Congenital heart disease (CHD) increases the risk of adverse maternal and neonatal outcomes. However, previous studies have included mainly women with low-risk features. A single-center, retrospective analysis of pregnant women with CHD was performed. Inclusion criteria were the following high-risk congenital lesions and co-morbidities: maternal cyanosis; New York Heart Association (NHYA) functional class >II; severe ventricular dysfunction; maternal arrhythmia, single ventricle (SV) physiology, severe left-sided heart obstruction and severe pulmonary arterial hypertension. Multivariate analyses for predictors of adverse maternal cardiovascular and neonatal outcomes were performed. Forty-three women reported 61 pregnancies. There were no maternal or neonatal deaths. Maternal cardiac (31%) and neonatal (54%) complications were frequent. The most frequent cardiac events were pulmonary edema, arrhythmia, and reduced NYHA class. Previous arrhythmia conferred a 12-fold increase in the odds of experiencing at least one major cardiac complication. Maternal SV physiology was an independent risk factor for low birth weight, risk of neonatal intensive care unit admission and lower gestational age. Maternal cyanosis and severe pulmonary arterial hypertension also predicted adverse neonatal outcomes. In conclusion, mothers without antepartum arrhythmia or functional incapacity are unlikely to experience arrhythmias or a decrease in NYHA class during pregnancy. In addition, SV physiology is a robust predictor of neonatal complications. Antepartum counseling and assessment of maternal fitness are crucial for the woman with CHD.
Advances in cardiothoracic surgery, cardiac anesthesia, and interventional techniques have dramatically improved the prognosis for congenital heart disease (CHD), and there are now more adults than children with CHD. With a growing number of women now reaching reproductive age, evidence-based counseling regarding the risks of pregnancy is imperative. Pregnancy, labor, and delivery place a significant hemodynamic toll on the gravid woman. Such cardiovascular adjustments are well tolerated in women with normal cardiac reserve. However, pregnant women with severe CHD may decompensate, increasing the risk of adverse maternal and fetal outcomes.
Previous studies have been largely retrospective and have mainly reported pregnancy outcomes for individual lesions. Contemporary reports have instead focused on qualitative measures of maternal cardiac fitness. Although these studies describe risk factors for adverse outcomes, they are comprised primarily of relatively healthy women with preserved functional status. Less is known about higher risk women, who are often steered toward elective termination because of anticipated high maternal or fetal risk. Studies restricted to women with severe congenital lesions or significant co-morbidities could provide more accurate information for pregnancy counseling.
Methods
We performed a single-center retrospective analysis of female patients followed at the Ahmanson/University of California Los Angeles (UCLA) Adult Congenital Heart Disease Center (ACHDC) who were included in the ACHDC pregnancy database from 1994 to 2012 and who had high-risk CHD. The center’s protocol for management of pregnant women includes consultation with an ACHD cardiologist and maternal fetal medicine specialist every trimester, fetal echocardiography at 18- to 22-weeks gestation, maternal echocardiography during each trimester, and a comprehensive individualized delivery plan. The selection of “high-risk” criteria was based on the available reports regarding cardiac lesions and co-morbidities associated with adverse maternal or fetal outcomes.
Maternal inclusion criteria were subsequently chosen as having at least one of the following preconception conditions: cyanosis (oxygen saturation at rest <90%); New York Heart Association (NHYA) functional class >II; systemic morphologic left ventricular ejection fraction (EF) ≤40%; systemic right ventricular EF ≤30%; subpulmonic ventricular EF ≤30%; symptomatic sustained arrhythmia (requiring intervention by medical therapy, cardioversion, electrophysiology study, or placement of a pacemaker); symptomatic nonsustained arrhythmia (palpitations); single ventricle (SV) physiology; the presence of any mechanical valve, severe left-sided heart obstruction (LHO), or severe pulmonary arterial hypertension (PAH, systolic pulmonary artery pressure on echocardiography or cardiac catheterization >60 mm Hg). SV physiology included those who had received Fontan palliation, a Glenn shunt alone and unpalliated patients. Ventricular dysfunction was defined as either quantitative (EF as specified for each ventricular morphology) or qualitative as assessed by the echocardiographer. Severe LHO was defined as follows: aortic valve stenosis (valve area <1 cm 2 , mean gradient >40 mm Hg, or peak velocity >4 m/s); subaortic stenosis (peak instantaneous gradient of >50 mm Hg or mean gradient >30 mm Hg); coarctation of the aorta (Doppler-derived peak instantaneous gradient at rest >40 mm Hg) or mitral stenosis (mitral valve area <1 cm 2 or mean gradient >10 mm Hg). Pregnancies were surveyed for maternal cardiac and neonatal adverse outcomes.
We interrogated the medical record for the following maternal cardiac outcomes: pulmonary edema, symptomatic sustained or nonsustained arrhythmia, decrease in NYHA functional class or need for urgent cardiac intervention, or maternal death. We included an additional composite outcome defined as the occurrence of any major cardiac complication during pregnancy or the postpartum period. Outcomes were assessed from the time of conception until 6 weeks postpartum. Major neonatal outcomes were defined as birth weight <2,500 g, 1- and 5-minute Apgar scores, admission to the neonatal intensive care unit (NICU), and length of stay in the NICU (measured in days). Cardiac outcomes were reported for all pregnancies. Neonatal outcomes were assessed for all gestations exceeding 20 weeks. NICU admission criteria were at the discretion of the neonatologist. The NICU was informed for all births known to have fetal intrauterine growth restriction, and a neonatologist was present at delivery if resuscitation was required. If the birth weight was <2,500 g and the infant had respiratory distress, they were taken to the NICU for observation. Infants <2,000 g were automatically admitted to the NICU for observation. The study was approved by the University of California Los Angeles Institutional Board Review. All investigators vouch for the integrity of the data and accuracy of the analysis.
Categorical variables are summarized as frequency (percentage), whereas continuous variables are summarized as median (interquartile range) for nonnormally distributed variables or as mean (standard deviation) for normally distributed variables. Outcomes were compared by level of each predictor using the exact chi-square test (categorical variables) or the Wilcoxon rank-sum test (continuous variables). Multivariate analyses for predictors of each binary outcome were carried out using the general estimating equations logistic regression model taking into account that observations are not independent. Candidate predictors for the logistic models were selected using p <0.25 in the bivariate screen. Final logistic models were selected using the backward stepwise procedure for variable selection with p <0.25 as the retention criterion. Reported are the adjusted odds ratios (OR), 95% confidence intervals (CI) and p values under the final logistic models. Missing values were imputed using the Markov Chain Monte Carlo method of imputation for the purpose of the multivariate analyses. Analysis was performed using SAS version 9.2 Copyright 2002 to 2008 by SAS Institute Inc. and SPSS Answer Tree version 3.0 Copyright SPSS Inc. 1998 to 2001. A p value <0.05 was considered significant. All tests were 2-sided.
Results
Two hundred women reporting pregnancies were identified among the outpatient services of the Ahmanson/UCLA ACHDC. Of these, 43 women with 61 pregnancies (mean age 27 ± 5 years at delivery, termination, or miscarriage) were identified as having high-risk congenital lesions or co-morbidities as described previously. Underlying maternal congenital cardiac lesions and maternal baseline characteristics are summarized in Tables 1 and 2 . Complete information on cardiac events was available for 92% of pregnancies and neonatal information for 100%.
| Congenital Lesion (n=61 pregnancies) | No. pregnancies (%) |
|---|---|
| Ventricular septal defect, surgically repaired | 2 (3%) |
| Atrioventricular septal defect, surgically repaired | 1 (2%) |
| Aortic stenosis ∗ | 18 (30%) |
| Unoperated | 10 (16%) |
| Transcatheter valvuloplasty | 1 (2%) |
| Aortic valve replacement | 2 (3%) |
| Following Ross procedure | 5 (8%) |
| Subaortic stenosis † | 5 (8%) |
| Coarctation of the aorta, ‡ unoperated | 1 (2%) |
| Mitral stenosis § | 2 (3%) |
| Shone’s complex following supramitral ring resection | 1 (2%) |
| Following mitral valve replacement | 1 (2%) |
| Transposition of the great arteries | 5 (8%) |
| D-Transposition of the great arteries after Mustard or Senning | 3 (5%) |
| D-Transposition of the great arteries following arterial switch | 1 (2%) |
| L-Transposition of the great arteries following mechanical tricuspid valve replacement | 1 (2%) |
| Single ventricle with Fontan physiology | 14 (23%) |
| Truncus arteriosus after mechanical truncal valve replacement | 2 (3%) |
| Ebstein anomaly after tricuspid valve replacement | 1 (2%) |
| Pulmonary atresia following Glenn shunt | 2 (3%) |
| Tetralogy of Fallot, unoperated | 3 (5%) |
| Patent ductus arteriosus with Eisenmenger physiology, unoperated | 1 (2%) |
| Nonrestrictive ventricular septal defect with Eisenmenger physiology, unoperated | 4 (7%) |
∗ Aortic stenosis with valve area <1 cm 2 , mean gradient >40 mm Hg, or peak velocity >4 m/s.
† Subaortic stenosis with mean gradient >30 m Hg or peak gradient >50 mm Hg.
‡ Coarctation of the aorta with peak gradient >40 mm Hg.
§ Mitral stenosis with valve area <1 cm 2 or mean gradient >10 mm Hg.
| Clinical Predictor | No. Pregnancies (%) |
|---|---|
| Oxygen saturation < 90% | 10 (16%) |
| NYHA functional class | |
| Class I or II | 55 (90%) |
| Class III or IV | 6 (10%) |
| Past medical history | |
| Sustained arrhythmia (atrial or ventricular) | 13 (21%) |
| Decompensated heart failure | 0 (0) |
| Systemic morphologic left ventricular ejection fraction < 40% | 1 (2%) |
| Systemic right ventricular ejection fraction < 30% | 2 (3%) |
| Subpulmonic ventricle ejection fraction < 30% | 0 (0) |
| Single ventricle physiology (operated or unoperated) | 16 (26%) |
| Mechanical valve (any) | 10 (16%) |
| Severe left-sided heart obstruction | 19 (31%) |
| Severe pulmonary arterial hypertension ∗ | 10 (16%) |
∗ Severe pulmonary arterial hypertension defined as systolic pulmonary arterial pressure >60 mm Hg.
All pregnancies were singleton. There was no association between maternal age and any maternal cardiac or neonatal outcome (including spontaneous abortion [SAB] or therapeutic abortion [TAB], gestational age at delivery, birth weight, NICU admission or Apgar scores). There were no maternal strokes, deaths, or episodes of cardiac arrest. Five subjects were admitted before delivery for maternal hypoxia or subjective dyspnea. One subject developed atrial fibrillation and one subject developed ventricular tachycardia. One subject had an antepartum pulmonary embolism. One subject was admitted multiple times for a bleeding placenta previa. Two subjects required cerclage placement. A total of 10 subjects received systemic anticoagulation: 8 (1 with 2 pregnancies) for mechanical valves, 1 for a history of pulmonary embolism and 1 for atrial arrhythmias in the context of SV physiology.
Analysis of maternal cardiac outcomes, listed in Table 3 , was performed on all 61 pregnancies, including the 13 pregnancies with SAB or TAB. At least one maternal cardiac event occurred in 31% of pregnancies (n = 19). Eight patients experienced pulmonary edema, 9 had a significant decrease in NYHA functional class, 8 reported nonsustained symptomatic arrhythmias, and 4 reported sustained symptomatic arrhythmias requiring therapy.
| Maternal Cardiac or Neonatal Outcome | No. Pregnancies (%) |
|---|---|
| Aborted pregnancy | |
| Spontaneous | 7 (12%) |
| Induced | 6 (10%) |
| Maternal cardiac complications (at least 1 of the following): | 19 (31%) |
| Pulmonary edema | 8 (13%) |
| Sustained symptomatic arrhythmia requiring therapy | 4 (7%) |
| Significant decline in New York Heart Association functional class | 9 (15%) |
| Need for urgent invasive intervention | 3 (5%) |
| Non-sustained symptomatic arrhythmia | 8 (13%) |
| Neonatal event (at least 1 of the following) | 26 (54%) |
| Intrauterine growth restriction | 4 (29%) |
| Neonatal intensive care unit admission | 19 (40%) |
| Preterm delivery | 9 (19%) |
| Low birth weight ∗ | 20 (42%) |
| Preterm premature rupture of membranes | 1 (6%) |
| 1 minute Apgar score < 7 | 9 (19%) |
| 5 minute Apgar score < 7 | 4 (8%) |
Four of 13 women (31%) with previous arrhythmia suffered subsequent sustained arrhythmias compared with 0 of 47 women without any previous arrhythmias (p = 0.01). Similar findings were found in subjects with symptomatic nonsustained arrhythmias before conception: 5 of 13 (39%) women with previous arrhythmia had further arrhythmias while pregnant compared with 3 of 47 without (6%, p = 0.002). Arrhythmias were managed at the discretion of the cardiologist. Twelve subjects were treated with β blockers, 4 with digoxin, and 2 with flecainide. One subject developed third degree heart block at 16 weeks gestation and required implantation of a permanent pacemaker. One subject developed supraventricular tachycardia requiring treatment with adenosine twice. Another subject with atrial flutter underwent radiofrequency ablation. One subject with multiple recurrences of atrial flutter underwent direct-current cardioversion for one episode, received flecainide for the second episode, and had to be overdrive paced for the third episode.
Reduced functional capacity before conception was associated with continued decrease during pregnancy. Three of 6 women (50%) with NYHA class III/IV deteriorated compared with 6 of 53 (11%) of women with previous NYHA class I or II (p = 0.04). There was no other significant association between other maternal cardiovascular predictors (cyanosis, SV physiology, mechanical valves, severe LHO, reduced EF, or severe PAH) and individual cardiovascular outcomes.
Multivariate logistic analysis demonstrated that a previous history of maternal arrhythmia conferred an increased risk of any major cardiac complication in women with a term pregnancy (adjusted OR 12, 95% CI 2.2 to 65.3, p = 0.004; Figure 1 ). The presence of SV physiology or preconception NYHA class of >II also had a trend toward increased risk of any major cardiac complication although these additional associations did not reach statistical significance (p >0.05).
