Author + information
- Received May 19, 2003
- Revision received October 24, 2003
- Accepted November 24, 2003
- Published online July 7, 2004.
- Gruschen R Veldtman, MBChB, MRCP*,
- Heidi M Connolly, MD, FACC*,* (, )
- Martha Grogan, MD, FACC*,
- Naser M Ammash, MD, FACC* and
- Carole A Warnes, MD, MRCP, FACC*
- ↵*Reprint requests and correspondence:
Dr. Heidi M. Connolly, Division of Cardiovascular Diseases and Internal Medicine, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905, USA.
Objectives We sought to determine pregnancy outcomes in patients with tetralogy of Fallot (TOF).
Background Pregnancy outcomes in patients with TOF are incompletely defined.
Methods Clinical, hemodynamic, and obstetric data were reviewed for women with TOF and prior pregnancy.
Results Of 72 respondents, 43 (mean age, 26 years) had 112 pregnancies (range, 1 to 5); 82 pregnancies were successful. Eight women had unrepaired TOF at the time of their 20 successful pregnancies. At first assessment (age ≥18 years), six patients had pulmonary hypertension, three had moderate or severe right ventricular (RV) systolic dysfunction, and 13 had severe RV dilation due to pulmonic regurgitation. Sixteen patients had 30 miscarriages (27%) and one term stillbirth. Mean overall birth weight was 3.2 kg (range, 2.1 to 4.2 kg). Unrepaired TOF (p = 0.05) and morphologic pulmonary artery abnormality (p = 0.03) were independently predictive of infant birth weight. Six patients had cardiovascular complications during pregnancy: supraventricular tachycardia in two, heart failure in two, pulmonary embolism in a patient with pulmonary hypertension, and progressive RV dilation in a patient with severe pulmonic regurgitation. Five infants (6%) had congenital anomalies.
Conclusions Patients with TOF have an increased risk of fetal loss, and their offspring are more likely to have congenital anomalies than offspring in the general population. Adverse maternal events, although rare, may be associated with left ventricular dysfunction, severe pulmonary hypertension, and severe pulmonic regurgitation with RV dysfunction.
Tetralogy of Fallot (TOF) is the most common cyanotic congenital heart defect and accounts for 5% to 6% of congenital heart malformations (1). Its hallmark anterior and superior infundibular septal displacement gives rise to the tetrad of ventricular septal defect, aortic override, infundibular obstruction, and right ventricular (RV) hypertrophy (2,3). Intracardiac repair, performed since the 1950s, has permitted survival into the childbearing years and excellent quality of life (4). Long-term complications usually relate to functional competence of the RV outflow tract and its secondary effects on ventricular and atrial myocardial function (5,6). Pulmonary regurgitation or stenosis may result in RV dysfunction and failure, progressive tricuspid valve regurgitation, atrial and ventricular arrhythmias, and sudden cardiac death (7,8). Less common problems include persistent left-to-right shunts and pulmonary hypertension (PHT), ascending aortic aneurysm, and aortic valve regurgitation.
Although successful pregnancy has been documented in small series (9–13), data are incomplete for maternal and fetal outcomes in women with TOF. We reviewed the Mayo Clinic records of 147 female patients with TOF to determine the outcomes of pregnancy.
A total of 147 women (≥18 years old) with a diagnosis of TOF were identified from the Mayo Clinic congenital heart disease database. Medical and surgical records were reviewed for anatomic characteristics before repair and for details of surgical repair and reoperation. Among patients with pregnancies, we determined hemodynamic and clinical status from echocardiography, cardiac catheterization, and clinical reports. Status was determined before and after each pregnancy when available, at first Mayo Clinic assessment as an adult, and at latest clinical assessment. Hemodynamic data considered for analysis were from within two years before the last menstrual period before pregnancy and within two years after delivery.
Patients were contacted by telephone or mail. All patients provided an obstetric history, including details of pregnancy, delivery, and fetal outcome. All newborns with congenital heart disease were seen by a pediatric cardiologist, and medical records were reviewed when appropriate. Diagnosis of TOF was verified for all patients at the time of echocardiography, cardiac catheterization, or surgery. Patients with pulmonary atresia were excluded. The protocol was reviewed and approved by the Mayo Foundation Institutional Review Board.
Mean, standard deviation, median, and range were determined for continuous variables when appropriate; frequencies were determined for nominal and ordinal variables. Histograms were used to assess the distributions of continuous variables and to determine whether the parametric Student ttest or the nonparametric Wilcoxon rank-sum test was more appropriate for comparison between groups. Univariate analyses were used to examine associations between clinical and hemodynamic factors and to determine outcome parameters affecting infant birth weight and the number of spontaneous pregnancy losses. Candidates found to be statistically significant or nearly statistically significant (p < 0.13) were considered as candidate variables in the multivariate models: linear regression for birth weight and binary logistic regression for pregnancy loss. If candidate variables were found to be highly correlated with each other (i.e., r ≥ 0.8), the most significant factor was used in the multivariate regression model.
A total of 147 female patients with TOF were identified. Of these, 17 had died (no deaths occurred peripartum), 21 were lost to follow-up, and 2 were institutionalized for developmental delay. Of the 107 patients invited to participate in the study, 72 (67%) responded and participated in this investigation; 43 had pregnancies.
All patients had anatomically confirmed TOF. The most common associated anomalies were pulmonary artery (PA) anomalies (hypoplastic, absent, or disconnected PA or ductal origin of PA) in 9 patients, coronary artery anomalies in 5, and right aortic arch anomalies in 11 (Table 1).
In 20 patients (47%), TOF was initially palliated with an aortopulmonary shunt; 1 patient had a second shunt procedure. A Blalock-Taussig shunt was used in 14 patients and a Waterston or Potts shunt in six patients. Five of eight women with unrepaired TOF were cyanotic during 12 pregnancies. All patients had eventual repair, during which the ventricular septal defect was closed and the RV outflow tract repaired. Among the 43 patients, 24 (56%) had patch reconstruction of the RV outflow, and 17 of these reconstructions (40%) were transannular; of the remaining 19 patients, 17 had only infundibular resection (Table 2). Sixteen patients had a second operation (Table 3), and five patients had a third operation.
Genitourinary tract abnormalities
Four women had acquired anomalies of the genital tract, including endometriosis (n = 2), fibromyoma (n = 1), and uterine carcinoma (n = 1). Two other patients had congenital anomalies, including uterus didelphys and vaginal agenesis.
Obstetric and neonatal outcome
Fifteen patients who were counseled against pregnancy did not comply with medical advice and had 36 pregnancies, of which seven (19%) terminated spontaneously. Obstetric and neonatal outcomes are summarized in Table 4. Among 43 respondents, 112 pregnancies occurred, and 82 were successful. Twenty uncomplicated pregnancies occurred in eight women with unrepaired TOF. Mean hemoglobin concentration at first presentation was 16.4 g/dl (range, 14.1 to 18.6 g/dl). Obstetric and neonatal outcomes are described in Table 5. Five of the seven infants (71%) who were small for gestational age were born to women with unrepaired TOF.
No multiparous births were reported. The mean maternal age at first pregnancy was 26 years (range, 19 to 35 years). All deliveries occurred at term except for one elective cesarean delivery at 35 weeks for perceived maternal cardiovascular risk.
Sixteen respondents had 30 pregnancy losses, three occurring in two women with unrepaired TOF. All but three pregnancy losses occurred spontaneously during the first trimester. The remaining pregnancy losses included a stillbirth at term in a mother with unrepaired TOF, a therapeutic abortion at 20 weeks for fetal hypoplastic left heart syndrome, and one elective termination during the first trimester. Three patients with repaired TOF had three or more (range, 3 to 6) spontaneous first trimester pregnancy losses. At initial assessment, the first patient had severe RV hypertension (RV systolic pressure, 91 mm Hg) due to RV outflow obstruction with associated moderate-to-severe RV enlargement and dysfunction. The other two patients had moderate elevations in their RV systolic pressure (54 and 50 mm Hg) due to supravalvar pulmonary stenosis; RV systolic function was mildly impaired, and left ventricular (LV) function was normal. No gynecologic data were available about the competence of their genital tracts.
By univariate analysis, heavier birth weight correlated with the absence of morphologic PA abnormalities (r = 0.425), lower RV systolic pressure (r = 0.358), and older age at delivery (r = 0.303). Women without surgical repair had smaller infants than women with prior surgical repair (2.573 ± 0.445 kg vs. 3.258 ± 0.466 kg; p = 0.001). Of the seven infants (8.5%) who were small for their gestational age, five were born to women with unrepaired TOF at the time of pregnancy.
Ten women (23%) had cesarean deliveries; four were performed for obstetric reasons and six for perceived cardiovascular risk to the mother. Obstetric and neonatal outcomes for the group are summarized in Table 6.
Obstetric outcome in relation to maternal hemodynamic data is summarized in Table 7. Moderate and severe RV dilation resulted in lower infant birth weight, but the trend was not statistically significant. Six patients had PHT (mean PA systolic pressure, 68 mm Hg; range, 40 to 122 mm Hg). Among these patients, PHT was related to a previous Potts or Waterston shunt in four, probable chronic pulmonary thromboembolic disease in one, and an absent left PA in one. Mothers with PHT tended to have smaller infants than mothers with normal or only mildly increased PA systolic pressure, but this trend was not statistically significant (Table 7). Of the six patients with PHT, one had four first trimester miscarriages, one had peripartum pulmonary embolism, and one had supraventricular arrhythmia during pregnancy.
The statistically significant univariate factors that corresponded to birth weight (unrepaired TOF, PA abnormality, age at primary correction, and RV systolic pressure) were entered into a multiple variable regression analysis model. Unrepaired TOF (p = 0.05; parameter estimate = −0.36) and morphologic PA abnormality (p = 0.03; parameter estimate = −0.20) were found to be independent predictors of birth weight. The overall model characteristics were as follows: R2= 0.37; parameter estimate = 2.66; p < 0.009. For spontaneous pregnancy losses, we were unable to identify any risk factors on either univariate or multivariate analyses.
Obstetric complications included blood group incompatibility (n = 1), antepartum hemorrhage (n = 2), gestational diabetes mellitus (n = 1), and small bowel perforation (n = 1) during cesarean delivery at another institution. Gestational proteinuric hypertension developed in two patients.
Congenital anomalies in offspring
Congenital anomalies occurred in five infants (6%): one had a muscular ventricular septal defect that closed spontaneously by age 12 years; one had hypoplastic left heart syndrome, and the pregnancy was terminated; one was born with mitral valve prolapse; one had pyloric stenosis; and one had clubbed feet, bilateral strabismus, and a cleft lip and palate. No unifying genetic syndrome was identified. One mother had a documented 22q11 deletion syndrome. Her pregnancy was uneventful, and the fetus did not have congenital anomalies.
Table 8summarizes functional and hemodynamic status before each pregnancy (n = 16) or at first assessment after pregnancy (n = 66) when antepartum hemodynamics were not available (mean interval from pregnancy to first assessment, 16.9 years; range, 0.8 to 38.4 years). Among the 66 pregnancies for which no antepartum hemodynamic data were available, four patients had cardiac reoperation after the pregnancy but before the first assessment.
Six women (7%) with repaired TOF at the time of pregnancy reported cardiovascular complications during pregnancy. Two women had increased frequency of palpitations. In one, documented supraventricular tachycardia was noted, and digoxin therapy controlled the heart rate. This patient had severe PHT with an absent left PA, an RV systolic pressure of 122 mm Hg, and severe RV dilation. No records were available to establish underlying cardiac rhythm in the other patient, but she had severe RV dilation and dysfunction from severe pulmonic regurgitation.
Pulmonary edema occurred one day after delivery in a patient with peripartum LV dysfunction of uncertain cause and gestational proteinuric hypertension. She was treated medically, and her LV function returned to normal. Before pregnancy her RV systolic pressure was 62 mm Hg because of hypoplastic obstructive PAs. In another patient, severe LV dysfunction developed at 38 weeks of gestation; emergent delivery was required. The LV dysfunction was most likely related to peripartum cardiomyopathy because normal systolic LV function was noted antepartum. At latest follow-up, LV ejection fraction remained depressed (45%).
One patient with free pulmonic regurgitation had progressive RV dilation and tricuspid valve regurgitation during pregnancy. She required pulmonary valve replacement 32 months after the delivery. She had an uncomplicated peripartum period. In another patient, pulmonary embolism and circulatory collapse developed after delivery. The patient survived, and at the first assessment after her latest pregnancy, she had severe PHT (RV systolic pressure, 89 mm Hg) suggestive of chronic pulmonary emboli.
Aortic root diameter was not significantly different between women who had pregnancies and women who did not (35 ± 6 mm vs. 35 ± 10 mm; p = 0.21). No aortic complications during pregnancy were reported. Among the eight women without surgical repair, one had a pregnancy-related cardiovascular complication. This patient with pulmonary embolism and circulatory collapse was discussed above. In 3 of the 15 women counseled against pregnancy, cardiovascular complications developed during pregnancy.
Pregnancy places a physiologic load on the cardiovascular system. Plasma volume increases because of conservation of renal salt and water, cardiac output increases up to 50% above the prepregnant level because stroke volume and heart rate are augmented, pulmonary vascular resistance usually falls while PA pressures are maintained despite enhanced flow, and systemic arterial resistance usually decreases (14).
The hemodynamic burden of pregnancy combined with the residual cardiovascular lesions after TOF repair is potentially of concern because of the following: progression of RV dysfunction, atrial and ventricular dysrhythmias, thromboembolic phenomena, maternal or fetal death in the presence of severe PHT, progressive aortic root dilation, intrauterine growth retardation, and endocarditis. Also, offspring of mothers with TOF are more likely to have congenital heart disease, with a reported incidence of approximately 3.1% (15,16).
In our series, patients who had cardiovascular complications during pregnancy also had severe RV dilation or RV dysfunction (or both), RV hypertension due to outflow obstruction, and PHT. Unrepaired TOF and the presence of morphologic PA abnormality (hypoplastic or disconnected PA or ductal origin of PA) were independently predictive of infant birth weight.
Before successful intracardiac repair of TOF was introduced in the 1950s, few patients reached childbearing age, and successful pregnancy was uncommon. Pregnancies were characterized by spontaneous abortions, stillbirths, and premature deliveries. Presbitero et al. (17)demonstrated that the most important risk factor for adverse fetal outcome in cyanotic patients was the degree of cyanosis. These authors suggested that an arterial oxygen saturation >85% and a hemoglobin concentration ≤18 g/dl were more likely to result in live birth, whereas hemoglobin concentrations >20 g/dl were associated with adverse fetal outcome.
Although premature delivery was uncommon in our series (1%), 8.5% of infants were small for their gestational age, six of the seven (86%) being born to women with unrepaired TOF. Our eight patients with unrepaired TOF (mean hemoglobin concentration 16.4, 14.1 to 18.6 g/dl) had smaller infants than the patients with repaired TOF. Recently, in a series of women with cardiovascular disease, Siu et al. (18)reported a 4% incidence of infants who were small for their gestational age. The fetus was at highest risk when the mother had a combination of obstetric risk factors and cardiac risk factors. The incidence among healthy controls was only 2%. Although we did not examine obstetric risk factors, we were able to demonstrate a negative correlation between infant birth weight and maternal unrepaired TOF status, morphologic PA abnormality, higher RV systolic pressure, and younger age at primary surgical repair. Unrepaired TOF and the presence of morphologic PA abnormality (hypoplastic or disconnected PA or ductal origin of PA) were independently associated with infant birth weight. Among infants with low birth weight, ≥50% of their mothers have chronic disease that unfavorably affects maternal-placental blood flow (19). One can speculate that PA abnormalities, particularly in the presence of pulmonic regurgitation, may adversely affect augmentation of maternal cardiac output at rest or during exercise. This may result in depressed placental blood flow and, subsequently, intrauterine growth retardation.
Obstetric and neonatal outcome
Patients in our series had a rate of spontaneous fetal loss of 24%, which is substantially higher than the expected national average of 10% (20). Multiple factors may be implicated, including: 1) the high prevalence of genital tract abnormalities in our series; 2) residual hemodynamic lesions such as RV outflow tract obstruction and PHT; and 3) the known increased incidence of congenital malformations in the offspring of women with congenital heart disease.
In our cohort, 23% of the women had cesarean deliveries, a rate very similar to the national rate of up to 21.8%. All but one of the cesarean deliveries were performed for perceived cardiovascular risk to the mother or fetus.
Congenital anomalies in offspring
In our series, 6% of the infants had congenital anomalies, including congenital heart disease, pyloric stenosis, and multiple anomalies (in one infant). This rate is higher than the prevalence of congenital anomalies in the general population, which is approximately 3% (21). The recurrence rate of congenital heart disease alone in our cohort was similar to previously reported rates of approximately 4% (15).
Six women in our series had adverse cardiovascular events, which included supraventricular arrhythmia, heart failure, pulmonary embolism, and progressive RV dilation. Previously published series have demonstrated little or no risk to the mother with a repaired TOF. Singh et al. (9)reported 40 successful pregnancies in 27 women, with no adverse maternal cardiovascular events. Similarly, Nissenkorn et al. (12)demonstrated no maternal complications in five women who had nine pregnancies, and the authors stressed the importance of alleviating a significant residual hemodynamic burden before pregnancy. The earlier studies all had relatively few patients, and the authors examined limited hemodynamic parameters. Antepartum, five of our six patients with cardiovascular complications had hemodynamically significant lesions, including severe PHT with marked RV dilation, severe pulmonic regurgitation with RV dysfunction, RV hypertension due to hypoplastic PAs, and peripartum LV dysfunction (worsened by gestational hypertension in one patient). These data reemphasize that women with important hemodynamic or structural heart disease during pregnancy are at risk of adverse cardiovascular events during pregnancy or the puerperium.
Although this study is the largest series to date of women with TOF and pregnancies, there are several limitations. The hemodynamic data were collected retrospectively, and the obstetric data were collected directly from the patients. Hemodynamic data before pregnancy were often not available. Hemodynamic data at first assessment were used to estimate likely hemodynamic forces during pregnancy, when there was no intervening operation. We obtained available medical records from referring physicians.
The 22q11 deletion is a recognized risk of congenital defects in offspring. This deletion was recognized in only one mother, but routine screening for this abnormality was not performed in our series. We also did not specifically acquire 22q11 deletion status among infants with congenital anomalies. Thus, the frequency of 22q11 microdeletion in the mothers and the frequency of congenital defects in their offspring remain unknown.
Only 67% of eligible women responded to our questionnaire. Although this low response rate potentially introduced bias to our results, we expected the effect to be small. Patients may not have responded because they had uneventful pregnancies or because their pregnancies resulted in fetal loss.
Adult survival in patients with repaired TOF is now expected. Pregnancy in patients with repaired TOF is possible but carries an increased risk of fetal loss, cardiovascular complications during pregnancy, and congenital heart disease in offspring. Pregnancy complications are related to important maternal hemodynamic disturbances (severe PHT, severe pulmonic regurgitation with RV dysfunction, and LV dysfunction). This reemphasizes the critical importance of a comprehensive prepregnancy evaluation. From the cardiovascular standpoint, vaginal delivery is preferred for most TOF patients. Women with any combination of the above hemodynamic disturbances may benefit from surgical repair, if feasible, before pregnancy. Women with PHT or significant LV dysfunction should be counseled about their cardiovascular risk and advised against pregnancy.
The authors acknowledge the contributions of Douglas J. Kocer, Jayne M. Roth, and Effie Veldtman.
- left ventricle/ventricular
- pulmonary artery
- pulmonary hypertension
- right ventricle/ventricular
- tetralogy of Fallot
- Received May 19, 2003.
- Revision received October 24, 2003.
- Accepted November 24, 2003.
- American College of Cardiology Foundation
- Singh H,
- Bolton P.J,
- Oakley C.M
- Zuber M,
- Gautschi N,
- Oechslin E,
- Widmer V,
- Kiowski W,
- Jenni R
- Perloff J.K,
- Koos B
- Presbitero P,
- Somerville J,
- Stone S,
- Aruta E,
- Spiegelhalter D,
- Rabajoli F
- Siu S.C,
- Colman J.M,
- Sorensen S,
- et al.
- Cunningham F.G,
- MacDonald P.C,
- Gant N.F,
- et al.
- Dastgiri S,
- Stone D.H,
- Le-Ha C,
- Gilmour W.H