Author + information
- Received February 22, 2010
- Revision received August 18, 2010
- Accepted August 26, 2010
- Published online January 18, 2011.
- Melanie Vogel, MD⁎,
- Doff B. McElhinney, MD⁎,
- Louise E. Wilkins-Haug, MD, PhD†,
- Audrey C. Marshall, MD⁎,
- Carol B. Benson, MD‡,
- Amy L. Juraszek, MD⁎,
- Virginia Silva, MSN†,
- James E. Lock, MD⁎,
- Gerald R. Marx, MD⁎ and
- Wayne Tworetzky, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence
: Dr. Wayne Tworetzky, Department of Cardiology, Children's Hospital Boston, 300 Longwood Avenue, Boston, Massachusetts 02115
Objectives The objective of this article is to review anatomic, physiologic, and clinical features of fetuses and neonates with severe mitral regurgitation (MR) in conjunction with aortic stenosis (AS) and left ventricular (LV) and left atrial (LA) dilation and to present preliminary results of pre-natal intervention for this condition.
Background Severe fetal valvar AS with an abnormal mitral valve (MV) and MR can lead to left heart dilation, with consequent compression of the right ventricle (RV); hydrops and low cardiac output are often associated.
Methods This is a retrospective review of fetuses diagnosed with AS, severe MR, and LA dilation (2002 to 2009) and neonates with the same combination of abnormalities (1988 to 2009).
Results Fourteen fetuses and 7 neonates were investigated. Eleven fetuses had severe hydrops; all had polyhydramnios and a structurally abnormal MV, abnormal MV inflow pattern, restrictive/intact atrial septum, retrograde flow in the transverse aortic arch, and compression of the right heart. The mean indexed RV output was 326 ± 160 ml/kg/min, lower than the normal average fetal combined ventricular output of 550 ± 150 ml/kg/min. Ten fetuses underwent pre-natal cardiac intervention: aortic valvuloplasty (n = 8) and/or atrial septal dilation/stenting (n = 5). Seven of these, and 11 overall, were live born. Nine patients died (median age 6 days), and 2 patients are currently alive. All 7 patients diagnosed in the neonatal period died (median age 1 day).
Conclusions Aortic stenosis associated with significant MR in the fetus can cause severe LA and LV enlargement, leading to low cardiac output and hydrops. Despite the potential advantages of early pre-natal diagnosis and both fetal and neonatal interventions, this rare complex of anomalies carries a poor prognosis.
Clinically significant mitral regurgitation (MR) in fetal life is rare. It may occur in isolation or in conjunction with other structural and functional abnormalities, such as aortic stenosis (AS) with global left ventricular (LV) dysfunction and dilation (1–4). Mitral regurgitation can also occur in conditions with high fetal cardiac output, such as anemia or extracardiac arteriovenous malformations, or with arrhythmias (5–8). Mild to moderate MR is most commonly observed in evolving or established hypoplastic left heart syndrome (HLHS) (1–4). In rare instances, AS with an abnormal mitral valve (MV) and MR can lead to left atrial (LA) dilation with consequent MV annular dilation, with further exacerbation of MR. The aim of this study was to review the anatomic, physiologic, and clinical features of fetuses and neonates with the rare but distinct combination of severe MR, AS, and severe LA enlargement who were managed at Children's Hospital Boston and to report our preliminary experience with pre-natal intervention for this condition.
We included all fetuses with valvar AS, severe MR, and severe LA dilation evaluated by echocardiography at Children's Hospital Boston from 2002 to 2009. This is a different cohort than the fetuses with AS and evolving HLHS or with established HLHS and an intact or restrictive atrial septum, which have been the subject of prior reports from our group (1–4). We also evaluated patients that did not undergo fetal echocardiography but were diagnosed in the first week of life with the same combination of abnormalities from 1988 to 2009. The study was performed in conjunction with a protocol that was approved by the Children's Hospital Committee for Clinical Investigations.
Cardiovascular and noncardiovascular ultrasound
All fetal patients underwent at least 1 detailed obstetric sonogram and echocardiogram, during which a full anatomic and Doppler survey was performed. Evaluations included fetal biometry, anatomic measurement of left and right heart structures, and cardiac and extracardiac Doppler measurements.
Fetal biometry consisted of measurement of the biparietal diameter, head circumference, femur length, and abdominal circumference. In fetuses with ascites, the abdominal circumference measurement was taken in the standard fashion, recognizing that this would likely lead to an overestimate of fetal weight. Fetal weight was estimated using the method of Hadlock et al. (9). Gestational age–based z-scores were calculated for fetal weight from equations reported by Doubilet et al. (10) and for head circumference, biparietal diameter, femur length, and abdominal circumference from equations reported by Hadlock et al. (11). Because measurement of abdominal circumference was confounded by the presence of ascites and weight calculated by the Hadlock method was above the 50th percentile for gestational age in almost all cases, the mean (50th percentile) weight for gestational age was used to index right ventricular (RV) output to avoid underestimation of output. Other factors recorded on noncardiovascular ultrasound included the presence of polyhydramnios, fetal hydrops, skin edema, ascites, pericardial or pleural effusion, and fetal sex (10).
Anatomic assessment included measurement of the following left and right heart structures: LA dimensions, MV annulus in diastole, LV length (diastole and systole), LV volume (using the 5/6 area × length method), aortic valve and ascending aorta diameters, tricuspid valve (TV) and pulmonary valve annulus diameter, and RV end-diastolic length. The MV was described as anatomically normal or abnormal, and the presence of accessory chordal attachments and echogenic papillary muscles was recorded. The LV was further characterized as normal or dilated, and the presence or absence of endocardial fibroelastosis was noted. The pulmonary veins were categorized as normal, dilated, or compressed, and the patency and size of the foramen ovale was recorded. All reported z-scores are based on gestational age and were calculated from unpublished normative data collected at Children's Hospital Boston between 2005 and 2007 on 232 normal fetuses.
Cardiac Doppler evaluation included measurement of MV and TV inflow patterns and durations, MV and TV regurgitant jet color Doppler vena contracta width, LV pressure (maximum instantaneous MR jet velocity), and maximum instantaneous AS gradient and color Doppler jet width. The direction of patent foramen ovale flow was described. The RV output was calculated as: the product of pulmonary valve velocity-time integral (m) × heart rate (beats/min) × valve area (cm2).
Extracardiac Doppler measurements included velocities and flow patterns in the middle cerebral artery, umbilical artery, umbilical vein, and ductus venosus.
Because of the small number of fetuses and neonates, data are presented primarily in descriptive fashion. Anatomic and physiologic variables for which internally derived z-scores were available were compared with normal (z = 0) using 1-sample t test. The RV output indexed to calculated weight and to the 50th percentile weight for gestational age (assuming that hydrops would confound estimation of weight) was compared with published normal values of fetal combined ventricular output (429 ± 100 ml/kg/min to 550 ± 150 ml/kg/min) (12–15) and with previously published cardiac output data for fetuses with HLHS (16) using 1-sample t test. Indexed RV output was also compared with unpublished pre-intervention data in 75 fetuses that underwent pre-natal aortic valvuloplasty for AS with evolving HLHS at our center, using the Wilcoxon rank sum test. In all cases, LV output was negligible; therefore, the RV was assumed to provide essentially all cardiac output. No other statistical analysis was performed. Data are presented as frequency (%), median (range), or mean ± SD. For 1-sample t test analysis, only p ≤ 0.01 was considered significant.
Fourteen fetuses with severe MR, AS, and severe LA and LV dilation underwent echocardiography at a median gestational age of 28.6 weeks (range 21.6 to 33.3 weeks) (Table 1). Four of these 14 patients were from the usual referral base of Children's Hospital, and 10 were referred from elsewhere. Eight of these fetuses were included in a prior report dealing with fetal cardiac instability during pre-natal intervention, but none of the data included in this report were presented in the prior report (17). One patient was known to have an abnormal karyotype (mosaic Turners [XY, XO]). Two patients had a significant family history: 1 had a sibling with trisomy 18 and another had a sibling that underwent heart transplant for dilated cardiomyopathy. Thirteen fetuses were male and 1 was female.
In addition to the above fetuses, 7 neonates with a median age of 1.0 day (1.0 to 8.0 days) presented to Children's Hospital between 1988 and 2000 with an enlarged LA, severe MR, and AS. Five were male and 2 female.
Anatomic and physiologic findings
Noncardiovascular Fetal Ultrasound
Median z-scores for fetal biometric measurements were 1.39 (0.13 to 3.58) for weight, 1.80 (−1.77 to 3.28) for biparietal diameter, −0.55 (−2.79 to 1.70) for femur length, 1.45 (−0.82 to 7.79) for abdominal circumference, and 0.04 (−1.71 to 1.64) for head circumference.
Eleven of 14 fetuses had severe hydrops: 9 had skin thickening, 9 had ascites, 7 had a pericardial effusion, and 4 had a pleural effusion. Polyhydramnios, ranging from mild to severe, was present in all 14 cases. Two of the 3 fetuses without hydrops at the time of diagnosis were the youngest in the cohort (21.6 and 23.3 weeks' gestation).
All fetuses had severe MR with AS and marked LA and LV dilation (Fig. 1). The most notable anatomic and physiologic features included a structurally abnormal MV (Figs. 1D and 2), abnormal MV (Fig. 3) and TV inflow patterns, dilated central veins, a restrictive or intact atrial septum that was bulging left to right, and retrograde flow in the transverse aortic arch. Anatomic and Doppler findings are summarized in Table 2.
Left Heart Anatomy and Physiology
The pulmonary veins were compressed in all but 2 fetuses, in which they were dilated. As a result, pulmonary venous Doppler data were only available in 4 fetuses; 1 of these was normal, and 3 were abnormal, with evidence of ventricular systolic reversal. All fetuses had severe LA dilation (Fig. 1, Table 2). On average, the LA to right atrial (RA) lateral dimension ratio was approximately 3:1. The atrial septum was bulging left to right in all patients, and there was no detectable right to left or left to right flow across the foramen ovale in 7. The LV was dilated and empirically larger than the RV in the lateral dimension in all fetuses (Table 2). The LV systolic function and generated pressure varied (Table 2). Left ventricular endocardial fibroelastosis was present in 13 fetuses (Fig. 1C). The MV was anatomically abnormal in all fetuses, with echogenic papillary muscles and a bright, immobile anterior leaflet (Fig. 1D). The MV papillary muscles were attached more basally than normal, with shortened immobile chordae and the appearance of a mitral arcade (18) (Fig. 2). The MV annulus was dilated (z-score >2) in 9 fetuses. The MR was severe in all fetuses, with a median MR jet Doppler vena contracta width of 3.6 mm (Fig. 1B, Table 2). All fetuses had an abnormal MV Doppler inflow pattern, and the inflow time was abnormally short in 1 (Fig. 3, Table 1). In all 11 fetuses with partial E-A wave fusion on the MV inflow signal, there was a dominant E wave. All fetuses had demonstrable flow across the aortic valve, with a median maximum instantaneous AS gradient of 17 mm Hg. Four fetuses had antegrade flow, and 10 had retrograde flow in the ascending aorta; all had retrograde flow in the transverse aortic arch and right to left flow across the ductus arteriosus. One fetus had mild aortic regurgitation.
Right Heart Anatomy and Physiology
In all fetuses, the RA and RV appeared distorted and compressed by the dilated LA and LV (Figs. 1A and 1B). The TV annulus diameter was within the normal range in 12 fetuses and just above normal in 2 (Table 2). The TV inflow Doppler was abnormal in 13 fetuses: partially fused in 7 and monophasic (fused) in 6. Among the 7 patients with partially fused TV inflow signal, the E-wave was dominant (with an E/A ratio >2) in 4, and the A-wave was dominant in 3 (E/A ratio 0.58 to 0.78). The TV inflow time was lower than normal, with a median z-score of −1.9 (−3.1 to 0.4). Four fetuses had moderate (n = 3) or severe (n = 1) TV regurgitation.
The average RV output indexed to estimated fetal weight was 309 ± 96 ml/kg/min (148 to 555 ml/kg/min). If all fetuses were assumed to be at the 50th percentile of weight for gestational age, the indexed RV output was 326 ± 160 ml/kg/min (127 to 617 ml/kg/min). Both of these values are significantly lower than published normal values for combined ventricular output, which range from 429 ± 100 to 550 ± 150 ml/kg/min (both p < 0.001), and than the cardiac output of 382 ± 77 ml/kg/min reported by Szwast et al. (16) for fetuses with HLHS (p = 0.014 for calculated weight) (12–16). They were also significantly lower than pre-intervention RV output in 75 fetuses with AS that underwent pre-natal aortic valvuloplasty for evolving HLHS at our center (436 ± 142 ml/kg/min; p = 0.002 for calculated weight).
In 11 patients, Doppler flow patterns in the umbilical artery, umbilical vein, ductus venosus, and/or middle cerebral artery were abnormal. One patient had reversed end-diastolic flow in the umbilical artery, 6 had notching of the umbilical vein, 5 had flow reversal in the ductus venosus (2 not measured), and 6 had an abnormally low middle cerebral artery pulsatility index (3 not measured) (19–21). Doppler data were not obtained for flow in the inferior vena cava.
All 7 patients diagnosed in the newborn period had severe MR, AS, and severe LA enlargement. The median aortic valve and ascending aortic z-scores were −4.1 (−4.4 to 0.8) and −2.0 (−4.3 to 2.3), respectively. All had severe LV dysfunction, variable amounts of endocardial fibroelastosis, and right heart compression.
Fetuses and Fetal Intervention
Ten of 14 fetuses underwent technically successful pre-natal cardiac intervention (Table 1), including aortic valvuloplasty in 8 and atrial balloon dilation in 5 (1 with atrial septal stent placement). Seven of these 10, and 11 overall, were live born, all before 37.8 weeks' gestational age. The parents of 1 patient elected termination of pregnancy, and 2 had a fetal demise the day after fetal intervention. Nine of the 11 live-born patients died at a median age of 6 days (1 to 94 days), including 4 on the first day of life. The median weight at birth or fetal demise was 1,900 g (979 to 4,000 g). Two patients are currently alive at 3 years and 3 months, respectively; both had a successful fetal intervention. None of the 3 patients without hydrops at the time of diagnosis survived beyond the neonatal period.
All 7 patients diagnosed in the neonatal period died soon after birth, at a median age of 1 day (1 to 86 days). The median birth weight was 2,550 g (1,500 to 3,700 g). All 7 neonates underwent cardiac catheterization with the intention to perform a balloon aortic valvuloplasty. Five patients underwent balloon dilation of the aortic valve, with additional balloon atrial septal dilation in 1 patient. Two others underwent initial balloon atrial septal dilation but deteriorated during the catheter procedure before aortic valve dilation could be performed.
This article reports the findings of a cohort of fetuses and neonates with the unique constellation of severe MR, AS, restrictive or intact atrial septum, and severely dilated LA, usually in association with endocardial fibroelastosis of the LV and MV tension apparatus. These cardiac abnormalities were associated with hydrops in 11 fetuses. Fetuses with these findings have been described previously but only in isolated reports (22–25). We contend that this is a unique complex of defects that should be thought of as a distinct clinical entity, which may be on the severe end of the mitral arcade spectrum (18). The primary anatomic anomaly in these fetuses is most likely an abnormal MV resulting in MR, AS leading to increased wall stress and dilation, primary endocardial fibroelastosis, or abnormalities of both aortic and mitral valves. The prognosis of this complex is very poor, with survival beyond the neonatal period in only 2 of the 21 patients. The consistent anatomic and physiologic features of these fetuses may allow insight into the pathophysiology of cardiac hydrops (26,27) in this setting and provide a rationale for pre-natal cardiac intervention for this dire condition.
We propose that the underlying cause of hydrops in fetuses with this constellation of findings is elevated systemic venous pressure, which is a consequence of right heart compression by the dilated LA and LV, which in turn is due to AS and severe MR. The restrictive or intact atrial septum contributes further to left heart dilation and increased LA pressure because there is no natural avenue for decompression of the left heart. The resulting volume load on the LV leads to MV annular dilation, which further exacerbates the MR. The dilated LA and LV can result in mechanical compression of the RA and RV, which might impede both the filling and emptying of the right heart. This may result in elevated systemic venous pressures, as indicated by flow reversal in the ductus venosus and notching on the umbilical venous flow signal, which may lead to hydrops, ascites, and skin edema given the fetus's low albumin and oncotic pressures (28). High central venous pressure also likely results in increased pressure in the thoracic duct, which may impair lymphatic drainage. In addition, both diastolic and systolic RV function appear to be impaired: diastolic dysfunction is suggested by an abnormal Doppler TV inflow signal (either partially or completely fused) and abnormal flow in the ductus venosus and umbilical vein, whereas systolic dysfunction is evidenced by low RV output, which may also be a function of reduced filling. Three fetuses in this series did not have hydrops, but there were no obvious cardiac functional differences between these fetuses and those with hydrops. Because 2 of the 3 fetuses without hydrops were the youngest in the cohort, it is possible that hydrops had simply not developed yet in the same hemodynamic context as the older fetuses with hydrops.
Although more than half of the fetal patients in this series underwent pre-natal intervention, the overall survival was still poor. With this complex cardiac anomaly, there are theoretic advantages to balloon dilation of both the aortic valve and the atrial septum, both of which we have reported in patients with evolving or established HLHS (those series did not include the patients in this report) (1–4). Given that severe MR is the most burdensome aberration, procedures that promote aortic antegrade flow, such as aortic valve dilation (1), might improve forward flow, reduce LV afterload, and therefore decrease the absolute and relative amount of MV regurgitant flow. In addition to relieving LV outflow obstruction, creation or enlargement of an interatrial communication has theoretic benefits (2,3). If an adequately sized hole can be created, this will allow LA decompression, which, in conjunction with improved antegrade flow through the LV, may ameliorate LA hypertension and enlargement. This will consequently relieve RA and RV compression and possibly reduce distortion or compression of other thoracic structures such as the lungs, airways, or esophagus. In addition, the left to right atrial flow might improve RV preload and consequently RV output. The LA decompression in utero may also have implications for post-natal hemodynamics, insofar as LA and pulmonary venous hypertension owing to MR, in conjunction with pulmonary venous compression by the dilated LA, may impede pulmonary blood flow in the newborn and contribute to impaired oxygenation, pulmonary hypertension, and pulmonary edema.
Although 4 of the 5 fetuses in which we performed atrial septoplasty did not undergo aortic valve dilation, atrial septoplasty alone is unlikely to be effective in these fetuses for the reasons discussed previously. Unfortunately, both aortic valvuloplasty and atrial septal opening procedures were performed successfully in only 1 of the fetal patients. In the others, we did not appreciate the potential importance of performing both procedures (early in our experience), had procedural adverse events preventing us from completing both procedures (17), or were technically unable to perform one or the other. Atrial septoplasty can be very challenging in these fetuses, given that the RA is severely compressed and can be difficult to enter.
The structure and orientation of the foramen ovale and the foramen flap valve allow blood to flow more readily from right to left, such that blood flow in most right heart lesions is adequately redirected to the left heart. In contrast, in fetuses with left-sided cardiac disease, impaired or absent LV filling may cause LA hypertension and consequent restriction or closure of the foramen ovale. Unlike HLHS, in which atrial septal restriction can lead to pulmonary venous dilation, severe LA enlargement caused compression of the pulmonary veins in the majority of fetuses in this series (12 of 14). Regardless of the appearance of the pulmonary veins, severe LA hypertension and inadequate LA egress in fetuses with this lesion, as in those with HLHS, may critically limit pulmonary blood flow in the perinatal period with lethal consequences. As such, immediate post-natal decompression of the LA should be anticipated in fetuses that survive to birth. Even with acute neonatal intervention and atrial septal opening, survival in our experience has been poor.
Aortic stenosis associated with significant MR in the fetus can cause severe LA enlargement, ultimately leading to low cardiac output and hydrops, with poor prognosis. The anatomic and physiologic findings presented in this report support a proposed etiology for fetal hydrops and a rationale for pre-natal cardiac intervention. Despite the potential advantages of early pre-natal diagnosis and both fetal and neonatal cardiac interventions, this rare complex of anomalies carries a poor prognosis. These findings will help inform parental counseling and may prompt innovative approaches to therapy, including fetal cardiac intervention.
This study was supported by contributions from the Kenrose Kitchen Table Foundation, Gig Harbor, Washington. Thomas P. Graham, Jr., MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- aortic stenosis
- hypoplastic left heart syndrome
- left atrial/atrium
- left ventricular/ventricle
- mitral regurgitation
- mitral valve
- right atrial/atrium
- right ventricular/ventricle
- tricuspid valve
- Received February 22, 2010.
- Revision received August 18, 2010.
- Accepted August 26, 2010.
- American College of Cardiology Foundation
- McElhinney D.B.,
- Marshall A.C.,
- Wilkins-Haug L.E.,
- et al.
- Marshall A.C.,
- Van der Velde M.E.,
- Tworetzky W.,
- et al.
- Tworetzky W.,
- Wilkins-Haug L.,
- Jennings R.W.,
- et al.
- Pedra S.R.,
- Smallhorn J.F.,
- Ryan G.,
- et al.
- Doubilet P.M.,
- Benson C.B.,
- Nadel A.S.,
- Ringer S.A.
- Mielke G.,
- Benda N.
- Layman T.E.,
- Edwards J.E.
- Nomura R.M.,
- Brizot Mde L.,
- Liao A.W.,
- Andrade J.Q.,
- Zugaib M.
- Rudolph A.M.