Impact of Pregnancy on Ventricular Strain in Women with Repaired Tetralogy of Fallot

Abstract

Increasing numbers of women with repaired tetralogy of Fallot (rTOF) are reaching reproductive age and seek counseling regarding their cardiovascular risks related to pregnancy. Therefore, the aim of this study was to characterize changes in left ventricular (LV) strain in women with rTOF during pregnancy and in the postpartum period. Seventeen pregnancies in women with rTOF were included (mean age at repair = 3.2 years ± 5.5 years; mean age at delivery = 32.7 ± 4 years). Echocardiograms from three time periods were analyzed; baseline (prior to conception or in the first trimester), third trimester, and 4–6 weeks postpartum. Sixty-five percent of the patients had at least mild pulmonary regurgitation. Eight patients (47%) had undergone at least one pulmonary valve replacement. There were no changes in LV ejection fraction (EF) or circumferential strain across the three time periods. Significant differences were present in longitudinal strain within the three time points (p = 0.01). Postpartum strain decreased in magnitude compared to the third trimester value (− 17.7 ± 4.1 vs. − 21 ± 5, p-value = 0.003) but was not different when compared to baseline strain (− 17.7 ± 4.1 vs. − 19.4 ± 3.4; p-value = 0.15). In conclusion, in women with rTOF, changes in longitudinal strain were observed during pregnancy with a return to baseline after delivery; EF did not change. These findings provide evidence that pregnancy does not adversely impact LV mechanics in the short term in this potentially vulnerable patient population.

Introduction

Left ventricular (LV) dysfunction has been reported in 20% of adults with repaired tetralogy of Fallot (rTOF) and is associated with a risk of early mortality [1]. While the etiology of this dysfunction is not completely understood, it is known that LV dysfunction is associated with increased risk of sudden cardiac death in patients with rTOF [2–4]. Although LV ejection fraction (EF) remains the most commonly utilized measure of ventricular systolic function in repaired congenital heart disease patients, increasingly, other mechanical indices have been introduced. Impaired two dimensional (2D) longitudinal strain, for example, has been associated with risk of sudden cardiac death and life-threatening ventricular arrhythmias in patients with rTOF [5].

In recent years, increasing numbers of women with rTOF are reaching reproductive age and seeking counseling from their providers regarding pregnancy risk [6]. However, there are limited data on the impact of pregnancy on ventricular function and remodeling in women with rTOF [7, 8]. Moreover, ventricular function is usually assessed by echocardiography through systolic and diastolic parameters such as EF and tissue Doppler values. Ventricular strain using 2D speckle-tracking echocardiography has emerged as a sensitive tool to assess subclinical LV dysfunction in patients with normal EF in a number of clinical scenarios [9–13]. In addition, strain has been shown to be a more sensitive parameter of LV systolic function than LVEF, is used to evaluate the efficacy of cardioprotective drugs in heart failure with reduced EF, and is a better predictor of adverse outcome in heart failure with reduced EF than LVEF for LV remodeling and acute heart failure [14–16].

We sought to characterize changes in LV longitudinal and circumferential strain in women with rTOF, from the preconception to the postpartum period, to help inform counseling of these women regarding the impact of pregnancy on ventricular mechanics.

Materials and Methods

Pregnant women with rTOF who were enrolled in the Standardized Outcomes in Reproductive Cardiovascular Care (STORCC) registry from 2011–2016 were included [17]. Data collected at the first visit included baseline demographics, cardiac anatomy, surgical history, comorbid conditions, and medications. New cardiac or obstetric symptoms, changes in clinical status, medications, and cardiac and obstetric outcomes were collected prospectively at each clinic visit, during all admissions, and for up to one year following delivery, as previously described [17]. Women who had incomplete cardiac imaging or did not complete a pregnancy were excluded. All women provided informed consent, and this protocol was approved by the Institutional Review Boards at the Brigham and Women’s Hospital and Boston Children’s Hospital.

Data Collection

Echocardiograms were performed at three time points: baseline (within 18 months prior to conception or in the first trimester), during the third trimester, and 4–6 weeks after delivery. LV volumes were calculated using the 5/6 × area x length method. Additional measurements included pulmonary valve peak and mean gradients as well as qualitative assessment of pulmonary regurgitation (PR) and tricuspid regurgitation (TR) severity [18]. For consistency, all the echocardiograms were reviewed by a single reader (VED).

LV longitudinal strain was assessed by 2D speckle tracking of the apical 4-chamber view (two and three chamber views were not analyzed as these are not uniformly acquired as part of our imaging protocol for TOF). Circumferential strain was evaluated using the parasternal short axis view at the mid-papillary muscle level. These parameters were analyzed offline on standard 30 frame per second ultrasound DICOM images in a standard fashion utilizing commercially available software (2D Cardiac Performance Analysis 1.3.0.147, TOMTEC Imaging Systems, Unterschleissheim, Germany). All the post-processing strain measurements were performed by a single reader (VED).

Statistical Analysis

Continuous variables are presented as a mean and standard deviation. Categorical variables are presented as number (percent) percentage of the study cohort. Means were compared across time using repeated measures analysis of variance. A p-value of < 0.05 was considered significant. Comparisons of longitudinal and circumferential means across the three time points (baseline, third trimester, and postpartum) were performed using repeated measures analysis of variance; additional comparisons of changes from baseline to postpartum were performed using the paired t-test. Bonferroni corrected p-values are reported.

Results

Demographic Data

Sixteen women and 17 pregnancies were analyzed (one subject had two pregnancies included). The mean age at repair was 3.2 years ± 5.5 years and the mean age at delivery was 32.7 ± 4 years. Baseline characteristics of these women are summarized in Table 1. Twenty-nine percent of the women had associated pulmonary atresia, 53% had undergone an initial transannular patch repair, 18% had a right ventricular to pulmonary artery (RV-PA) conduit, and 47% had undergone at least one pulmonary valve replacement. Sixty-five percent of the patients had PR in the baseline echocardiogram, of which 63% was moderate or greater in severity.

Table 1.

Baseline Patient Characteristics

Variable	n = 17
Age at repair, years	3.2 ± 5.5
Age at delivery, years	32.7 ± 4.0
Diagnosis
TOF	12 (71%)
TOF, PA	5 (29%)
Type of initial TOF repair
Transannular patch	9 (53%)
RV-PA conduit	3 (18%)
Other	5 (29%)
QRS duration, msec	136 ± 15
LVEF (baseline)*	61 ± 6
Pulmonary Regurgitation (baseline)*
None	6 (35%)
Mild	4 (24%)
Moderate	1 (6%)
Severe	6 (35%)
Pulmonary valve replacement	8 (47%)

^*Four patients had baseline echocardiograms during the first trimester

LVEF left ventricular ejection fraction, PA pulmonary atresia, RV-PA right ventricle to pulmonary artery conduit, TOF tetralogy of Fallot

Data are represented as means ± standard deviation or numbers (percentages)

Echocardiogram Data

The baseline echocardiograms were performed in the preconception time period in 13 cases and the first trimester in 4 cases (mean gestational age: 10 ± 2.2 weeks). All patients had imaging during the third trimester and 4–6 weeks postpartum.

Table 2 summarizes 2D echocardiographic findings. There was no significant change in LVEF when comparing the baseline echocardiograms with those performed in the third trimester or with those performed postpartum. In addition, there was no significant difference in the degree of PR and TR in the three time points.

Table 2.

Temporal changes in traditional echocardiographic parameters

Variable	Baseline	Third trimester	Postpartum	p-value
LV end-diastolic volume (ml)	130 ± 26	133 ± 47	126 ± 41	0.7
LV end-systolic volume (ml)	51 ± 14	54 ± 18	52 ± 16	0.8
LV ejection fraction (%)	61 ± 6	60 ± 4	58 ± 5	0.1
Stroke volume (ml)	78 ± 18	83 ± 26	77 ± 23	0.5
PV peak gradient (mm Hg)	19 ± 15	25 ± 19	21 ± 17	0.6
PV mean gradient (mm Hg)	10 ± 9	14 ± 10	12 ± 10	0.6

N = 17. LV left ventricle, PV pulmonary valve

Data are represented as means ± standard deviation

The mean longitudinal and circumferential strain measurements are summarized in Fig. 1a statistically significant change in longitudinal strain was present when considering all three periods studied (p-value = 0.01); the magnitude of postpartum strain is significantly reduced compared to the third trimester (− 17.7 ± 4.1 vs − 21 ± 5, p-value = 0.003) but was not different when compared to baseline strain (− 17.7 ± 4.1 vs − 19.4 ± 3.4; p-value = 0.15).There was no significant difference in circumferential strain across the time period of study (p = 0.69) as shown in Fig. 1b.

Fig. 1.

Changes in longitudinal and circumferential strain during pregnancy. a Longitudinal strain at all three time points of the study; which differs over the three time periods. A significant difference is present between third trimester and postpartum (p-value = 0.003), but not between baseline and postpartum values (p-value = 0.48). b Circumferential strain at all three points of study; no significant differences are present among any of the timepoints. *p-values represent comparison between all three points of study measurements. Strain measurements are shown as means ± one standard deviation and compared across time using repeated measures analysis of variance

Clinical Outcomes

Five women (29%) had adverse clinical outcomes. Of these, 1 subject had degeneration of a bioprosthetic pulmonary valve during pregnancy. The initial peak valve gradient was 50 mmHg and increased to 80 mmHg in the 3rd trimester of pregnancy. During the postpartum period, there was a small decline in the gradient; however, it did not return to baseline at 6 weeks postpartum (60 mmHg). Other adverse clinical outcomes included retained placenta, gestational diabetes, postpartum hemorrhage, and cholestasis. The majority of women completed a vaginal delivery (70%). Three cesarean deliveries were performed for obstetric indications (prior cesarean).

All women had a fetal echocardiogram between 18 and 22 weeks gestation. There was one prenatal diagnosis of fetal TOF; otherwise there were no postnatal diagnoses of structural heart disease of the newborn. There were five cases of intrauterine growth restriction (29%). There were no cases of maternal or fetal death.

Discussion

We found that the magnitude of LV longitudinal strain is increased in the third trimester compared to postpartum. This change may be explained by the physiologic changes of pregnancy, which include increased preload with recruitment of the Frank-Starling mechanism and concomitant decrease in systemic vascular resistance and afterload, leading to enhanced longitudinal ventricular function. However, the expected increases in ventricular volumes during pregnancy were not confirmed in this small population. Notably, the increase in strain magnitude during the third trimester involved the longitudinal, but not circumferential, component of strain. The fact that these changes were present in longitudinal but not circumferential strain measures is consistent with prior work that has shown that changes in ventricular function may become manifest in longitudinal fibers before circumferential ones [9, 19]. These findings might be related to the characteristics of LV architecture and the distribution of the circumferential fibers predominantly in the mid wall and of the longitudinal fibers in the subepicardium and subendocardium [20]. These data provide evidence that pregnancy does not have a negative lasting impact on any of the LV functional parameters measured, including EF, longitudinal, or circumferential strain. Where no changes in EF were detected in this cohort, subclinical changes during pregnancy were present as evidenced by changes in longitudinal strain.

To the best of our knowledge, this is the first study to examine myocardial deformation parameters during pregnancy in women with rTOF. Recently, several small studies using cardiac magnetic resonance (CMR) have evaluated ventricular performance in women with rTOF during pregnancy. The data remain limited, and conflicting. Egidy Assenza et al. reported an increase in right ventricular (RV) volume at a faster rate in women with rTOF who completed a pregnancy in comparison to nulliparous women; however, there was no significant change in LV size [8]. Cauldwell et al. did not find a significant change in RV volumes in women with rTOF during pregnancy; however, they did demonstrate increased LV volumes in this cohort [7]. Our observations further suggest that pregnancy in this patient population does not negatively impact LV mechanics in the short term.

There are several limitations to this study, including the variation in the severity of TOF, type of surgical repair, age at surgical repair, and sample size. RV strain was not analyzed as part of this project due to the fact that views of this cardiac chamber (particularly the free wall) were not optimal in all subjects, a usual limitation in echocardiographic images of adult TOF patients, particularly during pregnancy. Due to the limited number of subjects, we were not able to examine any associations between strain and clinical outcomes.

Conclusion

An understanding of the impact of pregnancy on ventricular mechanics in women with underlying congenital heart disease is required for practitioners to be able to provide appropriate counseling. Our study shows that while changes in LV mechanics do occur in pregnant women with rTOF, these alterations are neither adverse nor sustained in our follow-up time. Further studies examining the unique physiologic changes of pregnancy and the effects of these hemodynamic changes on ventricular mechanics in women with congenital heart disease will lead to improved counseling and management of this growing population.