Introduction

Hypoplasia of one or several cardiovascular structures is a common feature of many cardiac malformations detected during fetal life. Based on hypoplasia location and severity, this finding may bear important and life-long consequences for postnatal treatment and prognosis. Materno-fetal hyperoxygenation (HO) in the third trimester results in fetal pulmonary vasodilation [8] and increased venous return to the fetal heart. Similar to findings in chicks, in which increased ventricular preload is compensated by myocyte proliferation in hypoplastic left ventricles [1], the resulting increase in cardiac preload from pulmonary vasodilation might, with time, also improve hypoplastic fetal cardiovascular dimensions in humans. The purpose of this report is to describe the impact of chronic intermittent materno-fetal HO on the dimensions of hypoplastic cardiovascular structures in 15 human fetuses with a variety of cardiac malformations.

Patients and Methods

The studies were performed in 15 pregnant women and their fetuses between 33 weeks + 5 days and 38 weeks + 0 days of gestation; all fetuses had normal karyotypes. Thirteen of the 15 fetuses had ≥1 hypoplastic left heart structures; of these, 1 fetus had a severely hypoplastic tricuspid valve and right ventricle associated with anatomically corrected transposition of the great arteries, and another had markedly hypoplastic pulmonary vessels and lung blood flow associated with tetralogy of Fallot (Table 1).

Table 1 Clinical data, baseline values, Doppler findings, and dimensional changes of selected fetal cardiovascular structures during the course of materno-fetal HO in 15 late-gestation fetuses with cardiac malformationsa
Full size table

Informed Consent and Treatment Protocol

Based on fetal echocardiographic findings, the 15 expectant mothers were offered chronic intermittent materno-fetal HO during the final weeks of gestation. The individual procedures were performed after receiving approval by the heads of the local committee on human research and maternal informed consent.

On the first day, the women underwent a detailled multimodal prenatal sonogram. Then they were requested to wear a rebreathing oxygen mask. The oxygen flow was adjusted to a maternal arterial \( P_{{{\text{O}}_{2} }} \) between 250 and 300 mm Hg by ultrasound-guided puncture of the maternal femoral artery using a 26-gauge needle.

On the first day, oxygen was provided for approximately 6 hours. After 1 to 2 hours of materno-fetal HO, the response of the pulmonary vascular circulation and the loading effect of increased venous return to both ventricles on fetal hemodynamics were assessed by detailled sonography as previously outlined.

From the second day onward, for a total between 8 and 33 days (mean 15.5 and median 14), the women underwent 3 × 4 h HO/d until delivery. The timing and length of materno-fetal HO were determined by gestational age at referral, response to therapy, and obstetric considerations. No attempts were made to prolong gestation in cases for whom spontaneous delivery ensued.

The fetuses were monitored by fetal echocardiography and cardiotocography and the expectant mothers by daily clinical examination and interrogation. After delivery, the mothers were requested to undergo chest X-ray and pulmonary function test to monitor for maternal pulmonary morbidity from the experimental treatment approach.

Measurements

Multimodal fetal echocardiography was performed using a high-end ultrasound system (Acuson Sequoia 512; Acuson-A-Siemens-Company, Erlangen, Germany). Measurements of ventricular area as well as atrioventricular valve diameter were taken in the four-chamber plane at end-diastole. Semilunar valve diameter was measured during ventricular ejection. In addition, the largest diameter of great vessel regions of interest were measured during early systole.

Fetal cardiovascular flow regions of interest were assessed by color and pulsed Doppler echocardiography at insonation angles <20°. Before and during materno-fetal HO on the first treatment day, fetal pulmonary venous blood flow mean velocity as well as quantitative changes in color Doppler signal of the lung vessels were compared from still and motion files, respectively, to document pulmonary vasodilation (Table 2).

Table 2 The effect of materno-fetal HO on fetal left heart filling and emptying velocitiesa
Full size table

Measurements of fetal cardiovascular structures of interest were taken at four time points (Table 1). The first and second sets of measurements were taken on the first treatment day right before and during the first hours of materno-fetal HO to document whether oxygen-induced fetal pulmonary vasodilation does in fact load the fetal heart. The third set of measurements represents fetal cardiovascular dimensions taken during the last 2 days of materno-fetal HO. The fourth set of measurements was taken by pediatric cardiologists blinded to the prenatal measurements after delivery during the first 2 days of life.

Data Presentation and Analysis

To compare changes in cardiovascular dimensions of the hyperoxygenated fetuses with normal ones, the data were expressed as gestational age–related Z-scores based on data provided by Schneider et al. [9] and Pasquini et al. [7] employing a Web site calculator available at http://www.parameterz.com. In addition, for enhanced visual demonstration, selected cardiovascular dimensions of the 13 fetuses with hypoplastic left heart structures were plotted on fetal cardiac growth curves for normals provided by Schneider et al. [9] and Matsui et al. [5].

Statistical analysis on fetal cardiac dimensions, atrioventricular and semilunar valve flows, and central pulmonary vein mean flow velocities during loading of the fetal heart was performed using paired Student t test (Table 2). In addition, the same test was performed on semiquantitative data derived from color Doppler motion files of fetal lung flow before and during materno-fetal HO, and p ≤ 0.05 was considered statistically significant. The study addressed the following questions:

  1. 1.

    Is chronic intermittent materno-fetal HO—by increasing fetal pulmonary venous return and, hence, cardiac preload—associated with an increase in fetal hypoplastic cardiovascular dimensions?

  2. 2.

    Are there any adverse maternal events during and after materno-fetal HO?

  3. 3.

    Are there any adverse fetal or neonatal complications?

Results

Within the first 2 h of initiating treatment, materno-fetal HO resulted in statistically significant increases of pulmonary venous mean flow velocity (p = 0.0007) and mitral valve late diastolic (“a-wave”) flow velocity (p = 0.05) (Table 1 and Fig. 1). During this left heart loading phase, significant increases in the dimensions of the mitral valve (p = 0.02), ascending aorta (p = 0.03), and aortic isthmus (p = 0.03) were observed.

Fig. 1
figure 1

Typical effect of materno-fetal HO on fetal lung blood flow as demonstrated by color and pulsed Doppler echocardiography in a fetus with hypoplasia of left heart cardiovascular structures. Within minutes of initiating treatment, materno-fetal HO results in marked color Doppler changes and increases in pulmonary venous flow velocity (Color and pulsed Doppler frames before (top panel) and during (bottom panel) materno-fetal HO). (Color figure online)

Full size image

The following changes in fetal cardiac dimensions and the following changes in Z-scores were observed in the fetuses during the course of chronic intermittent materno-fetal HO between 8 and 33 days (mean 15.5 and median 14) (Table 1 and Figs. 2–4).

Fig. 2
figure 2

Dimensions of left ventricular area and mitral valve diameter in the four-chamber plane before and at the end of materno-fetal HO in 13 fetuses with cardiovascular hypoplasia of ≥1 left heart structures. Black solid lines indicate mean values ± 2 z-based on data provided by Schneider et al. [9]. Black dots to the right of the figure are provided to facilitate the comprehension of even small dimensional changes on flow

Full size image

Ventricular Area

Changes in areas and Z-scores of left ventricles (n = 13), as well as a left-sided right ventricle (n = 1), were available in 14 of 15 fetuses. In 7 of these fetuses, marked area and Z-score increases for left ventricular area were observed in the cardiac four-chamber view (Fig. 2) during the course of therapy. In five fetuses, the Z-scores for left ventricular area developed in parallel to the growth curve of normals. In one fetus, the Z-score for the left ventricular area, and in one fetus, the Z-score for the right ventricular area decreased progressively during the course of therapy.

Atrioventricular Valve Diameter

Changes in diameters and Z-scores of the left-sided mitral or tricuspid valves were available in 14 of 15 fetuses. In 9 of these fetuses, marked increases of dimension and Z-scores of mitral valve (n = 8) or tricuspid valve (n = 1) diameter were observed during the treatment period (Fig. 2). In two fetuses, the mitral valve diameter developed in parallel to the growth curve of normals; in three fetuses, the mitral valve diameter remained unchanged, hence decreasing progressively below the lower limit of normal during the course of therapy.

Semilunar Valve Diameter

Changes in diameters and Z-scores of aortic (n = 13) or pulmonary valves (n = 1) were available in 14 of 15 fetuses. In 10 of these fetuses, marked diameter and Z-score increases were observed (Fig. 3) during the course of therapy. In two fetuses, the aortic valve diameter developed in parallel to the growth curve of normals. In two fetuses, the dimensions of the aortic valve remained unchanged such that their Z-scores decreased progressively during the course of therapy.

Fig. 3
figure 3

Diameters of the aortic valve and ascending aorta in the four-chamber plane before and at the end of materno-fetal HO in fetuses with cardiovascular hypoplasia of ≥1 left heart structures. Black solid lines indicate mean values ±2 z-based on data provided by Schneider et al. [9]. Black dots to the right of the figure are provided to facilitate the comprehension of even small dimensional changes on flow

Full size image

Great Artery Diameter

Changes in diameters and Z-scores of the ascending aorta (n = 14) or main pulmonary artery (n = 1) were available from all fetuses. In 10 of these fetuses, marked diameter and Z-score increases of the ascending aorta (n = 19) or main pulmonary artery (n = 1) were observed (Fig. 3) during the course of therapy. In three fetuses, the ascending aorta diameter developed parallel to the growth curve of normals. In two fetuses, the diameter of the ascending aorta remained unchanged such that their Z-scores decreased progressively during the course of therapy.

Aortic Isthmus and Right Pulmonary Artery Diameter

Changes in diameters and Z-scores of the aortic isthmus diameter were available from 13 of 14 fetuses with left heart malformations and of the right pulmonary artery from 1 fetus with Fallot tetralogy and poor pulmonary blood flow. In 11 of the fetuses with left heart malformations, marked diameter and Z-score increases of the aortic isthmus were observed (Fig. 4) during the course of therapy. In two fetuses, the diameter of the isthmus remained unchanged or decreased slightly such that their Z-scores decreased progressively during the course of therapy. In the fetus with Fallot tetralogy, marked diameter and Z-score increases of the right pulmonary artery were observed.

Fig. 4
figure 4

Diameters and Z-scores of the aortic isthmus in the four-chamber plane before and at the end of materno-fetal HO in fetuses with cardiovascular hypoplasia of ≥1 left heart structures. Z-scores for normals are based on data provided by Matsui et al. [5]. Black dots to the right of the figure are provided to facilitate the comprehension of even small dimensional changes on flow

Full size image

In fetuses in whom individual hypoplastic cardiovascular structures exhibited little or no increase in dimension during the course of chronic intermittent materno-fetal HO, associated cardiac malformations—such as left superior caval vein-to-coronary sinus, unrestrictive ventricular septal defects, obstructed or stenotic atrioventricular valves, ventricular outflow obstruction, or semilunar valve stenosis—were observed (Table 1).

Maternal and Fetal Safety

Adverse maternal, fetal, or neonatal events during and after chronic intermittent materno-fetal HO were not observed during or after any of the treatment attempts. More specifically, neither detrimental changes to uterine and fetoplacental blood flows, fetal cardiac failure, nor preterm constriction of the ductus arteriosus or postnatal pulmonary hypertensive events occured in any of the fetuses. After HO, the mothers exhibited normal chest X-rays and pulmonary function tests.

Discussion

This analysis found that chronic intermittent materno-fetal HO administered in late gestation may be associated with improvements of hypoplastic cardiovascular dimensions in fetuses with a variety of cardiac malformations. This finding may benefit postnatal treatment options and prognosis in suitable cases.

Reliable increases in hypoplastic cardiovascular dimensions during the course of chronic intermittent materno-fetal HO were observed in most fetuses with small ventricles that filled and emptied across normally built valves. These increases in the dimensions of cardiac structures were usually accompanied by increases in great vessel dimensions and even the aortic arch (Fig. 5). In contrast, the effect of materno-fetal HO was ameliorated or neutralized by the presence of unrestictive ventricular septal defects as well as obstructions to ventricular filling or emptying (e.g., left superior caval vein to the coronary sinus with absence of the left brachiocephalic vein, mitral valve stenosis, ventricular outflow tract obstruction, aortic valve stenosis).

Fig. 5
figure 5

Fetus with mild left heart hypoplasia and severe tubular hypoplasia of the transverse and distal aortic arch (top) at 38 + 0 weeks of gestation (Table 1, case no. 1). (Top panel) Note the sudden decrease in diameter from the ascending aorta (AAo) toward the transverse aortic arch (AoA). (Bottom panel) After 13 days of materno-fetal HO, the increase in left heart flow resulted in marked increases in the aortic isthmus diameter from 2.3 mm (–4.9 Z-score) to 3.5 mm (–2.7 Z-score). Ruler to the right of the figure is provided to facilitate the comprehension of these small dimensional changes on flow. LA left atrium, RA right atrium, LV left ventricle, RV right ventricle, DA ductus arteriosus, MPA main pulmonary artery; same scaling used for both great vessel images

Full size image

Pending the absence of these ameliorating or neutralizing factors, the smaller a cardiovascular dimension was at the beginning of therapy, the larger its increase during the treatment course. In contrast, the more normal a cardiovascular dimension was at the beginning of materno-fetal HO, the smaller its increase during the treatment course. The latter two observations may be explained by the fact that the increased volume load from materno-fetal HO must have a larger impact on smaller than on larger cardiovascular structures.

The oxygen concentrations required to achieve marked fetal pulmonary vasodilation in this study were well below concentrations known to be harmful for adults, even when chronically administered. In contrast, little is known about fetal safety and the long-term effects of increasing fetal blood oxygen content by materno-fetal HO in late gestation. Therefore, careful observation of fetuses subjected to this novel treatment approach will be required not only before but also after birth.

Among the few data available, fetoscopic blood gas analysis in human fetuses between 17 and 22 weeks of gestation before and after materno-fetal HO with 100% oxygen showed that increases in fetal umbilical vein oxygen tension (\( P_{{{\text{O}}_{2} }} \)) from 39 ± 4 to 82 ± 3 mmHg and increases in oxygen saturation from 73% ± 6% to 95% ± 4% can be achieved by this strategy [4]. These fetal oxygen values resemble normal postnatal values for oxygen tension and saturation obtained in room air. However, because the maternal inspiratory oxygen concentration administered to achieve sufficient fetal pulmonary vasodilation and left heart growth in this study was approximately 45%, it seems unlikely that oxygen tensions and concentrations that might be toxic to a mature fetus at appoximately 34 weeks of gestation can be achieved by this approach. Furthermore, detrimental changes in uterine and fetoplacental blood flows, cardiac failure, brain-sparing, or preterm constriction of the ductus arteriosus did not occur in these first cases.

My major concern after manipulation of fetal pulmonary circulation by materno-fetal HO so close to delivery was whether postnatal cardiopulmonary adaption in the first minutes of life would proceed normally and whether pulmonary hypertensive crises might occur in the coming hours and days. Fortunately, all fetuses adjusted normally to the postnatal environment without evidence for untoward pulmonary vascular events. However, until a much larger number of patients has been studied, neonates who were subjected to materno-fetal HO during fetal life should be carefully monitored for the occurence of pulmonary hypertension.

Limitations of Study

Given the variety and the small number of fetuses—most of which recruited under the umbrella of hypoplastic left heart complex—and the observation that the acute response as well as that over time of the pulmonary circulation to materno-fetal HO may differ depending on various anatomic factors (e.g., oval foramen closure), it was not possible to match the treatment group with an appropriate control group of fetuses that would have mirrored the spectrum of cardiac anomalies observed in the treated cases. I believe that sufficiently large numbers of cases with nearly identical lesions for a randomized trial protocol can only be provided in a timely fashion by multicenter efforts.

Alternatively, the dimensional changes from before to the end of materno-fetal HO in the treated fetuses were overlayed on growth charts derived from normal fetuses from the literature to provide some insight into the dimensional changes with time that are usually seen for the various structures. In addition, if one considers the rapid changes in cardiovascular dimensions that result from altered fetal hemodynamics from experiments of nature, such as twin–twin transfusion syndrome or vascular steal from large tumors, it becomes easier to conceive that similar changes might be achieved by iatrogenic manipulation of fetal cardiovascular flows employing oxygen.

Implications

The therapeutic potential of chronic intermittent materno-fetal HO as a noninvasive tool for fetal cardiac intervention may now be tested in a large spectrum of prenatally detected fetal cardiac malformations with clinically relevant hypoplasia of cardiovascular structures. Apart from cases with hypoplastic left heart complex, many other cardiac malformations that are accompanied by severe underdevelopment of chambers, atrioventricular and semilunar valves, and great vessels (e.g., pumonary atresia with ventricular septal defect, double-inlet left ventricles, double-outlet right ventricles, balanced and unbalanced atrioventricular septal defects) might benefit from the method. In addition, improved left heart loading might benefit recovery of left ventricular function and growth after technically successful aortic balloon valvuloplasties. Given its simplicity, universal availability, and potential benefits for large numbers of patients, intensive research at dedicated centers is now required with the goals to optimize hyperoxygenation schedules, define all suitable cardiac malformations, and assess short and long-term safety of this novel therapeutic approach [2, 3].