Effects of uterine Doppler on midbrain growth and cortical development in late onset fetal growth restricted fetuses: a prospective cross-sectional study

Abstract

Objective

To investigate midbrain growth, including corpus callusum (CC), cerebellar vermis (CV) and cortical development in late fetal growth restriction (FGR) depending on uterine artery (UtA) Pulsatility Index (PI) values.

Methods

This was a prospective study including singleton fetuses with late FGR characterized by abnormal cerebral placental ratio (CPR). According to UtA PI values, the FGR fetuses were subdivided into normal ≤95th centile) and abnormal (>95th centile). Neurosonography was performed at 33–44 weeks of gestations to assess CC and CV lengths and the depth of Sylvian fissure (SF), parieto-occipital (POF) and calcarine fissures (CF). Neurosonographic variables were normalized for fetal head circumference size.

Results

The study cohort included 60 fetuses with late FGR, 39 with normal UtA PI and 21 with abnormal PI values. The latter group showed significant differences in CC (median (interquartile range) normal 35.9 (28.49–45.53) vs abnormal UtA PI 25.31(19.76–35.13) mm; p < 0.0022), CV (normal 25.78 (18.19–29.35) abnormal UtA PI 17.03 (14.07–24.16)mm; p = 0.0067); SF (normal 10.58 (8.99–11.97)vs abnormal UtA PI 7.44 (6.23–8.46) mm; p < 0.0001), POF (normal 6.85 (6.35–8.14) vs abnormal UtA PI 4.82 (3.46–7.75) mm; p < = 0.0184) and CF (normal 04.157 (2.85–5.41) vs abnormal UtA PI 2.33 (2.49–4.01)); p < 0.0382).

Conclusions

Late onset FGR fetuses with abnormal UtA PI showed shorter CC and CV length and delayed cortical development compared to those with normal uterine PI. These findings support the existence of a link between abnormal brain development and changes in utero placental circulation.

Keywords:

• Fetal growth restriction
• corpus callosum
• cerebellar vermis
• cortical folding
• neurosonography
• uterine artery PI

Introduction

Fetal growth restriction (FGR) complicates approximately 3–5% of all pregnancies and is associated with increased perinatal morbidity, with both short-term and long-term neurological sequelae [1–3]. Despite their better perinatal outcome, late onset FGR fetuses, defined as those diagnosed > 32 weeks, still show suboptimal neurological outcomes [4,5].

Advances in prenatal imaging techniques have led to a comprehensive assessment of fetal brain maturation [6,7]. Previous studies have reported significant changes in fetal brain growth and cortical maturation in different clinical scenarios such as, congenital heart diseases [8], in vitro fertilization [9], pre-eclampsia [10] and mild ventriculomegaly [11] which can potentially correlate with a higher risk of a sub-optimal neuropsychological performance after birth.

Recently the presence of brain abnormalities has also been reported in late onset FGR fetuses, including changes in corpus callosum (CC], cerebellum, white matter and cortical development [12–14]. These changes seem to be more pronounced in case of feto-placental circulation abnormalities characterized by an altered cerebro-placental ratio (CPR). UtA Doppler represents an indirect indicator of placental function and reflects maternal perfusion of the utero-placental unit [15]. An increase in UtA impedance has been shown to be associated with reduced oxygen supply to the fetus that may affect its growth and development [16]. However, there are currently no reports of how this might impact on brain maturation/development.

The objective of this study was to examine the growth of midbrain structures and cortical development in late onset FGR according to the presence or absence of UtA PI abnormalities.

Methods

Study population

This was a prospective observational case control study, performed at the Department of Obstetrics and Gynecology of the University of Roma Tor Vergata, from February 2022 to August 2023. It included pregnancies complicated by late FGR (i.e. an estimated fetal weight < 10th centile) and a CPR value < 5th centile as defined by the Delphi consensus [17]. All pregnancies were accurately dated by first trimester crown-rump length measurement. Birthweight and CPR centiles were calculated using local reference curves [18–20]. These late FGR pregnancies were subsequently subdivided as follows: (a) UtA PI < 5th centile for our reference limits (b) UtA PI ≥ 5th centile [19]. Exclusion criteria included multiple pregnancies, fetal structural or chromosomal abnormalities, perinatal infection, pregnancy conceived by in vitro fertilization, chronic maternal diseases, smoking habit or alcohol consumption during pregnancy, presence of any kind of complications during pregnancy (including hypertension, pre-eclampsia, gestational diabetes mellitus) and loss at follow-up. Clinical management followed the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) guidelines on FGR [21]. This study was conducted according to the Declaration of Helsinki for Medical Research involving Human Subjects. The study protocol on fetal neurosonography was approved by the Ethical Committee of the Policlinico Tor Vergata on January 2022 protocol n° prot.0001446/2022). All pregnant women provided their written informed consent to participate to the study.

Neurosonography

Every woman underwent a detailed neurosonographic exam at 32–34 weeks using a Samsung Hera W10 ultrasound device (Samsung Medison Co. Ltd. Seoul. South Korea). According to the study protocol, multiplanar assessment of the fetal brain using axial, sagittal and coronal views was initially performed to rule out fetal CNS anomalies, following the ISUOG neurosonographic guidelines[22]. Then all women underwent biometric assessment of the fetal brain as follows: CC and cerebellar vermis (CV) lengths were obtained by using a mid-sagittal plane of the fetal brain as previously reported [23,24]. Fetal cortical development was assessed following a previously reported methodology described by Alonso et al. [8,25–27]. Briefly, sylvian fissure (SF) length was measured on a trans-thalamic plane by tracing a vertical line perpendicular to the midline, starting from the insula’s external border. The parieto-occipital fissure (POF) was measured in a trans ventricular plane at its point of maximum length and symmetry compared to the contralateral fissure, beginning from the midline and avoiding inclusion of cortical tissue in the measurement. Finally, the calcarine fissure was identified in a coronal view of the trans-cerebellar plane and measured by tracing a line from the apex of the fissure to the midline. The planes were obtained using a transvaginal approach in fetuses with cephalic presentation and transabdominally in fetuses with breech presentation.

Measurements were performed offline using a DICOM viewer by an experienced examiner who was blinded to the study groups (IM).

The inter-observer and intra-observer reproducibility was assessed in 20 fetuses, and the results have been previously reported [8].

Statistical analysis

Normality of the considered variables was tested using the Shapiro-Walk test, and according to their distribution, parametric or non-parametric tests were applied.

Data were presented as numbers and percentages for categorical variables and as median values and interquartile range (IQR) or mean and standard deviation (SD) for continuous variables.

Neurosonographic measurements were evaluated as absolute values and as multiple of medians (MoM). MoM were calculated as the ratio between the measured value and the expected median obtained from our previously constructed reference limits for fetal head circumference (HC) [26]. The choice of HC as a dependent variable instead of gestational age was based on our previous findings showing a better fitting of the former variable with neurosonographic values.

The comparison among groups was performed by unpaired t test or Mann Withney U Test in function of the data distribution, followed by post-hoc analysis.

Analysis was performed using SPSS 29 (SPSS, Chicago, IL, USA) and Prism 10 (GraphPad Software Boston, MA USA) softwares. A two-sided p value < 0.05 was considered statistically significant.

Results

An initial sample of 66 pregnancies was considered for the study but 6 were then excluded due to inadequate neurosonographic images leaving 60 fetuses for the analysis, 39 with normal UtAPI and 21 with abnormal PI.

Their baseline and perinatal characteristics are reported in Table 1. Maternal characteristics and gestational age at the time of neurosonographic examination did not differ among groups while FGR fetuses with abnormal UtA PI were delivered earlier, had a higher incidence of emergency Cesarean Section and admission to neonatal special care unit. No differences were found in UA PI, MCA PI and CPR centiles values among the 2 groups.

Table 1. Baseline maternal and perinatal characteristics of pregnancies with normal and abnormal uterine artery (UtA) PI values.

Variable	UtA PI Normal N = 39	UtA PI abnormal N = 21	p*
Maternal age (years)	31 (27–34)	31 (27.5–36)	0.577
Pregestational BMI (kg/m^2)	24 (21–28)	23 (20.5–25.5)	0.147
Caucasian ethnicity N (%)	33 (84.6)	16 (76.2)	0.492
Nulliparous N (%)	24 (61.5)	15 (71.4)	0.572
GA at ultrasound (weeks)	32.9 (32.3–33.43)	33 (34.6–36.6)	0.361
HC centile	4 (3–8)	4 (3–7)	0.998
UA PI centile	93 (90–98)	94 (91–98)	0.223
MCA PI centile	3 (1–6)	2 (1–5)	0.591
CPR centile	2 (1–3)	2 (1–3)	0.994
UtA centile	67 (52–77)	98 (96–99)	0.0001
GA at delivery (weeks)	36.7 (35.6–37.3)	35.3 (34.6–36.9)	0.028
Emergency Cesarean Section N (%)	8 (20.5)	11 (52.4)	0.019
Birthweight (g)	2070 (1810–2290)	1780 (1625–2100)	0.014
Male gender N (%)	19 (48.7)	11 (52.5)	0.999
5-min Apgar score < 7	0 (0)	2 (9.5)	0.118
Neonatal admission to the special care unit n (%)	2 (5.1)	5 (23.8)	0.044

BMI: body mass index; SFA: small for gestational age; FGR: fetal growth restriction; GA: gestational age; HC: head cirv;UA umbilical artery; MCA: middle cerebral artery; PI: pulsatitilty index; CPR: cerebral placental ratio. *<Comparison performed using Mann–Whitney U test, Fisher’s exact or chi-square tests. Data are given as median (interquartile range) or number (%).

Table 2 summarizes the neurosonographic findings. FGR fetuses with abnormal UtA PI showed significantly smaller CC (U p < 0.0022) and CV (p = 0.0067) length compared to those with normal uterine Doppler (Figure 1). Similarly, the depth of SF (p < 0.0001), POF (p = 0.0183) and CF (p = 0.0382) were significantly reduced in FGR fetuses with abnormal UtA PI (Figure 2), Correction for HC did not affect statistical significance (Table 2) for both midbrain structures (Supplementary Figure 1) and cortical fissures (Supplemental Figure 2). No associations were found between the evaluated Doppler indices and neurosonographic findings.

Figure 1. Violin plots showing distribution of corpus callosum and cerebellar vermis in growth restricted (FGR) fetuses with normal or abnormal uterine artery (UtA) PI. Circles represent individual points, horizontal lines indicate median (continuous), and interquartile range (IQR) (dotted). Comparison among groups was performed by Mann–Whitney U test.

Figure 2. Violin plots showing distribution of Sylvian fissure, parieto occipital fissure and calcarine in growth restricted (FGR) fetuses with normal or abnormal uterine artery (UtA PI). Circles represent individual points, horizontal lines indicate median (continuous), and interquartile range (IQR) (dotted). Comparison among groups was performed by Mann–Whitney U test.

Table 2. Neurosonographic findings in fetuses with normal and abnormal uterine artery (UtA) PI values.

Variables	UtA PI normal N = 39	UtA PI abnormal N = 21	p*
CC mm	35.90 (28.49–45.53)	25.31 (19.76–35.13)	0.0022
CC MoM	0.90 (0.69–1.14)	0.61 (0.49–0.89)	0.0020
CV mm	25.78 (18.19–29.35)	17.03 (14.07–24.16)	0.0067
CV MoM	1.13 (0.76–1.26)	0.74 (0.59–1.02)	0.0025
SF mm	10.58 (8.99–11.97)	7.44 (6.23–8.46)	0.0001
SF MoM	0.81 (0.69–0.93)	0.55 (0.47–0.65)	0.0001
POF mm	6.85 (6.35–8.14)	4.82 (3.46–7.75)	0.0184
POF MoM	0.79 (0.62–0.91)	0.55 (0.38–0.89)	0.0121
CF mm	4.157 (2.85–5.41)	2.33 (2.49–4.01)	0.0382
CF MoM	0.81 (0.56–1.01)	0.72 (0.48–0.76)	0.0198

CC: Corpus callosum; CV: cerebellar vermis; SF: sylvian fissure; POF: parietal occipital fissure; CF: calcarine fissure; MoM: multiple of medians. *Comparison performed using Mann–Whitney U test. Data are given as median (interquartile range).

Discussion

Main findings

In this study we evaluated neurosonographic variables in a population of isolated late FGR fetuses by excluding potential confounding variables such as type of conception, chronic or pregnancy induced maternal disease, smoking habit or alcohol consumption, that may directly affect brain development. We demonstrated that abnormal UtA PI in late FGR affects midbrain structures resulting in shorter length of CC and CV and delayed cortical development. It is noteworthy that these differences were also achieved following data normalization for HC values thereby limiting the possible direct effects of fetal head dimension on brain growth.

Strengths and limitations

The main strengths of this study include its prospective design, the assessment of brain developmental and cortical maturation using an a priori designed protocol, and blinded assessment of the included fetuses. Furthermore, the restricted analysis ofa homogeneous group of late FGR fetuses, obtained after exclusion of potential confounding variable, allowed us to investigate the direct impact of UtA PI on brain development.

Cross-sectional design, lack of comparison with newborn imaging and the absence of postnatal follow-up represent its major limitations. However, previous studies using magnetic resonance in late-onset FGR fetuses have shown that the changes in cortical development are associated with a worse neurobehavioral performance after birth [28–30].

Moreover, due to skull shadowing, our cortical measurements were limited only to the distal brain hemisphere. However, it is unlikely that in late SGA an asymmetry in brain development is present, thus limiting this potential bias. We also acknowledge the absence of data concerning the concomitant hemodynamic of different fetal vascular districts, such as aortic isthmus or lung hemodynamics, that may impact neonatal adaptation [31,32].

Finally, it should also be observed that the differences between the neurosonographic variables among the two groups are subtle and their individual clinical significance still remains to be elucidated.

Comparison with existing literature and clinical implications

Our results are consistent with previous reports showing a reduction of CC growth and cortical development in late FGR fetuses [12–14]. However, establishing a direct relationship between growth defect and impaired brain maturation is challenging as other cofactors can affect brain development in these fetuses, such as the co-existence of maternal disease, type of conception and underlying cause of the growth defect. In order to limit this potential bias this study considered a homogeneous cohort of late FGR fetuses without any other evident associated confounding factors. Furthermore, the neurosonographic measurements obtained in this study were evaluated using reference limits constructed with HC as an independent variable in order to minimize the influence of fetal head size among groups [27]. Our findings support the hypothesis that those cases exposed to higher levels of hypoxia, as expressed by Doppler changes in UtA, suffer a greater impact on brain development. The CC is the major brain commissure [33], its measurement is considered a proxy of overalll brain development and it is used as a biomarker for monitoring cerebral growth in preterm infants [34]. Cortical folding is a complex organizational process which develops in an exponential way, most importantly during the third trimester, when the cortical surface evolves into a complex system of sulci and gyri to increase its surface and gain new functions in its different areas. Parieto-occipital, cingulate and calcarine fissures deepen progressively with advancing gestational age [6], The Sylvian fissure increases its size and undergoes a so called “opercularization” process [35]. Abnormal sulci changes and gyral formation during the third trimester may determine neurological disorders and development delay later in life [11,28]. Therefore, it is possible to speculate that in our population of late FGR with abnormal impedance to flow in uterine artery, are characterized by a more severe degree of hypoxia that may induce the delay in the neurodevelopment found.

Irrespective of the pathophysiology behind this process, the findings from this study showed that neurosonography can also be applied in fetuses not showing brain abnormalities and questions whether detailed assessment of fetal brain should be undertaken in pregnancies at a higher risk of fetal and post-natal neurocognitive impairment adding a clinical perspective to neurosonography [7].

Further longitudinal studies aimed at assessing fetal brain maturation in pregnancies complicated by late FGR are needed to confirm the results of this study and identify a sub-set of fetuses at a higher risk of impaired neuropsychological development after birth. The identification of thesefetuses opens the window to the possibility to test the role of changing maternal lifestyle, behavior and postnatal interventions to modify fetal and neonatal neurodevelopment [36,37].

Conclusions

In conclusion, late onset FGR with abnormal UtA PI values had significantly shorter CC and CV lengths and showed delayed cortical maturation when compared to matched FGR fetuses with normal uterine Doppler. These differences are evident in these fetuses after correction for HC, suggesting a delayed brain development in the presence of abnormal utero placental Doppler, independent of head growth. These findings support the role of third trimester neurosonography as a tool for evaluating fetal brain development. Future research correlating prenatal neurosonographic findings with postnatal follow-up are required and such studies are currently in progress.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data available from Authors on reasonable request.

Additional information

Funding

The study was supported by a grant of the MUR PE00000006 MNESYS PNRR [MUR PE00000006 MNESYS PNRR].