Behaviour of the Foramen Ovale Flow in Fetuses with Intrauterine Growth Restriction

Background Foramen ovale (FO) flow may be altered in IUGR. This study was designed to test this hypothesis. Methods Forty pregnant women (24–38 weeks) were divided into 3 groups: group I (IUGR), group II (adequate growth and maternal hypertension), and group III (normal controls). Impedance across the FO was assessed by the FO pulsatility index (FOPI): (systolic velocity − presystolic velocity)/mean velocity. Statistical analysis utilized ANOVA, Tukey test, and ROC curves. Results Mean FOPI in IUGR fetuses (n = 15) was 3.70 ± 0.99 (3.15–4.26); in the group II (n = 12), it was 2.84 ± 0.69 (2.40–3.28), and in the group III (n = 13), it was 2.77 ± 0.44 (2.50–3.04) (p=0.004). FOPI and UtA RI were correlated (r = 0.375, p=0.017), as well as FOPI and UA RI (r = 0.356, p=0.024) and, inversely, FOPI and MCA RI (r = −0.359, p=0.023). Conclusions The FO flow pulsatility index is increased in fetuses with IUGR, probably as a result of impaired left ventricular diastolic function.


Introduction
Intrauterine growth restriction (IUGR) is a signi cant clinical problem, a ecting up to 10% of all pregnancies [1] and even 15% of all monochorionic twin pregnancies [2], with high perinatal morbidity and mortality rates due to fetal hypoxia [3,4]. Placental insu ciency is the etiology in most cases [5,6], although infections, congenital anomalies, and drug misuse are other associated conditions [7]. Monitoring the consequences of fetal hypoxia is the basic obstetric management, given that the only current treatment for IUGR is delivery [7,8].
e fetal heart is a central organ in adaptive mechanisms to hypoxia, and cardiac dysfunction is recognized as the pathophysiologic determinant of clinical deterioration in both early-and late-onset IUGR [4]. Biophysical pro le and ductus venosus impedance become abnormal only in advanced stages of fetal compromise [7], and for this reason, alternative parameters, in particular cardiac function, could provide earlier markers with higher sensitivity [7,9].
In normal pregnancy, a progressive decrease in umbilical artery (UA) impedance occurs, which allows appropriate fetal cardiovascular development [4]. IUGR fetuses have abnormal placental changes, with increased placental vascular resistance and progressive deterioration of the UA ow [10]. UA Doppler is indicated for early detection of placental insu ciency [11]. Fetal circulatory response primarily bene ts the systemic cardiac output (CO), providing an adequate oxygen supply to vital organs [12,13]. With progression of fetal pulmonary and systemic vasoconstriction, an increased right ventricular afterload and a shift of cardiac output to the left ventricle occur [8], with impact on left diastolic function and possible increased impedance to ow through the foramen ovale (FO) [14]. Due to its triphasic ow pattern, the vascular pulsatility index (PI) may represent its impedance, as already demonstrated in other situations such as gestational diabetes [15]. e analysis of the foramen ovale ow dynamics in fetuses with IUGR has not been previously assessed. e purpose of this study was to compare the foramen ovale pulsatility index (FOPI) in IUGR fetuses with the FOPI in fetuses classi ed as of appropriate growth, with and without maternal hypertension. Correlations of the FOPI with maternal, fetal, and placental Doppler indices were tested.

Methods
A cross-sectional, controlled, nonblinded study of 40 single fetuses with 24 weeks of gestational age (GA) or more was designed.
e sample included 15 growth-restricted fetuses (group I), 12 fetuses with normal weight for gestational age from hypertensive mothers (group II), and 13 fetuses with normal weight for gestational age from healthy mothers (group III). Maternal hypertension was classi ed according to Guidelines of the American College of Obstetricians and Gynecologists [16]. Fetuses with other abnormalities or fetuses whose mothers used drugs or tobacco were excluded from the study.
All pregnant women provided written informed consent to participate in the study, which was approved by the Medical Research Ethics Committee of the Institute of Cardiology of Rio Grande do Sul.
Gestational age was determined in all fetuses by rst trimester ultrasound. After 24 weeks, a morphological ultrasound was performed for weight estimation (according to the Hadlock method) [17], placental assessment, amniotic uid volume determination, and ow velocimetry of the umbilical, middle cerebral, and uterine arteries. Under fetal apnea, the resistance index (RI) was obtained by Doppler ow analysis and determined by the ratio (systolic velocity − diastolic velocity)/systolic velocity. e amniotic uid index was measured by the sum of the pockets of the four quadrants of the maternal abdomen.
IUGR with placental insu ciency was de ned by fetal weight below the 10th percentile for gestational age in combination with abnormal Doppler indices (either umbilical artery RI > 95th centile, middle cerebral artery RI < 5th centile, or uterine artery RI > 95th centile for gestational age).
Fetal echocardiography was performed using an Acuson Aspen (Acuson, Mountain View, CA, USA) ultrasound system, with a multifrequencial transducer (3 to 5 MHz). Cardiac structural assessment used segmental sequential analysis. Flow in the FO was obtained at a four-chamber view, with the pulsed Doppler sample volume placed at the left atrial surface of the ori ce in the central portion of the color ow mapping, with an angle of less than 20° [15,18]. Flow impedance was assessed by the pulsatility index, obtained by the ratio (systolic velocity − presystolic velocity)/mean velocity [19] (Figure 1). e statistical package SPSS version 15.0 (SPSS Inc., Chicago, IL) was used for data analyses. Quantitative analyses were reported as mean ± standard deviation (SD). Analysis of variance (ANOVA) was used for each variable.
e Tukey test was applied for the individual assessment of each group when signi cant di erences were present. e chi-square test of the observed frequencies was used for qualitative analyses. e correlation of the FOPI with fetal arteries' RI was tested by the Pearson test. e critical signi cance level was p < 0.05. e sample size was not calculated since no previous studies were available to assess di erences. Nevertheless, the study power was calculated based on the pulsatility index of the foramen ovale, considering a signicance level of 5%, standard deviations of 0.44 and 0.99, and a di erence of 0.93 in the pulsatility indices of groups I and III, achieving a study power of 84%. e reproducibility of measurements of the FOPI was tested. Intraobserver variability was assessed in twelve nonconsecutive normal fetuses by repeating the measurements on two occasions (2 days apart) with the same conditions. Interobserver variability was also assessed, with measurements repeated in 10 normal fetuses on the same day by a second observer blinded to the results of the rst examination. Intraclass correlation coe cients were calculated to measure the degree of consistency among measurements. e Bland-Altman plots were created to show the mean of di erences between measurements. Table 1 presents the characteristics of the sample, showing that the groups were comparable except for maternal age and body mass index (BMI) in group II when compared to the other two groups. Gestational age ranged from 24 to 38 weeks. Minimum maternal age was 14 years and maximum 38 years.

Discussion
In this study, we assessed the foramen ovale ow dynamics in fetuses with IUGR and with adequate growth of both normotensive and hypertensive women. It was observed that IUGR fetuses had an increased impedance to ow through the FO, represented by an increased FOPI compared to control groups, probably as a result of impaired diastolic function. e same e ect has been demonstrated   in fetuses of diabetic mothers (FDMs) with myocardial hypertrophy, where the FOPI was higher than that in normal fetuses [15]. Several studies have demonstrated changes in diastolic function in FDMs, being the ductus venosus PI [20] and the pulmonary vein PI [21] signi cantly higher in fetuses with myocardial hypertrophy, probably due to a lower ventricular compliance. e mobility of the septum primum is lower in fetuses with IUGR [22] and myocardial hypertrophy [23] when compared to normal fetuses, due to a decreased ventricular compliance. e shortening fraction of the left atrium was also shown to be decreased in fetuses with myocardial hypertrophy, presenting a negative correlation with the septal thickness [24]. Left ventricular myocardial diastolic velocities by tissue Doppler imaging (TDI) were shown to be signi cantly higher in FDM, irrespective of the presence of cardiac hypertrophy, suggesting that maternal diabetes is associated with changes in left ventricle diastolic function even without myocardial hypertrophy [25]. TDI evaluation of IUGR fetuses showed that myocardial early and late diastolic velocity ratios were higher in IUGR in lateral and septal mitral annulus, when compared with normal growing fetuses [9]. A subsequent study [26] con rmed these ndings. Both studies concluded that TDI can probably be a more sensitive method for detection of diastolic cardiac dysfunction in IUGR fetuses than conventional mitral and tricuspid valve Doppler [9,26]. In a recent study in rabbits, cardiac weight reduction was demonstrated in response to IUGR, due to a decrease in the number of cardiomyocytes in both ventricles. However, an increase in the mean volume of cardiomyocytes occurred in the left ventricle. is demonstrates that the right and left ventricles respond di erently to placental insu ciency [27].
IUGR fetuses have impaired ventricular lling, with higher atrioventricular valve E/A ratios, lower aortic and pulmonary artery systolic peak velocities, left ventricular CO increase, and right ventricular CO decrease. ese hemodynamic changes favor ow into the left ventricle in order to maintain cerebral perfusion [7,18]. erefore, in the early stages of the disease, adequate levels of oxygen and substrates are maintained,  despite the reduction in placental transfer. During the process of centralization, fetal changes occur in cardiac afterload, decreasing left ventricle afterload due to cerebral vasodilatation and increased systemic vasoconstriction [28]. Furthermore, hypoxia may impair myocardial contractility, while polycythemia may increase blood viscosity [7]. In fetuses with IUGR, a faster presystolic left-to-right ow velocity through the FO is in accordance with a study that assessed ow in fetuses with left ventricle hypoplasia or left side obstruction, in which there was increased left atrial pressure [29].
Umbilical and uterine artery ow RIs were not di erent between groups. Correlation analysis of the RI of maternal and fetal ows with the FOPI showed a positive weak correlation with the UA and the UtA and a negative weak correlation with the MCA. Nonstrati cation according to the severity of placental insu ciency in group I could be a limitation and may explain these data, as a stronger correlation would be expected with more severe fetal compromise. Left diastolic function assessed by the FOPI could already be altered in cases of IUGR with placental insu ciency in early stages. Turan et al. [28] assessed the time of onset of placental dysfunction in fetuses with IUGR, and in two of the three study groups, patients were recruited when there was no placental dysfunction. is situation was diagnosed only two to four weeks later. e fetal cardiac output is redistributed during hypoxia, promoting ow to vital organs like brain, heart, and adrenal glands [30][31][32] and justifying the MCA resistance decrease associated with cardiac diastolic dysfunction assessed by FOPI. e possibility of maternal hypertension without IUGR is supported by other authors. Grisaru-Granovsky et al. [33] showed that the presence of IUGR in fetuses of hypertensive mothers was not correlated with worsening of the hypertensive disorder but re ected the individual predisposition of fetuses to abnormal development. Our study had a small sample size of group I, not allowing categorization in relation to the presence of associated maternal hypertensive disorder.
is study has limitations. It was not blind, allowing the potential occurrence of a mensuration bias. Being a crosssectional study, it was not possible to assess the sequential changes in ow through the FO with the evolution of pregnancy. e lack of strati cation according to the severity of fetal compromise and maternal hypertensive disorder in group I has already been mentioned. It was observed that group II had higher maternal ages and BMI than the other groups, but these are not expected to primarily interfere with fetal cardiac hemodynamics.
Despite the occurrence of oligohydramnios in one-third of the IUGR sample, technical aspects of assessment of the FOPI did not represent any limitation, as demonstrated by the strength of agreement between two sets of measurements obtained by trained operators.
It has been reported that IUGR fetuses may show di erent outcome patterns [28]. For this reason, it is important to perform serial tests in an e ort to promote birth with the lowest possible morbidity and mortality. e use of other techniques, such as fetal echocardiography, may help monitoring IUGR fetuses [34]. e present study suggests that the presence of an increased FOPI, especially above 2.95, may suggest abnormalities in fetal diastolic heart function, thus potentially in uencing the overall obstetric management.

Disclosure
Paulo Zielinsky takes responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.

Conflicts of Interest
e authors declare that there are no con icts of interest.