Left ventricular cardiac function in fetuses with congenital diaphragmatic hernia and the effect of fetal endoscopic tracheal occlusion




The pre-existing compression of the left ventricle in congenital diaphragmatic hernia (CDH) could be aggravated by the amplified lung growth after fetoscopic endoluminal tracheal occlusion (FETO). Our aim was to document left ventricular (LV) size and function in fetuses with isolated left-sided CDH and to document the effect of FETO on the fetal heart.


We determined cardiac axis, LV diameters, ejection fraction, shortening fraction, mitral E/A index and myocardial performance index (MPI) in 27 fetuses with isolated left-sided CDH, and compared these with values in a reference population (n = 117). In fetuses with severe CDH that subsequently underwent FETO and/or reversal of occlusion, additional measurements were obtained 24 h before and after each fetal intervention. We recorded fetal electrocardiograms non-invasively in six CDH fetuses and compared the duration of the QRS complex with data obtained from 12 controls.


LV end-diastolic diameter was 32% smaller in CDH fetuses than in controls (P < 0.0001) but LV function was comparable. QRS duration was no different between CDH and control fetuses. FETO did not affect cardiac size but reduced the MPI (P = 0.004). Reversal of FETO had no significant effect on cardiac size and function.


CDH fetuses have a smaller left ventricle than do healthy fetuses. There is no overall adverse impact of CDH and FETO on LV cardiac function. Copyright © 2009 ISUOG. Published by John Wiley & Sons, Ltd.


Fetuses with congenital left-sided diaphragmatic hernia (CDH) have variable degrees of visceral herniation into the thorax. The herniated abdominal organs compete for space with the developing lung but may also directly compress the fetal heart, leading to a small left ventricle1. It remains uncertain whether this is a primary or secondary phenomenon2. The latter theory is supported by the observation that the heart generally catches up in growth after surgical correction in the neonatal period, and that cardiac size is normal in long-term survivors of isolated CDH3. In the prenatal period, the left cardiac ventricle is often much smaller than the right one, but the clinical impact of this ‘ventricular smallness’ has been far less documented than the impact of pulmonary hypoplasia. Furthermore, information on prenatal ventricular function, as opposed to ventricular size, is lacking completely.

Overall, up to 30–40% of fetuses with prenatally diagnosed CDH succumb to the consequences of pulmonary hypoplasia despite advanced neonatal care4–6. One strategy to improve outcome in severe cases is in-utero intervention aiming to reverse pulmonary hypoplasia. Fetoscopic endoluminal tracheal occlusion (FETO) has been proposed as an experimental therapy for this purpose; it prevents egress of lung fluid which in turn induces tissue stretch acting as a signal for accelerated lung growth7. Presently, FETO can be achieved through a fetoscopic procedure during which the trachea is occluded with a small balloon (GVB16, Nfocus Neuromedical Inc., Palo Alto, CA, USA). Theoretically, inflation of this balloon causes an acute increase in intrathoracic pressure. Moreover, the subsequent fluid entrapment, which can be observed within the first 24–48 h after the procedure, promptly accelerates fetal lung growth. These factors could cause additional cardiac compression and impair normal cardiac function. FETO is typically reversed at around 34 weeks by fetoscopic balloon retrieval or percutaneous puncture to free the fetal airways and alleviate the need for an ex-utero intrapartum therapy (EXIT) procedure. The balloon deflation and evacuation of the accumulated pulmonary fluid from the lungs could again cause intrathoracic pressure changes with cardiac repercussions.

In the present study, we aimed (1) to document left cardiac ventricular size, function and electrical conduction in fetuses with isolated left-sided CDH during the mid- and early third trimester, and (2) to document the effects of FETO on the fetal heart. Data from a previous study in which we measured cardiac size and function in 117 normal fetuses at 20–36 weeks' gestation8 were used as the normal range.



This prospective study was carried out at the Fetal Medicine Unit of the University Hospitals Leuven between 1 July 2006 and 31 December 2007. The study protocol was approved by the Ethics Committee on Clinical Studies. Our unit serves as a tertiary referral center that offers in-utero therapy for selected cases with isolated severe CDH, i.e. fetuses with a predicted neonatal survival < 20% owing to an observed over expected (O/E) lung-to-head ratio (LHR) < 28% and liver herniation7. All mothers carrying a fetus with left-sided CDH without additional structural anomalies and with a normal karyotype gave written informed consent to participate. For this study, assessment of fetal left ventricular (LV) function was performed at the initial evaluation and at each follow-up visit. However, only one baseline observation was used for each patient. In fetuses that subsequently underwent FETO and/or reversal of occlusion, additional measurements were obtained 24 h before and after each fetal intervention, yielding observations further referred to as ‘pre-FETO’ and ‘post-FETO’, and ‘prereversal’ and ‘postreversal’, respectively.

Fetal heart and lung ultrasound assessment

All ultrasound examinations were performed using a Voluson 730 Expert (GE Medical Systems, Zipf, Austria) by two observers familiar with fetal cardiac ultrasound imaging (T.V.M. and L.G.). The following structural and functional cardiac measurements were obtained with the mother in voluntary suspended respiration and in the absence of fetal breathing movements. In CDH patients, the cardiac axis was measured as the angle between the ventricular septum and a line drawn between the spinal column and the sternum in the four-chamber view9. The ejection fraction (EF) was derived from the end-diastolic and the end-systolic diameters (EDD and ESD, respectively) of the left ventricle using Teicholz's formula10. EDD and ESD were both obtained from M-mode images with the cursor perpendicular to the interventricular septum, just below the tip of the mitral valve leaflets9. The shortening fraction (SF) was calculated as ((EDD − ESD)/EDD) × 100%. The mitral E/A index, i.e. the ratio of the transmitral flow during the early ventricular filling wave to the flow during the atrial contraction, was obtained by positioning the Doppler sample volume at the tip of the mitral valve leaflets in an apical four-chamber view9. Unlike in adults, the E/A index in fetuses is normally < 1 and a higher E/A index is considered to reflect better ventricular diastolic function. The myocardial performance index (MPI), also called the Tei index, was obtained by positioning the Doppler sample volume on the mitral and aortic valve in an apical five-chamber view, as described by Hernandez-Andrade et al.11. The isovolumetric contraction time (ICT), isovolumetric relaxation time (IRT) and ejection time (ET) were determined by using the valve click method. The MPI was calculated as (ICT + IRT)/ET. The MPI is considered to reflect a combination of both systolic and diastolic ventricular function, and a higher MPI corresponds to worse ventricular function8.

For Doppler measurements, the angle of insonation was kept below 10° in all cases. All measurements were obtained in triplicate and averaged. For gestational age-dependent indices (EDD, ESD, E/A index, IRT)9 measurements were expressed as a ratio of the observed value to the value expected in a normal fetus at the same gestational age (O/E)8. Cardiac measurements obtained in CDH fetuses were compared with normal values obtained from 117 control fetuses, which were described in a previous publication8. The cardiac axis in CDH patients was compared with values obtained in 183 normal fetuses published by Comstock12.

Fetal lung size was measured by means of the longest-axis LHR, as described by Metkus et al.13. Again, the effect of gestational age was adjusted for by calculating an O/E ratio using reference data obtained in normal fetuses14.

Fetal heart electric conduction

We obtained non-invasive fetal electrocardiograms (fECGs) in six patients with isolated CDH that had not undergone FETO and in 12 controls matched for gestational age in a 1 : 2 ratio. Fetal electrocardiography was performed using a Monica AN24 Fetal Holter monitor (Monica Healthcare, Nottingham, UK). This device records the fECG over 18 h through four electrodes (Ambu Blue Sensor VL, Ambu, Ballerup, Denmark) placed on the maternal abdomen and a fifth electrode on the back. The feasibility of this technique has been demonstrated recently15. The QRS duration can be calculated offline on the recorded signal from averaged waveforms with software provided by the manufacturer.

Data analysis

We used NCSS 2004 software (NCSS, Kaysville, UT, USA) and Prism for Windows version 5.0 (GraphPad Software, San Diego, CA, USA) for data analysis. The distribution of the data was checked using the D'Agostino Omnibus test. In the control group all variables were normally distributed except ICT. In the CDH group all variables were normally distributed except EDD and LHR, and in the CDH plus FETO group all data were normally distributed except ICT. The data from CDH and control fetuses were compared using the two-sample Student's t-test if normally distributed, or the Mann–Whitney U-test if not. Data obtained before and after intervention were compared using paired t-tests if normally distributed, or Wilcoxon matched pairs tests if not; two-sided P < 0.05 was considered significant. Correlations between the cardiac axis, cardiac diameters, cardiac function and LHR were assessed using Pearson's correlation analysis for normally distributed data or the Spearman's rank correlation analysis for non-normally distributed data.

Waveform data from fECGs were obtained by averaging a mean of 6590 raw fECG complexes per fetus. The fECG QRS duration of the CDH fetuses was compared with the averaged value of its two gestational age-matched controls using the Wilcoxon matched pairs test.


Fetal cardiac size and function

Twenty-seven fetuses with isolated left-sided CDH were included, with a median gestational age at first evaluation of 26 (range, 21–34) weeks. Twenty-four fetuses (89%) showed liver herniation and the median O/E LHR was 27% (range, 18–52%). Following prenatal counseling, five women chose to terminate their pregnancy, 12 underwent FETO and 10 were managed conservatively as they were not eligible for FETO according to the above-mentioned criteria. Survival at discharge from the neonatal intensive care unit was 33% (n = 4) in the small cohort that underwent FETO and 50% (n = 5) in the conservatively managed group.

Cardiac ultrasound data are presented in Table 1 and Figure 1. The LV end-diastolic and end-systolic dimensions were on average 32% and 37% smaller, respectively, in CDH fetuses compared with control fetuses (P < 0.0001), and there was significant displacement of the cardiac axis (P < 0.0001). In contrast, ventricular systolic and diastolic function, as reflected in the EF, SF, MPI and E/A index, were comparable to those in controls. There were no significant correlations between cardiac axis, LV dimensions, EF, SF or E/A index and the LHR (data not shown). However, the MPI was inversely related to the LHR (R = −0.49, P = 0.01). Looking further into this relationship, we found an inverse correlation between the ICT and LHR (P = 0.04), and a trend towards an inverse correlation between the IRT and LHR (P = 0.09). None of the cardiac function indices differed between fetuses that survived and those that did not (all P > 0.05, data not shown). The mean ± SD QRS duration was 60 ± 10 ms and 62 ± 7 ms in the CDH group and the control group, respectively (P = 0.44).

Figure 1.

Left ventricular end-diastolic diameter (EDD) plotted against gestational age in congenital diaphragmatic hernia (CDH) fetuses (○) and controls8 (▪). The regression line (_____) and the estimated 2.5th and 97.5th percentiles (- - - -) for the normal fetuses are plotted. EDD is on average 32% smaller in CDH fetuses than in controls (P < 0.0001).

Table 1. Comparison of left ventricular cardiac parameters between fetuses with congenital diaphragmatic hernia (CDH) and normal controls
ParameterCDH (n = 27)Controls (n = 117)P
  • Data are presented as mean ± SD.

  • *

    Obtained from Comstock (n = 183)12.

  • EDD, end-diastolic diameter; EF, ejection fraction; ESD, end-systolic diameter; MPI, myocardial performance index; O/E, observed over expected; SF, shortening fraction.

O/E EDD (%)69 ± 16101 ± 16< 0.0001
O/E ESD (%)66 ± 19104 ± 21< 0.0001
O/E E/A index (%)99 ± 1499 ± 110.73
MPI0.32 ± 0.050.34 ± 0.050.18
EF (%)63 ± 1060 ± 90.13
SF (%)40 ± 937 ± 70.06
Cardiac axis (°)8 ± 1545 ± 10*< 0.0001

Effect of in-utero intervention

The median gestational age at FETO (n = 12) was 27 (range, 25–29) weeks and the median gestational age at reversal (n = 10) was 33 (range, 32–34) weeks. Two patients who underwent FETO were delivered and underwent emergency EXIT procedure because of spontaneous preterm labor before elective removal of the balloon. This precluded echocardiography before and after reversal. Preintervention and postintervention data are displayed in Table 2. O/E LHR before FETO ranged from 21% to 27.5%. As expected, FETO clearly improved the LHR, which declined somewhat after reversal but remained higher than pre-FETO. FETO reduced the MPI in all but one patient (P = 0.004), but reversal had no significant effect on MPI (Figure 2). Looking further into the relationship between FETO and MPI, we found a shortening of the ICT from 29.6 ± 8.4 ms before FETO to 21.8 ± 7.0 ms after occlusion (P = 0.009), but no change in IRT (35.1 ± 8.5 ms before FETO vs. 35.6 ± 8.4 ms after occlusion, P = 0.77) or ET (178.1 ± 9.0 ms before FETO vs. 179.5 ± 7.2 ms after occlusion, P = 0.64). None of the other indices was affected by FETO or its reversal.

Figure 2.

Individual patient data showing the myocardial performance index before and after fetal endoscopic tracheal occlusion (FETO) (n = 12), and before and after its reversal (n = 10).

Table 2. Left ventricular cardiac parameters before and after fetal endoscopic tracheal occlusion (FETO) and its reversal
ParameterPre-FETO (n = 12)Post-FETO (n = 12)PPrereversal (n = 10)Postreversal (n = 10)P
  • Data are presented as mean ± SD.

  • *

    P = 0.03,

  • P > 0.05 vs. pre-FETO (Wilcoxon signed rank test).

  • EDD, end-diastolic diameter; EF, ejection fraction; ESD, end-systolic diameter; LHR, lung-to-head ratio; MPI, myocardial performance index; O/E, observed over expected; SF, shortening fraction.

O/E EDD (%)64 ± 768 ± 90.2961 ± 1363 ± 100.74
O/E ESD (%)61 ± 1264 ± 140.4557 ± 1859 ± 140.70
O/E E/A (%)99 ± 16100 ± 110.8995 ± 1588 ± 150.16
MPI0.35 ± 0.060.31 ± 0.060.0040.35 ± 0.050.38 ± 0.060.12
EF (%)63 ± 1261 ± 90.6569 ± 1164 ± 50.21
SF (%)40 ± 1038 ± 70.5845 ± 1140 ± 50.21
O/E LHR (%)24.0 ± 2.733.8 ± 7.5< 0.00150.5 ± 13.635.6 ± 5.6*0.01
Cardiac axis (°)11 ± 108 ± 50.247 ± 104 ± 100.95


This study demonstrated that fetuses with isolated left-sided CDH show a displacement in cardiac axis and a smaller left ventricle compared with gestational age-matched controls. However, these anatomical changes were not accompanied by adverse effects on LV function and conduction. In addition, the present study documented no adverse changes in cardiac axis, LV size and cardiac function in fetuses with severe CDH that underwent FETO and subsequent in-utero reversal.

In the 27 fetuses with CDH, LV size before any interventions had been performed was reduced by about 30% in comparison to controls, but this reduction was unrelated to lung size. The observed biometric cardiac changes did not have an adverse impact on the measured parameters of ventricular systolic and diastolic function. The present data are in line with experimental data obtained in rats, showing that ventricular mass is decreased yet myocardial cell size and function are preserved in CDH16. The combination of a smaller total ventricular size but a normal EF in CDH fetuses infers a reduced cardiac output. Indeed, using Doppler interrogation of the aortic and pulmonary valves, Allan et al.1 observed a redistribution of cardiac output toward the right ventricle. Although there was no overall adverse effect on LV function, we documented an inverse relationship between the MPI and the O/E LHR. Thus, LV function appears to improve with increased lung size (albeit remaining within the normal range), probably because blood flow through the lungs is a function of their size17. Smaller lungs are associated with impaired pulmonary blood flow, resulting in augmented shunting through the ductus arteriosus and higher impedance to the LV outflow. Moreover, bigger lungs, and hence a higher pulmonary blood flow, lead to an improved preload of the left ventricle.

We have produced the first antenatal data on ventricular conduction time in CDH fetuses, using commercially available non-invasive fetal Holter monitoring technology. We found that the conduction time was normal in CDH fetuses, which is in line with postnatal observations; dextroposition and axis deviation appear to be the only striking electrocardiographic findings in neonates with CDH18. The QRS complex duration measured with the AN24 device in this series of fetuses was longer than values previously reported using magnetocardiography19 or alternative non-invasive electrocardiographic devices20. Clearly, further work is required to compare different technologies and disease states.

Severe CDH is currently treated by antenatal FETO in several centers around the world. Clinical experience with FETO exceeds more than 200 cases, and we have encountered only one unexplained immediate postoperative fetal demise21. This is similar to what is observed in the natural history of the disease22, yet a theoretical concern of FETO is that the acute increase in intrathoracic pressure, owing to the expanded balloon and/or the rapid increase in lung volume, might compromise cardiac function. However, the data obtained in this study do not support such theoretical concerns. Although the LHR was indeed increased following FETO, LV size and function were not impaired. We actually observed a highly significant decline (i.e. improvement) in the MPI following FETO, coinciding with an increase in LHR; reversal of the occlusion resulted in an opposite trend. We suggest that the increased pulmonary blood flow after FETO (owing to FETO-induced lung growth) reduces shunting through the ductus arteriosus and reduces LV afterload. Besides, an increased pulmonary flow leads to an improved preload of the left ventricle.

We recognize that there were some limitations to this study. First, M-mode imaging (which provides two-dimensional information) was used to determine ventricular size. It cannot be excluded that, in analogy with the lung23, the LV configuration in CDH fetuses is different from that in normal fetuses, i.e. that the shape of the ventricle is altered differently in the longitudinal and the transverse axes. In that case, our measurements might underestimate the true LV volume. Three-dimensional evaluation might clarify this issue but is technically challenging. Second, there is most probably a selection bias in this series. As a referral center for fetal surgery, we see proportionally more severe cases; indeed, there was an 89% liver herniation rate and a median LHR of 27% in the present cohort. Finally, this study was cross-sectional rather than longitudinal. Most measurements were obtained in the second and early third trimester (which is the typical time point for referral for FETO). Cardiac growth continues throughout in-utero life, and the pulmonary venous return becomes an important determinant of LV growth only late in the third trimester24, 25. This could explain why ventricular measurements obtained at a later gestational age are more meaningful as a predictor of neonatal outcome26.

In summary, we have demonstrated that in isolated left CDH the cardiac axis is shifted and the LV size is decreased. However, there is no impairment of LV function. In fetuses undergoing FETO, the intervention does not impair LV growth or function.


The European Commission has supported this work in its Sixth Framework and Marie Curie Fellowship Programme (EuroSTEC; LSHC-CT-2006-037409; MEST CT2005 019707) with fellowships for T.V.M., L.G. and E.D. We thank the consultants of the Department of Obstetrics and Gynecology of the UZ Leuven for their generous funding of a 1-year fellowship for T.V.M. Monica Healthcare supported this study by providing us with a prototype of the AN24 device.