Long-Term Neurologic Consequences following Fetal Growth Restriction: The Impact on Brain Reserve Open Access

Fetal growth restriction (FGR) describes a fetus’s inability to attain adequate weight gain based on genetic potential and gestational age and is the second most common cause of perinatal morbidity and mortality after prematurity. The identification of FGR is, however, not straightforward, as fetal growth cannot be assessed through a single biometric evaluation of the fetal size, and growth potential is hypothetical. Size is used as a proxy with the most common definition of small for gestational age (SGA) being an estimated fetal weight or abdominal circumference below the 10th centile [1].

However, not all fetuses who fall below the 10th centile are growth restricted and conversely it is not plausible that none of the fetuses above the 10th centile have not suffered some form of placental insufficiency and growth restriction. In clinical practice, the identification of FGR relies on a combination of size, growth velocity and Doppler studies [2]. In addition, there are 2 phenotypes of FGR, early and late, commonly distinguished by the time of diagnosis before or after 32 weeks, respectively [3]. Early FGR tends to have a more severe placental problem; the fetuses are smaller in size and are more likely to have abnormal umbilical Doppler. The late FGR are not necessarily very small or have abnormal umbilical Doppler studies; their placental insufficiency is not as severe and has an impact on fetal growth velocity only at a later stage. Low pulsatility index in the middle cerebral artery, when discovered, reflects the redistribution of blood from the periphery to the central organs (heart, brain and lungs) – the so-called brain sparing effect. Late FGR management commonly includes induction of labor with the additional risk of acute deterioration during delivery. Both early and late stage, FGR has been associated with poorer long-term outcome when compared with normally grown fetuses [4].

The causes of being small in the antenatal period may be divided into 3 categories for simplicity: (1) the normal small (constitutional small with expected good outcomes), (2) the abnormal small (including fetuses with chromosomal anomalies, genetic conditions and fetal infections) and (3) the starved small (the ones truly growth restricted due to placental insufficiency). In the absence of fetal infection, genetic abnormalities and congenital malformations, the majority of cases of FGR are thought to be related to compromised uterine circulation to the placenta [5].

Various histopathological changes are observed in the placenta in FGR including villous infarction, maternal vascular changes, and various alterations of villous morphology [6]. However, a quarter of placentas from FGR lack any morphological abnormality on routine examination. The morphology of FGR detected later in pregnancy is more heterogenous and less specific and the histologic changes less dramatic. Other features noted in the placenta of FGR include an increased placental thickness between the first and second trimester, placental lakes in the second trimester and grade 3 calcification [7].

The theory of fetal programming suggests that environmental influences during critical periods of fetal development can have a profound impact on an individual’s long-term health and disease risk. For instance, when the fetus experiences reduced nutrient and oxygen supply due to placental insufficiency, it undergoes various adaptive changes aimed at improving its chances of survival. Fetal programming is well established in the context of cardiovascular health, with FGR associated with alterations in cardiac structure, blood vessel development, increased risk in cardiovascular disease later in life [8] as well as increased risk of type 2 diabetes [9].

Yet congenital infection, copy number variant anomalies, malnutrition, toxic or fetal inflammatory exposure may be underreported with very little data available to date specifically for each condition. Up to 5.5% pathogenic copy number variants are identified when fetuses with growth restriction are analyzed systematically with chromosomal microarray analysis [10]. Similarly, cannabis has been associated with growth restriction yet well underreported and often overlocked with singular neurotoxic effects on cortical integrity and connectivity very different from chronic hypoxemia following placental insufficiency [11]. Its exposure may lead to a series of severe behavioral and significant psychiatric conditions in the offspring not necessarily found in other causes of growth restriction. Indeed, identification of underlying causes is not available in large epidemiological studies, so that proper stratification of long-term risk for brain impairment is currently not available.

Regarding Doppler studies in FGR of placental origin, the brain sparing effect with vasodilatation of middle cerebral artery to prioritize brain perfusion at the expenses of peripheral tissues has been associated with a positive adaptative phenomenon but some studies have associated this with poor cognitive and behavioral development. The subject is still under great controversy and further studies are awaited. It is likely that the degree of brain sparing and its association with gestational age are factors that might turn a positive compensatory mechanism into a potential deleterious side effect [12]. In early onset, FGR prematurity plays a major role in neurodevelopment independently of hemodynamic changes.

Low fetal oxygen exposure secondary to placental or environmental factors has been shown to induce mitochondrial damage, increase in reactive oxygen species, peroxynitrite and the release of damage-associated molecular patterns as well as cause a sterile inflammation with an increase in pro-inflammatory cytokines [13]. This leads to a vicious cycle with further remodeling of placental microvasculature and worsening fetal chronic hypoxia and an increase of fetal anaerobic metabolism.

At the fetal cerebral level, placental insufficiency has been shown to cause significant activation of astrocytes and microglial cells with a concurrent increase in IL1β and TNFα pro-inflammatory cytokines and a reduction in neurons in a piglet model when analyzed at birth [14]. In fetal sheep exposed to chronic hypoxemia, inflammation and oxidative stress were shown to include the brainstem [15].

Clinical presentations other than chorioamnionitis may also increase cytokine production. Indeed, obesity is now a well-established driver of systemic inflammation with increased levels of inflammatory cytokines explaining the counterintuitive association with poor fetal growth [16]. Increased inflammatory cytokines (TNFα and IL1β) have been confirmed in the circulation of intrauterine growth restricted newborns [17]. The increased level of circulating cytokines is well known to activate microglia and cause significant astrogliosis [18] known for their deleterious effects on brain development. In a preclinical rat model using lipopolysaccharide, a well-known inflammatory agent recognized to greatly increase these cytokines and generate significant gliosis, we recently showed a major impact on five hundred genes with methylation related to brain development causing their downregulation and a hypomethylation of genes related to inflammation causing their upregulation [19]. In a less dramatic preclinical model of growth impairment using food restriction, reduction methylation via a reduction of DNMT1 expression was sufficient to impair neurogenesis in offspring [20].

Furthermore, chronic fetal hypoxia has been shown to eventually lead to apoptosis with increase in caspase-3 activation and Bax expression together with a reduction of bcl-2, an anti-apoptotic protein [21]. Oxidative stress has been shown to cause a reduction of pre-oligodendrocytes [21]. At the macroscopic level, these major changes produce a reduction of cortical and deep gray matter volume, and hypomyelination, with recent evidence showing elevated cell death and reduced cell proliferation leading to gray and white matter deficits in the brainstem [15]. This is further complicated by cerebrovascular remodeling and loss of cerebral vasoreactivity as shown in the growth restricted fetal lamb, increased permeability [22] together with a reduced angiogenesis and vascular density with a reduction in VEGF immunoreactivity [23].

Interestingly in IUGR mice, decreased hippocampal volumes and neuronal progenitors were associated with an accelerated embryonic hippocampal dentate gyrus neurogenesis [24]. In a guinea-pig model of growth restriction, there was a reduction of CA1 pyramidal neurons [25]. Knowing the pivotal role of hippocampal integrity in long-term cognitive function [26], it is worrisome to see an abundance of preclinical data on its vulnerability. When the impact of antenatal hypoxia and fetal stress was carried out in offspring of 5xFAD mice, a reduction in brain weight was shown but was only transient [27]. Yet antenatal hypoxia was shown to exacerbate cognitive decline that was associated with a synaptic loss rather than a neuronal loss together with a reactive astro-microgliosis. Finally, prenatal hypoxia also leads to a deficit of peptidases that are involved in catabolism of amyloid-β peptide (Aβ), a hallmark of Alzheimer’s disease [13].

In a rabbit model, metabolomics analysis revealed several metabolites to be affected by growth restriction. N-acetylaspartylglutamic acid (NAAG), N-acetylaspartate (NAA) where among the metabolites most affected which is very similar to findings using magnetic resonance spectroscopy detailed below [28].

In recent years, magnetic resonance imaging (MRI) has emerged as a pivotal tool in obstetrics, facilitating groundbreaking studies in placental perfusion dynamics and fetal oxygenation during pregnancy providing new insight into intrauterine growth restriction. Volumetric study of the placenta has recently shown to correlate with total brain, cerebral, and cerebellar volumes [29]. Placental volume could even be used as a marker of behavioral deficits at term equivalent age [30]. Arterial spin labeling in MRI has been used successfully in pregnancies complicated by FGR restriction where it showed a reduction in basal placental perfusion [31]. However, its complex setup makes it challenging to be used clinically.

Leveraging sophisticated MRI techniques like blood oxygen level-dependent (BOLD) imaging, researchers have made significant strides in the non-invasive evaluation of placental blood flow and oxygen exchange. BOLD imaging, based on the paramagnetic properties of deoxyhemoglobin, allows for the assessment of oxygenation status in fetal organs and tissues and was shown first successfully in a preclinical rat model with major differences following hyperoxygenation [32] with similar findings in humans in 2022 with a decrease in fetal placental oxygenation capacity as well [33]. Melbourne et al. [34] solved the limited reliability of oxygen measurements at baseline related to the overlap between maternal and fetal blood by using a new imaging methodology mechanism (Diffusion-rElaxation Combined Imaging for Detailed Placental Evaluation; DECIDE). The same technic was later shown to identify significant decrease in the feto-placental oxygen saturation in fetuses with growth restriction [35]. Using the same approach, they demonstrated in a larger translational validation study, the reliability of DECIDE and how it can reliably identify significant drops in placental oxygenation during pregnancy at multiple timepoints as well as predict fetal weight 15–30 days later [36] with a striking correlation shown in humans between fetal oxygen saturation and fetal growth rate. Furthermore, they showed how FGR was associated with identifiable microstructural placental changes with hindered diffusivity that was confirmed in several studies [37].

Noninvasive MRI studies have provided valuable insights into the intricate interplay between placental function and fetal oxygenation, offering a deeper understanding of fetal development and pregnancy outcomes. Approaches such as DECIDE imaging will provide complimentary tools for not only a better understanding of the pathophysiology of FGR but also may be useful for better monitoring of these high-risk pregnancies in the clinic. The integration of MRI-derived metrics into prenatal care holds promise for enhancing risk stratification, early intervention, and personalized management strategies. As MRI technology continues to evolve, collaborative efforts between clinicians, researchers, and imaging specialists will be essential for translating these research findings into clinical practice and optimizing maternal-fetal health.

Fetal magnetic resonance spectroscopy has allowed the evaluation of brain integrity showing similar findings to that of preclinical data with poor neuronal growth revealed in utero by a reduction of NAA [38], a marker of neuronal integrity. In some severely affected fetuses, evidence of impaired oxygen delivery was also shown with increased levels of brain lactate [39].

A similar NAA reduction was described in preterm infants with growth restriction [40] and found to remain lower compared to controls at age one [41]. Interestingly, this population was found to have a small increase in striatal myoinositol [42], a well-known marker of gliosis. If a reduction of NAA alone may be related to a simple delay in brain maturation, the associated presence of myoinositol is far more worrisome. Indeed, perinatal brain inflammation with activation of astrocytes and microglia may carry hefty consequences [43, 44]. Most of the available imaging data to date is done using voxel positioning based as per researcher assumption. It is possible that increases in levels of myoinositol may be missed if the voxel was not placed in the right position. To prevent this from happening new spectroscopic imaging techniques are now available and will allow us to properly map these changes throughout the brain.

Unfortunately, the impaired brain metabolism is further complicated by impaired structure. We and others have described how growth restriction carries a negative impact on cortical growth and hippocampal volume at term equivalent [42, 45]. When evaluated at term equivalent these children already show remarkable neurobehavioral impairment that correlated with fetal brain volume [30]. Recent studies confirm these findings that were associated with a lower neurocognitive outcome at 22–24 months [46] and increased risk of autism [47] when compared to preterm infants without growth restriction.

From ages 4 to 7 children with impaired growth were found to retain a smaller brain volume with associated reduction of cortical gray matter, the right hippocampus and basal ganglia and absence of normal cortical thinning process in the superior and medial prefrontal cortex [48]. At age 6, growth restriction was associated with weaker interhemispheric connectivity (superior frontal gyrus, precuneus, and cingulate gyrus) and intrahemispheric connectivity (middle frontal gyrus and subcortical gray structures [putamen, globus pallidus, thalamus]) [49].

An interesting study combined three cohorts age 1, 6, and 10 showed how impaired microarchitecture following growth restriction was slowly improving over time [50] thus indicating a delay in maturation rather than a persistent dysmaturity in these children. Yet, significant and persisting impaired cognitive outcome at age 12 was confirmed in preterm infants with growth restriction in a meta-analysis based on data from over 50 thousand children [51]. At age 15, impaired fetal growth was found to still be associated with a reduction of cerebral cortex, white matter, thalamus, and the cerebellar cortex with no difference in the hippocampal volume [52, 53].

To date there are two main studies in young adults born SGA. A Norwegian cohort found that in a group of 58 adults born with a mean birth weight of 2.9 kg a persistent impact with a reduction in the volume of caudate nucleus and putamen and a reduction of cortical surface found in temporal lobe and anterior cingulate cortex but without any significant changes in IQ [54].

We found long-lasting changes in a cohort of 39 adults born in Bogota with a mean smaller birth weight of 1.8 kg [55]. Using voxel-based morphometry, significantly smaller clusters in the left frontal lobe, cingulate cortex, caudate nucleus, putamen, and thalamus, in the right temporal, parietal lobe, and cerebellum were noted. Surface-based morphometry identified reductions in the left orbitofrontal surface area in the right inferior/supramarginal parietal gyrus and the right inferior/middle temporal gyrus. The mean gestational age of the Bogota cohort was smaller. This may explain why voxel-based morphometry identified a much larger number of reduced clusters. Similarly to the Norwegian study, there were no changes in IQ but an impairment of several subtests of attention/executive functions using the Test-battery of Attentional Performance.

In contrast to the deficits identified, it is interesting to point out that the hippocampus volume deficits found initially seem to have disappeared when assessed at age 20 alluding to the known plasticity and neurogenesis of this structure. The normalization of hippocampal volumes might explain in part the preservation of IQ. Evaluation of white matter tracks in adults from the Norwegian study found significant white matter clusters still present following growth restriction most preeminent in the regions of the external and internal capsules as well as in uncinate fasciculus, inferior longitudinal fasciculus and inferior fronto-occipital fasciculus [56].

Having shown significant changes in brain metabolism in the fetus and impaired initial brain growth persisting in children and in adults, one may speculate about what will happen further along the line with aging (Fig. 1). A small historical cohort of 346 showed how higher head circumference at birth was associated with partial improvement of cognitive abilities [57]. The importance of head circumference was confirmed in a large epidemiological study: the Swedish Twin Registry born 1926–1960 with over 35 K individuals (ages 55–89) cross-referenced for dementia diagnosis. They discovered a 20% increased risk for dementia in newborns less than 2,500 g. Lower birth weight (independent of GA) carried an incremental 2% higher risk with each 100 g. When individuals had a small head circumference for gestational age, this increased the risk of dementia by 65% after controlling for multiples and confounding factors [58]. In a study of 215 individuals aged 66–75, birth head circumference was not shown as predictive of poor outcome, yet there was an interesting benefit of having a large head on cognitive assessment with a protective effect against decline in performance supporting the brain reserve hypothesis [59]. Similarly higher head circumference at birth was found in a cohort of 346 individuals to be associated at age 72–74 with cognitive subdomains [57].

These studies are challenging as they need large numbers in each category to be sufficiently powered. Even with a cohort of over 35 K, the Swedish Twin study had less than a thousand individuals with a small head circumference for gestational age [58]. Furthermore, the demographics of surviving children in 2023 have significantly altered compared to 70 years ago, increasing greatly the likelihood of major impairments of behavioral and cognitive outcomes in relation to poor fetal growth.

Obstetrical management of FGR is based mostly on serial ultrasound measurements of fetal biometry and Doppler studies to determine the best time to deliver. Achieving the right balance between intervention and the risk of prematurity versus expectant management and the risk of intrauterine death summarizes the role of an obstetrician. If delivery is indicated before 34 weeks, antenatal steroids for lung maturation and magnesium sulfate infusion for neuroprotection can improve neonatal outcomes [60]. There is not, however, a magic potion to promote fetal growth and improve placental insufficiency once the diagnosis is made. Aspirin, when started before 16 weeks in high-risk population, has showed to reduce the risk of early onset pre-eclampsia which is frequently associated with FGR [60].

Only a few and very different interventions to promote growth during pregnancy have been recently tested. In a cohort of over 8,200 participants, unconditional prenatal income supplement was shown to improve birth weight and reduce prematurity and improve neurodevelopment in First Nation children in Canada [61]. A randomized clinical trial that included 1,184 individuals with singleton pregnancies testing Mediterranean diet and mindfulness-based stress management showed that both interventions resulted in a significantly lower percentage of SGA newborns compared with usual care [61]. Protecting the fetus from tobacco [62], e-cigarettes [63, 64], and cannabis [11] exposure is self-evident. In a small cohort of 77 women with a known fetal growth restricted fetus, pomegranate juice [65] was found to improve brain microstructure in newborns, when compared with placebo. Postnatal nutrition strategies in IUGR are scarce and challenging [66, 67]. There is no clear data on therapeutic window width to date to maximize cognitive reserve in this population.

In utero growth restriction carries short, median, and long-term consequences. We need to invest more in the actual root causes analysis of growth restriction too often labeled as placental insufficiency. New population-based registry been built need to include more holistic data of the pregnancy and the neonatal period with inclusion of genetic, epigenetic and inflammatory biomarkers. Knowing the impact of growth restriction significant resources for brain reserve by programs to optimize pregnancy outcome and cognitive reserve by programs postnatally need to be implemented.

The authors have no conflicts of interest to declare.

The project was funded by the FRQSC/FRQS 2023-0PTA-322706 Action concertée (G.A.L.).

D.K.S., S.P., and G.A.L. wrote conceptualized, edited, and approved the submitted manuscript.