and cardiologists are familiar with the expected changes of pregnancy.’

Mitral valve disease

Mitral regurgitation is much better tolerated in pregnancy than mitral stenosis. As valve area remains unchanged during pregnancy (unlike valve gradient), it can be used as a tool to risk stratify a woman, whenever she presents . The severity of mitral stenosis has been shown to correlate with maternal and fetal outcomes with a mitral valve area of 1.5 cm ² or less, carrying a 78% chance of heart failure and a 33% chance of arrhythmias .

Aortic valve disease

Aortic regurgitation is much better tolerated in pregnancy than aortic stenosis. The risk of complications during pregnancy is related to the severity of the gradient across the aortic valve. Good left ventricular function, normal heart rate, and blood pressure responses during an exercise test are predictors of good maternal and fetal outcomes .

Mechanical prosthetic valves

Women with mechanical prosthetic valves in situ have a maternal mortality risk, which has been estimated between 1 and 4% . The level of risk is influenced by the type of prosthetic valve, the position of the valve (mitral valves are more prone to thrombosis than aortic valves), and compliance with anticoagulation treatment during pregnancy .

The underlying pathology which led to the valve replacement should also be considered when assessing risk. Women with congenital heart disease who have required prosthetic valve replacement often have complex cardiac lesions and may experience other cardiac complications such as arrhythmias. Small sized valves are often used if the valve replacement occurs in childhood, resulting in a functional stenosis in the adult, particularly with the increased cardiac output of pregnancy. Intravenous drug users who develop endocarditis and require prosthetic valve replacement may continue to have chaotic lifestyles and be poorly compliant with treatment. Other factors such as smoking, poor intravenous access, poor diet, and social issues increase the obstetric and neonatal risks (see also Chapter 4).

Debate continues on which anticoagulant regimen is better in pregnancy, and whether low molecular weight heparin is as efficacious as warfarin, but this is outwith the scope of this chapter.

Tetralogy of Fallot

Usually, women with Tetralogy of Fallot tolerate pregnancy well. Those with severe pulmonary regurgitation are at highest risk of complications, which can include right ventricular failure (1.6%); arrhythmias (6.5%); preterm labour (18%), and small for gestational age babies (19%) .

Transposition of the great arteries

The risks for women during pregnancy depend on the type of surgery the woman has undergone.

In atrial switch procedures (operations such as a Mustard or Senning procedure), the right ventricle supplies the systemic circulation. These women are prone to develop arrhythmias and heart failure during pregnancy. About 38% will deteriorate in pregnancy and, in 12%, this will be a permanent deterioration in cardiac function .

Over the past 30 years, arterial switch procedures have been carried out in preference to the atrial switch described above. In this operation, the great arteries are switched around, so that the left ventricle supplies the systemic circulation. As this is a more modern operation, only small numbers of pregnancies have been reported . It is thought that these women could potentially be prone to myocardial ischaemia and aortic root dilatation in pregnancy.

Fontan

All women with a Fontan type circulation are at high risk of maternal and fetal complications in pregnancy. Within this group, it is difficult to predict which women are at higher risk, as only small case studies have been reported . They were also under-represented in the cohorts studied by Siu et al. and Drenthen et al. when developing their scoring systems, as many were advised against pregnancy.

Pulmonary hypertension

All women with pulmonary hypertension are at high risk of maternal mortality. Bedard et al. carried out a systematic review of the literature and found 48 papers describing outcomes in 73 parturients who delivered between 1997 and 2007. An overall mortality rate of 25% was reported, but this was lowest (17%) in the women with idiopathic pulmonary hypertension and highest (33%) in women who had developed pulmonary hypertension secondary to an underlying medical condition. In women with congenital heart disease, it was 28%.