The impact of pulmonary valve replacement on pregnancy outcomes in women with tetralogy of Fallot

Abstract

Background:

Pregnant women with repaired tetralogy of Fallot (rTOF) are at increased risk of adverse outcomes. Although pre-pregnancy pulmonary valve replacement (PVR) may be considered in some women to attenuate risk, published data to support this practice are lacking. Our objective was to explore the impact of pre-pregnancy PVR on pregnancy outcomes in rTOF.

Methods:

Women with rTOF and cardiovascular magnetic resonance imaging (CMR) before and after pregnancy were included if CMR studies were completed within 3 years of pregnancy. Subjects were compared according to presence (+) or absence (−) of PVR at pre-pregnancy CMR. Pregnancy outcomes (cardiovascular, obstetric, and fetal/neonatal) were documented.

Results:

Of the 29 study women identified, 7 were PVR+ and 22 were PVR−. Post-pregnancy, the PVR− group demonstrated interval increase in indexed right ventricular end-diastolic volumes (RVEDVi) (157 ± 28 versus 166 ± 33 ml/m², p = 0.003) and end-systolic volumes (RVESVi) (82 ± 17 versus 89 ± 20 ml/m², p = 0.003) as compared with pre-pregnancy, but no significant change in RV ejection fraction, RV mass, or left ventricular measurements. In the PVR+ group, there were no interval changes in RV measurements pre-versus post pregnancy. Interval rate of change in RVESVi of PVR− exceeded PVR+ women (+3.7 ± 5.0 versus −2.2 ± 5.0 ml/m²/year, p = 0.03). Pregnancy outcomes did not differ in PVR+ versus PVR− women.

Conclusions:

Pregnancy outcomes did not differ according to PVR status in our cohort. While RV volumes remained unchanged in PVR+ women, interval RV dilation was observed in PVR− women. Additional study of a larger population with longer follow-up may further inform clinical practice regarding pre-pregnancy PVR.

1. Introduction

Tetralogy of Fallot (TOF) is the most common form of cyanotic congenital heart disease [1]. Notwithstanding successful repair in childhood, important residual hemodynamic sequelae, including pulmonary valve regurgitation (PR), are prevalent and have been associated with late morbidity and increased mortality [2]. Pulmonary valve replacement (PVR) is often performed to restore valve function, to facilitate RV reverse remodeling, and to mitigate risk of adverse outcomes, although published data to support the latter are relatively scarce [3,4]. Survival after primary repair of TOF (rTOF) has improved considerably over recent decades and, as a result, increasing numbers of women are reaching childbearing age. Even in the absence of cardiac disease, anticipated hemodynamic changes in the context of pregnancy are substantial. These include elevation in heart rate and augmentation of stroke volume, in order to achieve an increase in cardiac output of up to 50%, which typically peaks in the third trimester. Less is known about precise physiologic adaptations to pregnancy in women with rTOF.

While it is recognized that pregnancy is generally uncomplicated in the majority of women with rTOF, these women remain at increased risk of adverse pregnancy outcomes as compared with the general population [5–7], with the highest risk observed in those with RV dysfunction and/or severe PR [8–10]. Studies focused on imaging have demonstrated unfavorable RV remodeling as a result of pregnancy in select mothers with rTOF, particularly when the RV is known to be substantially dilated preconceptually [11–14]. Although some advocate for proactive preconceptual PVR in women with significant PR and concomitant RV dilation in order to attenuate risk of poor pregnancy outcomes, data to support this practice are lacking. Therefore, the objective of this study was to explore the impact of pre-pregnancy PVR on clinical outcomes of pregnancy and on serial CMR measurements. We hypothesized that mothers with pre-pregnancy PVR would have superior maternal/fetal outcomes and better preservation of ventricular dimensions following pregnancy as compared with mothers without PVR.

2. Methods

2.1. Study population

Women with rTOF receiving clinical care during pregnancy in Boston (Boston Children’s Hospital and Brigham and Women’s Hospital: Standardized Outcomes in Cardiovascular Care [STORCC] Initiative) or in Toronto (University Health Network and Mount Sinai Hospital) with sequential pre-pregnancy (baseline) and post-pregnancy (follow-up) CMR examinations between 2001 and 2018 were retrospectively identified. Candidates were included if a baseline and a follow-up CMR study were available for analysis within 3 years of the reference pregnancy, respectively, and if comprehensive care (antepartum, intrapartum, and postpartum follow-up to a minimum of 6 months following delivery) took place at one of the participating centers. Subjects were excluded (1) if the time interval between PVR and pre-pregnancy CMR was <6 months, or (2) if PVR occurred in the interval between pre- and post-pregnancy CMR. Subjects were divided into 2 groups according to presence or absence of PVR (PVR+ versus PVR−) at the time of the pre-pregnancy CMR. To explore whether potential change in RV volumes was reflective of disease progression as opposed to the impact of pregnancy, we studied a comparison group of nulliparous mothers identified from an existing TOF registry using a greedy matching algorithm [15]. The PVR− mothers were matched with the nulliparous control women based on (1) age at first MRI, (2) RVEDVi at first MRI, and (3) time interval between CMR studies. The study was approved by institutional research ethics board at all participating sites and the requirement to obtain informed consent from each patient was waived. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki.

2.2. Clinical data

Demographic characteristics, surgical history, and clinical data pertaining to each subject at the time of the pre-pregnancy CMR were collected from available medical records. Additionally, electrocardiogram, echocardiogram, and cardiopulmonary exercise study results within 6 months of the baseline CMR were recorded. Maternal cardiovascular outcomes were defined as death, stroke, New York Heart Association (NYHA) functional class deterioration (≥2 grades), heart failure requiring intensification of therapy, sustained tachyarrhythmia (atrial or ventricular) requiring therapy, endocarditis, and/or thromboembolism. Obstetric outcomes included premature rupture of membranes, pregnancy-induced hypertension, pre-eclampsia, premature labor (<37 weeks gestation), and postpartum hemorrhage (>500 ml for vaginal delivery and > 1000 ml at Caesarean section). Outcomes in offspring were defined as small for gestational age (birth weight < 10th percentile), intrauterine or neonatal death (within 28 days after birth), and/or neonatal intensive care unit (NICU) admission.

2.3. CMR technique

All CMR studies were performed on a 1.5 Tesla scanner (Magnetom Avanto Fit, Siemens Healthcare, Erlangen, Germany; or Ingenia Ambition, Philips Healthcare, Best, the Netherlands). Retrospectively gated cine steady-state free precession (SSFP) images were reviewed for assessment of ventricular volumes, systolic function, and ventricular mass. Right and left ventricular (LV) volumes were calculated in the short-axis plane (6–8 mm slice thickness and 0–2 mm inter-slice gap). Free breathing, ECG-gated, cine phase contrast flow measurements were obtained at the main pulmonary artery for assessment of PR.

2.4. CMR analysis

Ventricular endocardial and epicardial borders were contoured by a single experienced observer blinded to clinical information according to a standardized protocol for measurement of biventricular end-diastolic and end-systolic volumes, biventricular ejection fraction (EF), and biventricular mass, as previously described [16]. Intra- and interobserver agreement for CMR measurements in our laboratory have been recently published [17]. End-diastolic volume (EDV), end-systolic volume (ESV), and mass measurements were indexed to body surface area using the Mosteller equation. [18] The PR fraction was calculated as the percentage of backward flow over forward flow in the main pulmonary artery.

2.5. Statistical analysis

Demographic and clinical characteristics of subjects were compared according to presence or absence of PVR at the time of baseline CMR. Continuous variables were tested for normality using the Shapiro-Wilk test. Means (with standard deviation [SD]) or medians (with interquartile ranges [IQR]) were reported, as appropriate. For comparisons between groups, an unpaired t-test or Wilcoxon rank sum test was used for continuous variables and the chi-square test or Fisher exact test was applied for categorical variables. For comparisons within groups, a paired t-test or Wilcoxon signed rank test was used. The rate of change in ventricular volumes was calculated as the delta volume $(\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{t}-\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{g}\mathrm{n}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{y}\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}–\mathrm{p}\mathrm{r}\mathrm{e}-\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{g}\mathrm{n}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{y}\mathrm{m}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{u}\mathrm{r}\mathrm{e}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t})/\mathrm{t}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{b}\mathrm{e}\mathrm{t}\mathrm{w}\mathrm{e}\mathrm{e}\mathrm{n}\mathrm{C}\mathrm{M}\mathrm{R}\mathrm{s}\mathrm{t}\mathrm{u}\mathrm{d}\mathrm{i}\mathrm{e}\mathrm{s}$. A mixed effects model was used to compare rates of change over time of ventricular size and systolic function between groups. The correlation between the rate of change in ventricular size and other clinical parameters were assessed by either the Pearson or Spearman correlation coefficient. A two-tailed P-value <0.05 was considered statistically significant. Data were analyzed using JMP Pro 14 statistical software package (SAS Institute Inc. Cary. NC).

3. Results

3.1. Baseline demographics

Of the 29 women who met inclusion criteria, 7 underwent PVR before pregnancy (PVR+) and 22 did not (PVR−). Among the 7 women in the PVR+ group, 5 women had surgical PVR and 2 had percutaneous pulmonary valve implantation. The majority of women had an underlying diagnosis of TOF with pulmonary stenosis (n = 24) and a smaller number had more complex anatomy consisting of pulmonary atresia (n = 4) or absent pulmonary valve (n = 1). The baseline characteristics of the PVR+ and the PVR− groups are shown (Table 1). The number of women with a shunt prior to primary repair and type of TOF repair (specifically transannular, valve sparing or RV-pulmonary artery conduit) did not differ significantly between the groups. All of the women studied were asymptomatic in NYHA functional class I at study entry. Three women in the PVR− group were treated with cardiovascular medications at study entry (diuretic therapy n = 1, beta-blockade n = 1 and angiotensin converting enzyme inhibitor n = 1) as compared with no women in the PVR+ group, although this difference was not statistically significant (p = 0.45).

Table 1.

Baseline clinical characteristics at the time of the first antenatal visit.

	PVR+ n = 7	PVR- n = 22	P value
Age at pregnancy (years)	28.6 ± 4.1	32.2 ± 4.8	0.07
Gravida	2 (1–3)	2 (1–3)	0.54
Parity	1 (0–1)	1 (0–2)	0.48
Pre-pregnancy BSA (m [2])	1.79 ± 0.22	1.74 ± 0.21	0.59
TOF anatomy, n (%)			0.01
Pulmonary stenosis	3 (43)	21 (95)
Pulmonary atresia	3 (43)	1 (5)
Absent pulmonary valve	1 (14)	0 (0)
Prior shunt, n (%)	2 (29)	7 (32)	0.87
Age at primary repair, years	4.0 (0.7–4.0)	5.2 (2.9–5.9)	0.07
Type of primary repair			0.41
Transannular	4 (57)	15 (68)
Non-transannular	0 (0)	4 (18)
RV-PA conduit	3 (13)	3 (14)
Electrocardiogram
QRS duration (msec)	149 ± 30	136 ± 25	0.35
Echocardiography
≥ Moderate TR, n (%)	0 (0)	0 (0)	1.00
Peak RVOT Doppler, mm Hg	20 ± 8	22 ± 11	0.75
RV systolic pressure based on TR jet velocity, mm Hg	41 ± 12	35 ± 6	0.43
Cardiopulmonary exercise test
VO2 peak, ml/min/kg	19.4 ± 4.3	23.8 ± 6.8	0.15
Predicted VO2, %	61 ± 17	78 ± 14	0.12
VO2 AT, ml/min/kg	14.0 ± 1.2	14.7 ± 5.1	0.64
Predicted VO2 AT, %	45 ± 3	49 ± 13	0.38

Data are represented as mean ± SD, median (IQR) or frequency (percent).

ACEI, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; AT, an-aerobic threshold; BSA, body surface area; PA, pulmonary artery; PVR, pulmonary valve replacement; RV, right ventricle; TOF, tetralogy of Fallot; TR, tricuspid regurgitation; VO₂, aerobic capacity

3.2. CMR data

The mean time interval between sequential pre- and post-pregnancy CMR studies was 3.3 ± 1.1 years. In the PVR+ versus PVR− groups, there was no significant differences in time intervals between baseline CMR and delivery, and between delivery and follow-up CMR (2.3 ± 0.8 years versus 1.6 ± 0.7 years, p = 0.09; 1.5 ± 0.8 years versus 1.2 ± 0.7 years, p = 0.46). In the PVR+ group, the median time from PVR to baseline CMR was 1.6 years (0.6–3.8) and from PVR to delivery was 4.5 years (3.3–5.7). Compared with pre-pregnancy, the PVR− group demonstrated interval enlargements post-pregnancy in both in right ventricular (RV) EDVi (from 157 ± 28 to 166 ± 33 ml/m², p = 0.003) and ESVi (from 82 ± 17 to 89 ± 20 ml/m², p = 0.003) (Fig. 1), but no significant changes in RV ejection fraction, RV mass. In the PVR+ group, aside from indexed LV mass (pre-pregnancy 53 ± 10 and post-pregnancy 59 ± 4 g/m2, p = 0.02), we did not observe any other statistically significant changes in CMR measurements between pre- and post-pregnancy studies (Table 2).

Fig. 1.

Interval change in indexed right ventricular end-diastolic and end-systolic volumes before and after pregnancy in women with pulmonary valve replacement (PVR+) versus those without pulmonary valve replacement (PVR−). Each line demonstrates longitudinal change in right ventricular (RV) size for an individual patient. The rate of change in RV end-systolic volume indexed (RVESVi) was found to be significantly greater in PVR− versus PVR+ women while end-diastolic volume indexed (RVEDVi) rate of change was not (panel A). The rates of change in RVEDVi and RVESVi were greatest in the women with the largest RV dimensions entering pregnancy. Specifically, change in women with upper tertile ventricular dimensions (RVEDVi >161 ml/m² and RVESVi >85 ml/m²) had the greatest change in RV size (panel B).

Table 2.

Pre-pregnancy (baseline) and post-pregnancy (follow-up) CMR data for the PVR+ and PVR- women: within group and between group comparisons.

	PVR+			PVR-			PVR+ versus PVR-
	n = 7			n = 22			n = 29
	Pre-pregnancy	Post-pregnancy	P value	Pre-pregnancy	Post-pregnancy	P value	Pre-pregnancy P value	Post-pregnancy P value
RVEDVi, ml/m2	109 ± 27	102 ± 22	0.24	157 ± 28	166 ± 33	0.003	0.002	<0.001
RVESVi, ml/m2	62 ± 19	55 ± 16	0.14	82 ± 17	89 ± 20	0.003	0.04	0.001
RV EF, %	44 ± 6	46 ± 5	0.07	48 ± 3	47 ± 3	0.14	0.049	0.80
RV mass index, g/m2	25 ± 4	28 ± 2	0.18	35 ± 8	36 ± 7	0.81	<0.001	<0.001
RV mass:volume ratio, g/ml	0.20 ± 0.03	0.22 ± 0.05	0.08	0.19 ± 0.04	0.19 ± 0.04	0.50	0.0.63	0.06
PR fraction, %	3±3	3±3	1.00	43 ± 3	39 ± 10	0.68	0.001	0.004
LVEDVi, ml/m2	93 ± 14	83 ± 9	0.24	82 ± 12	81 ± 12	0.87	0.47	0.67
LVESVi, ml/m2	40 ± 7	38 ± 6	0.41	35 ± 7	36 ± 8	0.34	0.18	0.51
LV EF, %	54 ± 4	54 ± 4	0.79	57 ± 4	57 ± 5	0.83	0.07	0.13
LV mass index, g/m2	53 ± 10	59 ± 4	0.02	56 ± 13	55 ± 15	0.61	0.19	0.82
LV mass:volume ratio, g/ml	0.52 ± 0.05	0.56 ± 0.06	0.14	0.58 ± 0.11	0.58 ± 0.14	0.92	0.11	0.90

Data are represented as mean ± SD. Bold numbers reflect P values < 0.05.

While the rate of change in RVESVi in the PVR− group significantly exceeded that of the PVR+ group (+3.7 ± 5.0 versus −2.2 ± 5.0 ml/m²/year, p = 0.03), the rate of change in RVEDVi approached however did not meet statistical significance (+5.5 ± 8.5 versus −2.5 ± 4.9 ml/m²/year, p = 0.07). Among the PVR− women, those in the pre-pregnancy upper tertile of RV size (all with RVEDVi >161 ml/m² and RVESVi >85 ml/m²) had the greatest interval increase in RV volumes as compared with those in the lower tertiles (upper tertile interval change in RVEDVi +12.9 ± 10.1 ml/m²/year versus +2.0 ± 4.9 ml/m²/year in the lower tertiles combined, p = 0.03; upper tertile interval change in RVESVi +7.7 ± 6.2 versus +1.8 ± 3.1 ml/m²/year in the lower tertiles combined, p = 0.04) (Fig. 1). Rate of change in RVEDVi and RVESVi was correlated positively with gravidity (r = 0.55, p = 0.002; r = 0.43, p = 0.02, respectively). There were no significant correlations between rate of RV change and other relevant parameters (including age at primary repair, age at pregnancy, QRS duration on ECG, RV systolic pressure on echocardiography or aerobic capacity on cardiopulmonary study). Finally, although the rate of change in RV volumes appeared to be larger in the pregnant PVR− women as compared with the nulliparous controls the differences were not statistically significant (RVEDVi +5.5 ± 8.0 versus +0.1 ± 3.6 ml/m²/year, p = 0.27, RVESVi +3.7 ± 5.0 versus +1.3 ± 1.7 ml/m²/year, p = 0.25).

3.3. Maternal and fetal outcomes

Maternal cardiovascular, obstetric, and fetal/neonatal outcomes for the PVR+ and PVR− groups are shown (Table 3). There were no maternal, fetal, or neonatal deaths during this study. There were no statistically significant differences between groups in any of the recorded maternal or offspring outcomes. Heart failure was the only adverse maternal cardiovascular outcome observed in pregnancy. Heart failure requiring intensification of therapy and decline in NYHA functional class (non-mutually exclusive events) occurred at a similar frequency in the PVR+ and PVR− groups (n = 1 [14%] versus n = 3 [14%], p = 0.97; and n = 1 [14%] versus n = 2 [9%], p = 0.76, respectively). Gestational age and birthweight were similar between the groups (38.7 ± 1.4 versus 38.4 ± 1.6 weeks, p = 0.62; 2734 ± 430 versus 3079 ± 394 g, p = 0.10). The percentage of infants admitted to the NICU did not differ significantly between the PVR+ and PVR− groups (n = 1 [14%] versus n = 4 [18%], p = 0.40).

Table 3.

Clinical outcomes in PVR+ versus PVR- women.

	PVR+ n = 7	PVR- n = 22	P value
Cardiovasculara
Heart failure requiring intensification of therapy, n (%)	1 (14)	3 (14)	0.97
NYHA deterioration ≥2 grades, n (%)	1 (14)	2 (9)	0.76
Atrial or ventricular arrhythmia, n (%)	0 (0)	0 (0)	1.00
Endocarditis, n (%)	0 (0)	0 (0)	1.00
Thromboembolic event, n (%)	0 (0)	0 (0)	1.00
Maternal death, n (%)	0 (0)	0 (0)	1.00
Obstetric
Premature rupture ofthe membrane, n (%)	0 (0)	2 (9)	0.27
Pregnancy-induced hypertension, n (%)	1 (14)	0 (0)	0.09
Pre-eclampsia, n (%)	0 (0)	0 (0)	1.00
Infection, n (%)	0 (0)	1 (5)	0.44
Bleeding, n (%)	0 (0)	1 (5)	0.42
Premature labor, n (%)	0 (0)	3 (14)	0.15
Gestational age	38.7 ± 1.4	38.4 ± 1.6	0.62
Mode of delivery, n (%)			0.65
Vaginal	5 (71)	13 (62)
Caesarean section	2 (29)	8 (38)
Fetal/neonatal
APGAR score at 1 min	8 (5–9)	8.5 (7–9)	0.17
APGAR score at 5 min	9 (8–9)	9 (9–9)	0.50
Birthweight	2734 ± 430	3079 ± 394	0.10
NICU admission, n (%)	1 (14)	4 (18)	0.40

Data are represented as mean ± SD, median (IQR) or frequency (percent).

NICU, neonatal intensive care unit; NYHA, New York Heart Association Functional class; PVR, pulmonary valve replacement.

^aNon-mutually exclusive events.

4. Discussion

To our knowledge, this is the first study to explore the association between PVR status and pregnancy outcomes in women with rTOF. In keeping with previous studies, cardiovascular, obstetric, and fetal/neonatal outcomes in this cohort of mothers with rTOF were relatively favorable as compared with other forms of moderately complex congenital heart disease [19]. Furthermore, adverse events did not differ according to the presence or absence of pre-pregnancy PVR. A notable observation in this cohort was the difference in CMR measurement change between women who received PVR before pregnancy and those who did not. Women without PVR had a significant increase in RV volumes from pre- to post-pregnancy, whereas those with a PVR in situ prior to pregnancy demonstrated stability in RV size.

Review of the literature suggests that pregnancy may accelerate maladaptive change in the hearts of some women with rTOF. Specifically, those with significant RV dilation at the onset of pregnancy can be at the greatest risk of progressive enlargement following pregnancy. Egidy-Assenza and colleagues reported that 13 women with rTOF developed interval RV enlargement following pregnancy, with the greatest change observed in the patients in the top tertile of RV volumes, specifically those with at least moderate RV enlargement at baseline (RVEDVi >152 ml/m²); conversely those in the lower tertiles did not demonstrate interval change in RV size [12]. Similarly, Cauldwell et al. reported that pregnancy was not associated with important adverse RV remodeling in a cohort of 19 women with rTOF and mild-to-moderate RV dilatation [11].

The baseline CMR characteristics of the women studied in our cohort without PVR were in keeping with moderate-severe RV enlargement prior to pregnancy (mean RVEDVi 157 ± 28 ml/m²), a reflection of more severe enlargement in this as compared with other published cohorts of mothers with rTOF [11,12]. In comparison, we found that women who underwent PVR prior to pregnancy had only borderline enlarged RV volumes with no significant change in dimensions following pregnancy, findings which are in keeping with previous publications describing women with similar RV volumes, albeit in the absence of PVR. Of note, we observed that the greatest rate of change following pregnancy occurred in the PVR− women with the largest pre-pregnancy RV volumes (upper tertile RVEDVi >161 ml/m²and RVESV >85 ml/m²). The rate of change in the upper tertile was striking, and distinct from the lower quartiles, with a delta RVEDVi of 12.9 ml/m²/year and a delta RVESVi of 7.7 ml/m²/year. This finding highlights the presence of a high risk subset of mothers in whom pregnancy may impose a deleterious effect on ventricular dimensions, an observation which may further assist with selection of women who might specifically derive benefit from pre-pregnancy PVR to restore valve competence. It is noteworthy that the rate of change in RV volumes, both RVEDVi and RVESVi, was positively correlated with gravidity, suggesting that there may be a cumulative risk which is imposed on the RV in women with multiple pregnancies.

Contemporary guidelines on the management of the adult with congenital heart disease highlight the importance of serial imaging studies to guide timing for PVR referral in those with rTOF, particularly in the asymptomatic adult, as interval change in ventricular measurements may be considered a marker of cardiac instability, which may herald the need for pulmonary valve intervention [20–23]. One of the salient observations in our study is the statistically significant increase in RVESVi documented in the PVR− group on sequential CMR studies (+3.7 ± 5 ml/m²/year) which was distinct from the stability observed in the PVR+ group, suggesting that anticipated hemodynamic adaptations related to pregnancy may result in maladaptive changes at the ventricular level which can persist even after pregnancy has concluded. It is worthy of mention that in our cohort of PVR− women, post-pregnancy RVEDVi and RVESVi measurements were significantly larger as compared with pre-pregnancy dimensions, and in fact, the mean volumes following pregnancy exceeded contemporary imaging thresholds for PVR referral as a result of interval change in RV size following to pregnancy [20]. In keeping with a defined pre-operative volume threshold which is associated with successful RV reverse remodeling following restoration of pulmonary valve competence as previously established, we speculate that the extent of change in RV size as a result of pregnancy could be associated with suboptimal imaging outcomes following PVR in some mothers with the largest ventricles [24–27].

Our study supports the presence of a “vulnerable ventricle” in women with moderate-severe RV enlargement prior to pregnancy which appears to be uniquely predisposed to an increased magnitude of change as a result of pregnancy. This vulnerability is not apparent in those with mild or moderate RV enlargement at baseline [11,12]. The rate of change in RV volumes in pregnant PVR− women was larger than that observed in the nulliparous PVR− women although not statistically significant. This observation may reflect similarities in disease course between the groups or may be a reflection of small sample size and is therefore worthy of further study in a larger population.

While we advise caution with regard to extrapolation of our findings to support routine pre-pregnancy PVR, particularly as maternal and fetal outcomes are similar in our PVR+ and PVR− cohorts, we suggest that the accelerated change in ventricular volumes associated with pregnancy, which may indeed be irreversible, be incorporated into pre-conceptual counselling. Although our study did not have adequate follow-up to allow us to comment on the incidence of PVR referrals following pregnancy, we hypothesize that numbers of PVR referrals may be augmented in mothers based on post-pregnancy RV dimension thresholds in the PVR− group, coupled with the knowledge that rates of intervention are known to be increased following pregnancy in studies of women with other forms of valvular disease (specifically in the left heart) [28,29].

There are several limitations of this retrospective study which are worthy of mention. While baseline demographic characteristics of the PVR+ and PVR− groups were similar, underlying anatomy was the only feature which varied according to presence of PVR; we propose that the potential relevance of this difference is mitigated by the observation that primary repair type which is a main driver of outcomes was similar between groups. The small sample size and the relatively short post-partum follow-up are other notable limitations. As a result, the number of pregnancy outcomes was modest and we were unable to comment on longer term imaging and cardiovascular outcomes (including irreversibility of interval change in ventricular parameters, increased numbers of referrals for PVR after pregnancy, etc.). Additionally, we were unable to match pre-PVR CMR parameters between PVR+ and PVR− groups. Furthermore, selection bias may exist as CMR ineligible women could not be studied (i.e., those with indwelling devices/ leads were excluded) while those selected for CMR may represent a population of women thought to be an increased risk. Finally, wider generalizability of results may be limited given the tertiary care settings where women received care in Boston and in Toronto. Taken together we acknowledge that more data are needed to make informed recommendations regarding pre-pregnancy PVR in women with rTOF.

5. Conclusions

Clinical outcomes in rTOF mothers and their offspring did not differ according to PVR status. In women without PVR, significant enlargement in RV size was observed following pregnancy, whereas women with pre-pregnancy PVR had no appreciable change in CMR parameters after pregnancy. Future study of a larger population with a longer follow-up may further inform clinical practice regarding indications for pre-pregnancy PVR.