Background
Despite undisputable benefits, midtrimester prenatal surgery is not a cure for myelomeningocele (MMC): residual intracranial and motor deficits leading to lifelong handicap question the timing of prenatal surgery. Indeed, the timing and intensity of intrauterine spinal cord injury remains ill defined.
Objective
We aimed to describe the natural history of neuronal loss in MMC in utero based on postmortem pathology.
Study Design
Pathology findings were analyzed in 186 cases of myelomeningocele with lesion level between S1 and T1. Using a case-control, cross-sectional design, we investigated the timewise progression and topographic extension of neuronal loss between 13 and 39 weeks. Motor neurons were counted on histology at several spinal levels in 54 isolated MMC meeting quality criteria for cell counting. These were expressed as observed-to-expected ratios, after matching for gestational age and spinal level with 41 controls.
Results
Chiari II malformation increased from 30.7% to 91.6% after 16 weeks. The exposed spinal cord displayed early, severe, and progressive neuronal loss: the observed-to-expected count dropped from 17% to ≤2% after 16 weeks. Neuronal loss extended beyond the lesion to the upper levels: in cases <16 weeks, the observed-to-expected motor neuron count was 60% in the adjacent spinal cord, decreasing at a rate of 16% per week. Progressive loss was also found in the upper thoracic cord, but in much smaller proportions. The observed-over-expected ratio of motor neurons was not correlated with the level of myelomeningocele.
Conclusions
Significant neuronal loss is present ≤16 weeks in the exposed cord and progressively extends cranially. Earlier prenatal repair (<16 weeks) could prevent Chiari II malformation in 69.3% of cases, rescue the 17% remaining motor neurons in the exposed cord, and prevent the extension to the upper spinal cord.
Myelomeningocele (MMC) is a form of neural tube defect characterized by the protrusion of the meninges and spinal cord through open vertebral arches. Cerebrospinal fluid (CSF) leakage through the defect is believed to be responsible of a decrease in hydrostatic pressure leading to the descent of cerebellar tonsils through the foramen magnum which define Chiari II malformation. Hydrocephalus is associated with Chiari II malformation in 85% of cases. Spinal cord injury and intracranial anomalies result in lifelong paralysis, incontinence, and cognitive impairment. Risk factors such as maternal folate metabolism anomalies and obesity have been associated with open neural tube defects; however, the pathophysiology of MMC remains poorly understood. Despite efforts to reduce MMC incidence using folic acid fortification, it remains at around 1 per 1000 live births. Typically, the diagnosis is made by routine midtrimester ultrasound, although MMCs can be diagnosed as early as the first trimester. In Europe, prenatal diagnosis results in termination of pregnancy (TOP) in >60% of cases owing to the burden of disabilities associated with the condition.
Why was this study conducted?
Midtrimester prenatal surgery is not a cure for myelomeningocele (MMC): residual intracranial and motor deficits leading to lifelong handicap question the timing of prenatal surgery. We aimed to determine the kinetics of spinal cord injury in utero in a case-control study based on a large postmortem population.
Key findings
Significant neuronal loss is already present ≤16 weeks of gestation in the exposed cord and progressively extends cranially: before 16 weeks, the exposed cord displays 17% of normal motor neurons and falls to ≤2% after 16 weeks.
What does this add to what is known?
Current surgical strategies might benefit from earlier repair to rescue motor function within the exposed cord. The development of new therapeutic strategies for early in utero repair could significantly improve neurologic and motor prognosis of these children.
Prenatal midtrimester intrauterine repair of the defect improves outcome, thus confirming that spinal cord injury is progressive during gestation. However, midtrimester repair is not a cure as most infants treated in utero remain severely disabled. In the era of developing prenatal therapies, the timing of the spinal cord injury is instrumental for optimizing such strategies and identifying new therapeutic targets.
Based on a large collection of postmortem cases following TOP, we aimed to describe the natural history of neuronal loss associated with MMC. Using a case-control, cross-sectional design, we investigated both timewise progression and topographic extension of neuronal loss along the spinal cord.
Materials and Methods
Study population
We retrospectively collected all postmortem cases of spinal defects following TOP and confirmed by pathologic examination from 1998 to 2018. A signed consent for fetal autopsy was obtained for all cases.
Among these cases of spinal defects, we selected those that met the following features of MMC: open vertebral arch, open arachnoidal sheet, and open ependymal canal. Furthermore, to match the eligibility criteria of most trials on prenatal surgery, we selected cases with a normal karyotype or array comparative genomic hybridization (array CGH) and an upper level of the lesion located between the first thoracic vertebra (T1) and the first sacral vertebra (S1).
The controls were fetuses resulting from miscarriage or TOP performed for defects not involving the central nervous system (CNS) with a normal karyotype or array CGH and for which a consent to postmortem research had been signed before TOP ( Supplemental Table 1 ). Cases and controls were classified into 4 groups of gestational age: ≤16 +0 weeks, 16 +1 to 20 +0 weeks, 20 +1 to 24 +0 weeks, and >24 weeks.
Study design
We aimed to compare the number of alpha motor neurons of the spinal cord in MMC fetuses with controls ( Figure 1 ). In cases with MMC, spinal cords were examined at 3 levels: within the exposed spinal cord; at the level adjacent above MMC, defined as the last level where the spinal cord is surrounded by a closed vertebra above the defect; and at a distant, upper thoracic level (T1–T2), considered only in MMCs below T10. In controls, spinal cords were examined at 5 levels: T1–T2, lower thoracic level (T11–T12), upper lumbar level (L1–L2), L3 (until 24 weeks), and below L3 (until 20 weeks).
We isolated L3 as a specific level because it corresponds to the lumbar enlargement of the spinal cord (ie, the lumbosacral plexus) with a high density of motor neurons at that specific level.
Because of the difference in the relative rates of growth of the spinal cord and vertebral column, the spinal cord is detected below L3 at 20 weeks and above L3 at 24 weeks in normal fetuses. Moreover, the vertebral column grows faster than the spinal cord, resulting in a relative ascent of the conus medullaris with gestation from S2 at 12 weeks to L2–L3 at birth.
Procedures
Termination of pregnancy
In France, TOP is permitted by law at any gestational age for severe fetal conditions such as MMC. Vaginal delivery is achieved after cervical ripening by mifepristone and overnight cervical dilators before induction. Intravaginal prostaglandins are administered following the placement of an epidural catheter. In cases >21 +0 weeks, feticide is performed by cordocentesis or intracardiac lidocaine injection.
Macroscopic examination
Fetal pathologic examination was performed according to the French national guidelines. All cases underwent full pathologic examination and photographs of all macroscopic findings including the MMC before dissection, x-ray examination, and visceral sampling. Cases with signs of advanced lysis of the brain or birth trauma were not considered for neuropathologic examination. Chiari malformation was noted when the brainstem, fourth ventricle, and caudal tip of cerebellar tonsils crossed the foramen magnum, both at craniotomy and by the shape of these structures after formaldehyde preservation.
Cases with descent of cerebral structures but without herniation through the foramen magnum were not considered as Chiari malformation.
Hydrocephalus/ventriculomegaly was defined as an enlargement of the cerebral ventricles at >16 weeks. Hydrocephalus is not evaluated before 16 weeks, given the physiologically large cerebral ventricles during early gestation. Foot deformities and calf muscle hypotrophy were also recorded. All macroscopic recordings followed a triple assessment of postmortem pictures by 3 different observers. Discrepancies among observers were excluded.
Spinal sampling
Within the exposed cord, 5 tiered samples were taken on average ( Figure 1 ). All specimens were immersed in a 4% zinc formalin solution (pH=7.4) for preservation, for 1 to 5 weeks depending on gestational age. Samples at >18 weeks were decalcified by 7% hydrochloric acid (Surgipath’s Decalcifier II, Leica Biosystems, Richmond, IL) for 30 to 180 minutes depending on gestational age. All samples were cut transversally and embedded in paraffin for histology.
Motor neuron count
For histologic analysis, 7-μm-thick sections of the spinal cord were stained with hematoxylin and eosin (H&E). For cell count, we selected MMC cases that met quality criteria related to spinal cord sampling on histologic analysis: samples strictly cut transversally and with a continuous open ependyma, a thick ventral horn, and an uninterrupted funiculus (behind the ventral horns). Cases that lacked one of these quality criteria were excluded for cell count. Selected H&E slides were scanned (Hamamatsu slide scanner, Hamamatsu, Japan) at ×40 magnification and visualized using NDP.view2 software (Hamamatsu Photonics, Hamamatsu, Japan). For each case, within the exposed cord, the slide with the largest amount of ventral horn was chosen for counting. All counts were conducted with H&E staining, and only alpha motor neurons were counted. A morphologically normal alpha motor neuron was defined as a large cell showing either a large pale differentiated euchromatic nucleus with a prominent nucleolus surrounded by a light cytoplasm, or a stellate shape with light cytoplasm holding regularly scattered intracytoplasmic basophilic Nissl bodies. , All motor neurons showing nuclear and/or cytoplasmic characteristics of injury were excluded from cell count. Such features included nuclear shrinkage and increased nuclear basophilia without distinguishable nucleolus (pyknosis) or cytoplasmic eosinophilia without Nissl bodies (chromatolysis) ( Figures 1 and 2 ).
Inter- and intraobserver reproducibility of cell count based on these criteria was assessed by 2 blinded observers with excellent intraclass correlation coefficient (ICC) of 0.995 (0.989–0.998) and 0.998 (0.994–0.999), respectively. In a subset of 15 slides, the motor neuron count was compared with immunohistochemistry using an anti-NeuN antibody (1/500, clone A60, MAB377, Millipore), showing an excellent ICC of 0.992 (0.976–0.997).
Data analysis
Every case count was divided by the mean motor neuron count on control samples matched for spinal level and gestational age: this ratio represents an observed ( Figure 1 ) to expected (O/E) ratio and is presented in percentage. Mean O/E ratios are plotted with error bars of 1 standard error of the mean. Motor neuron counts by gestational age group and spinal level in control are presented in Supplemental Table 2 . All analyses were conducted using R (The R Foundation for Statistical Computing, Vienna, Austria).
Data availability statement
Data sharing is not applicable for this article because no new data were created in this study.
Results
Description of the study population
After exclusion of cases with defects at >T1 or <S1, closed defects, and cytogenetic anomalies, 186 cases were included in this study ( Figure 3 ). Gestational age at TOP ranged from 13 to 39.5 weeks. A significant proportion of cases were terminated at an early gestational age with 21 (11.3 %) and 36 (19.4%) cases at ≤16 weeks and 16–20 weeks, respectively ( Table 1 ), whereas most cases were terminated at >24 weeks after midtrimester ultrasound diagnosis. The time between delivery and autopsy in MMC cases was 12 (8-16) hours. Apart from the expected CNS, spinal, and limb anomalies related to MMC, 15 of 186 cases (8.06%) presented associated anomalies including caudal dysgenesis (n=4), multiple malformations (n=4), and cleft lip and palate (n=3).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
