Article Text

Download PDFPDF
Prevalence and timing of pregnancy termination for brain malformations
  1. Caroline Rouleau1,
  2. Adeline Gasner2,
  3. Nicole Bigi3,
  4. Alain Couture2,
  5. Marie Josée Perez3,
  6. Patricia Blanchet3,
  7. Jean Michel Faure4,
  8. François Rivier5,
  9. Pierre Boulot4,
  10. Annie Laquerrière6,
  11. Ferechté Encha-Razavi7
  1. 1Department of Pathology, Lapeyronie Hospital, Montpellier, France
  2. 2Department of Paediatric Radiology, Arnaud de Villeneuve Hospital, Montpellier, France
  3. 3Department of Medical Genetics, Arnaud de Villeneuve Hospital, Montpellier, France
  4. 4Department of Obstetrics and Gynecology, Arnaud de Villeneuve Hospital, Montpellier, France
  5. 5Department of Neuropaediatrics, Gui de Chauliac Hospital, Montpellier, France
  6. 6Department of Pathology, Rouen University Hospital, Rouen, France
  7. 7Fetal and Placental Pathology Unit, Necker Enfants-Malades Hospital, Paris, France
  1. Correspondence to Dr Caroline Rouleau, Department of Pathology, Lapeyronie Hospital, 371 avenue du doyen Gaston Giraud, 34295, Montpellier, France; r-caroline{at}hotmail.fr

Abstract

Objective To determine the prevalence and the timing of pregnancy termination relative to the type of central nervous system (CNS) malformations.

Design Retrospective cohort study.

Setting Multidisciplinary centre for prenatal diagnosis in the Languedoc-Roussillon region, France.

Population A cohort of 481 pregnancy terminations performed between 2005 and 2009.

Methods Detailed post-termination fetal and neuropathological analyses were carried out to identify the CNS malformations. Then, the prevalence and timing of pregnancy termination were assessed relative to the identified malformations.

Results About one-third of pregnancy terminations (143/481) were performed for severe CNS malformations. Up to 24 weeks of gestation (WG), pregnancy terminations (56.6%) were carried out mainly for defects occurring during the two major first steps of CNS development (neurulation and differentiation of cerebral vesicles). After 24 WG, pregnancy terminations (43.3%) were mainly performed for corpus callosum agenesis (16/17), vermian agenesis (10/12) and gyral anomalies (13/15). For hindbrain malformations and gyral anomalies, there was a significant relationship between the timing of pregnancy termination and the presence of a severe ventriculomegaly at prenatal diagnosis (p=0.002 and p=0.02, respectively).

Conclusion By classifying CNS malformations according to the neuropathological analysis, the authors show that the timing and prevalence of pregnancy termination are distributed in a manner that is consistent with what is currently known on the development of brain. They are also influenced by the French prenatal screening policy and the variable expressivity of the brain malformations and associated lesions.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

The assessment of fetal brain malformations requires an accurate knowledge of the different steps and features of the developing central nervous system (CNS) during gestation.1 Indeed, compared with other organs, fetal brain continues to evolve during and beyond the fetal period, and the gestational age is crucial for the development of certain structures.2

Malformations of the CNS remain a major source of developmental delay and neurological deficits. During the last decade, progress in brain imaging (ultrasound (US) and MRI) led to the identification of an increasing number of CNS malformations. However, their prevalence and exact phenotype remain poorly documented, because case series on CNS malformations are rare and based on imaging data alone. Therefore, fetal and neuropathological studies are essential for determining the exact phenotype of the brain malformations and the underlying cause.1 3

What is already known on this topic

  • The exact phenotype of brain malformations remains poorly documented from a clinical and radiological point of view.

  • Brain malformations are identified by descriptive morphological terms such as microcephaly, lissencephaly, hydrocephaly and so on that do not necessarily correspond to the aetiology.

  • Fetal and neuropathological studies are essential for determining their exact phenotype and the underlying causes.

What this study adds

  • This study comes from a country where ultrasound is mandated in early, mid and late trimester.

  • The timing of pregnancy termination is distributed in a manner that is consistent with what is currently known on the development of the brain.

  • This could help indicate when certain diagnoses might be considered, in agreement with the French policy for prenatal screening.

In France, during pregnancy, fetal US screenings are mandatory at 12 (11–13), 22 (22–24) and 32 (32–34) weeks of gestation (WG) as well as serum screening between the 14 and 18 WG.4 In case of fetal abnormalities, parents are referred to a multidisciplinary centre for prenatal diagnosis (MCPD). Identification of CNS malformations by US leads usually to complementary investigations including new US, performed by experts in brain imaging, and also MRI if necessary. Termination of pregnancy (TOP) may be carried out after neuropaediatric and genetic counselling and, in accordance with French legislation, up to term. Postmortem study of the fetus is recommended and performed after parental consents.

Accordingly, the aim of this retrospective study was to investigate the prevalence and timing of TOP relative to the type of CNS malformation identified during pregnancy by US (and possibly MRI) and confirmed postmortem by neuropathological studies in a cohort of pregnancy terminations performed between 2005 and 2009 in the Languedoc-Roussillon region, France.

Methods

We assessed a cohort of 481 terminations of pregnancy, performed between 2005 and 2009 in the Languedoc-Roussillon region, France. This cohort corresponded to the total number of medical terminations of pregnancy after 12 WG, and all were validated by experts of the regional MCPD. Voluntary terminations, which are allowed in France until 12 WG, were not included in the study. This cohort covered the whole population of the Languedoc-Roussillon region (ie, 2 616 000 individuals, according to INSEE).5 During the last decade, the mean number of live births (LBs) registered in the region was 29 000 per year, and this number was taken as the denominator.

A total of 143 terminations were carried out for CNS malformations, in which the post-termination detailed fetal and neuropathological analysis, validated by experts in fetal neuropathology, permitted the identification of the exact phenotype of the CNS malformation and of the associated anomalies.

Over the studied period, seven families with prenatal diagnosis of brain malformations refused TOP and were excluded from the study. On the basis of prenatal imaging, these malformations were anencephaly (one case), spina bifida with myelomeningocele (two cases), brain malformations with severe and isolated ventriculomegaly (two cases), corpus callosum agenesis (one case) and lissencephaly (LIS, one case).

For statistical analysis, we used StaView software (StaView 512). Baseline variables were compared using the χ2 analysis for categorical data or, when appropriate, the Fisher exact test. The Mann–Whitney U test was used to compare medians of non-parametric variables. A p value of <0.05 was considered statistically significant.

Results

Between 2005 and 2009, 481 terminations of pregnancy were performed at the MCPD of the Languedoc-Roussillon region with a prevalence of 0.41/1000 LBs. Of these, 143 were carried out for CNS malformations (prevalence=0.92/1000 LB). The pathological findings as well as the time (in WG) of detection and the time (in WG) of TOP are summarised in table 1. Depending on the type of CNS malformation, they were classified in six distinct groups, including neural tube defects (NTD) (33.5%), midline anomalies (MDA) (29.3%), midbrain malformations (MBM) (5.5%), hindbrain malformations (HBM) (15.3%), growth and gyral abnormalities (GA) (9.7%), and clastic lesions (CL) (6.2%).

Table 1

Frequency, timing of prenatal detection and of pregnancy termination for central nervous system malformations

NTDs were observed in 48 fetuses and consisted in anencephaly (15 cases; prevalence of 0.97/10 000 LB) and open spina bifida (33 cases; prevalence of 2.1/10 000 LB).

In the case of anencephaly, TOP was performed on average at 14 WG. All were detected between week 10 and 14 of gestation during the first US screening.

Most cases (15/33) of open spina bifida were discovered during the second US screening, which is normally performed at about 22 WG. TOP for open spina bifida with myelomeningocele was carried out at a median gestational age of 22 WG (range 14–32 WG). Specifically, termination was performed earlier (median age: 20 WG) (p=0.001) in women with abnormal AFP levels above the cut-off level of 2.5 multiples of the median,6 whereas in women with no AFP elevation or who did not do the test it was performed later, at a median age of 24 WG (p=0.0005).

Midline anomalies were identified in 42 fetuses (29.3%). They corresponded to holoprosencephaly (HPE) (25 fetuses) and corpus callosum agenesis (CCA) (17 fetuses). For HPE, the prevalence of TOP was of 1.6/1000 LB, in agreement with previous reports.7,,9

Among HPE, the whole range of severity (ie, alobar, semilobar and lobar HPEs) was found. The timing of pregnancy termination was different according to the type of malformation. In the case of alobar and semilobar HPE (n=22), it was performed at a median gestational age of 17.5 WG (range 15–24 WG), and there was a significant relationship between the timing of termination and that of prenatal diagnosis by US (p=0.001). The anomaly was detected during the first US screening in 10/22 cases as the presence of a single ventricle.7 Lobar HPE was infrequent (3/25), and TOP was performed at a median gestational age of 27 WG (range 27–32). In these cases, the anomaly was suspected during the second US screening and validated by additional US and/or MRI. TOP for CCA (n=17; prevalence of 1.1/10 000 LB) was performed at a median gestational age of 32 WG (range 25–36 GW).

Most CCAs were detected during the second US (table 1) and required additional investigations such as MRI that were performed at a median gestational age of 30 WG (29–31 WG). In agreement with previous studies, the information given to couples conveyed a good prognosis in 60% of isolated CCAs and poorer prognosis for most of the associated CCAs.10,,12

Midbrain malformations (atresia or stenosis of the aqueduct of Sylvius) were found in eight fetuses (5.5%) with a prevalence of 0.52/10 000 LB. TOP was carried out at a median gestational age of 25 WG (range 19 to 37 GW). Most cases (6/8) were detected during the second US screening due to the presence of severe ventriculomegaly. One case was detected earlier at 17 WG and another one later during the third US screening. Neuropathological studies allowed the exact identification of the midbrain anomaly.

Hindbrain malformations were present in 22 fetuses (15.3%), and the prevalence of TOP for such malformations was 1.43/10 000 LB. They were identified as rhombencephalosynopsis (RBS) in 10 cases (prevalence of 0.6 /10 000 LB), vermian agenesis in 11 cases (prevalence of 0.7/10 000 LB) and ponto-cerebellar hypoplasia in one case.

In RBS, TOP was performed at a median gestational age of 24.5 weeks (range 20–27 WG). All had severe ventriculomegaly (posterior horn of lateral ventricle >15 mm) at the second US screening. In addition, three fetuses had open spina bifida, detected between 18 and 21 WG. In all cases, the diagnosis of RBS was histological.

TOP for vermian agenesis was performed at a median gestational age of 29 weeks (range 21–34 WG), significantly later than for RBS (p=0.002). In most cases, the posterior fossa anomaly was suspected at the second US screening and investigated at 26–28 WG by MRI. In one case, vermian agenesis was detected by US at 19 WG in a woman with a high risk of trisomy 21.

Growth and gyral anomalies were found in 14 fetuses (9.7%), and the prevalence of TOP for such anomalies was of 0.9/10 000 LB. Growth and gyral anomalies corresponded to micropolygyria (six cases), LISII (two cases), LISI (four cases) and microcephaly vera (two cases). Independently from their aetiology, gyral anomalies were characterised by partial (pachygyria) or total (agyria) absence of circumvolutions and/or sulci over the brain surface.13 In all cases, the final diagnosis was neurohistological, as previously reported.14

In the case of micropolygyria, TOP was performed at a median gestational age of 28 weeks (24–37 WG). Severe ventriculomegaly was detected during the second screening in two cases. In four cases, ventriculomegaly was found during the third US screening and was associated with micro-calcifications in two cases.

In the case of LISII, TOP was performed at a median gestational age of 25.5 weeks (range 23–28 WG). Both cases were detected during the second screening because of severe ventriculomegaly.

In the case of LISI, TOP was performed at a median gestational age of 32 weeks (range 32–36 WG). In two cases, gyral anomalies were suspected at the second US screening, because of an abnormal Sylvian fissure. Although this is considered to be a good predictive sign of LIS,13 both cases were investigated by MRI at 30–32 WG. In three cases, LISI was detected during the third US screening.

CLs were found in nine fetuses (6.2%). This represents a prevalence of 0.5/10 000 LB. CLs corresponded to porencephalies (four cases), hydranencephalies (two cases) and posthaemorrhagic ventriculomegalies (three cases). TOP was performed at a median gestational age of 27 weeks (range 23–34 WG). Most cases were detected during the second US screening, whereas three cases were detected only during the third US, although they presented severe ventriculomegaly.

Discussion

Between 2005 and 2009, about one-third of terminations of pregnancy carried out in the Languedoc-Roussillon region were performed following prenatal findings of severe CNS malformations, yielding a prevalence of 0.92/1000 LB. Of these, 81 (56.6%) were performed before or at 24 GW.

Except for one case of CCA and two of vermian agenesis, these terminations were mostly for NTD (anencephaly or spina bifida) (42/48 cases), HPE (22/25 cases) and malformations with severe ventriculomegaly (14/29 cases). Termination before 24 GW was thus due to anomalies (NTD and HPE) that appear early, as they result from defects occurring during the two major first steps of CNS development, neurulation and differentiation of cerebral vesicles.15 Because of the gravity of the disease, most parents opted for TOP. The prevalence of terminations of pregnancy for spina bifida was 0.21/1000 LB. However, we did not include in this group three fetuses with spina bifida that, following neuropathological examination, were identified as RBS. Our results partially agree with the analysis of the European Surveillance of Congenital Anomalies (EUROCAT), in which 85% of terminations for spina bifida were performed before or at 24 WG.16 Spina bifida was identified later (median gestational age of 23 weeks) than in our study (21 WG), in which the timing of TOP was correlated with the detection of abnormal AFP levels. Terminations of pregnancy for anencephaly were less frequent than in the EUROCAT report. It is possible that we have underestimated the prevalence of such terminations. Indeed, anencephaly was mostly detected during the first screening, and it is thus likely that several women chose to have a voluntary TOP. In France, this is allowed until 12 WG, and the procedure is simpler compared with a medical termination.17

Ventriculomegaly is a descriptive term which refers to an abnormal enlargement of the ventricle size and is severe when the occipital horn measures more than 15 mm by US.18 Although ventriculomegaly is often present in many brain malformations, in our series, it was the sole prenatal finding in 29 unrelated brain malformations that were finally diagnosed as RBS (10 cases), midbrain anomalies (eight cases), CLs (nine cases) and LISII (two cases). Owing to the timing of their detection, pregnancy termination was performed earlier for RBS and LISII than for other hindbrain malformations and gyral anomalies (p=0.002 and p=0.02, respectively). Our findings are similar to those of a French study, in which 77% of pregnancy terminations for RBS (40 fetuses in total) were performed in the second trimester.19 The neuropathological studies not only confirmed the presence and the extent of the ventriculomegaly but also established the final cause by revealing specific lesions affecting precise CNS areas. This was particularly the case for pregnancy terminations performed before 24 GW, when prenatal investigations of the CNS remain limited.

After 24 WG, 62 terminations of pregnancy were performed (43.3%). Except for severe ventriculomegaly (15/29 cases), NTD (6/48 cases) and HPE (3/25 cases), these terminations were mainly performed for CCA (16/17), vermian agenesis (10/11) and gyral anomalies (12/14).

Most CCA were detected during the second US screening (22–24 WG), as previously reported.20 21 But in contrast to these reports, where TOP for CCA (when decided) was performed before 24 GW, in our study, 53% of these terminations were performed after 31 GW. This discrepancy could be explained by the fact that, as the prognosis of CCA is conditioned to the presence of associated anomalies,9,,11 fetal brain MRI was systematically proposed. It was performed at a median gestational age of 30 WG (29–31 WG), a period that has been reported to be a good balance between gestational age and gyration/posterior fossa development.22 Similarly, although vermian agenesis can be recognised by US at 20 GW, most cases were investigated by MRI at a median gestational age of 26–28 GW. Indeed, prenatal diagnosis of vermian agenesis, especially in its partial form, relies on MRI findings, such as the identification of cerebellar fissures that may not be seen until 25–26 WG.23 Conversely, TOP for LISI and microcephaly vera was carried out at a median of 33.5 WG, which is later (p=0.02) than for LISII and micropolygyria (median age of 27.5 WG), owing to, in most cases, the late timing of detection. As the neurological prognosis is often poor, couples usually opted for late TOP.

In our series, most CNS malformations were isolated (119; 83%). Multisystem problems were present in 24 fetuses (16.7%), corresponding to chromosomal abnormalities (nine cases), syndromes (eight cases) and polymalformative complexes (seven cases). Renal dysplasia (10/24; 41%) and cardiac malformation (8/24; 33%) were the most frequently encountered non-CNS malformations.

Conclusion

The timing of TOP for CNS malformations varied considerably during gestation. By classifying CNS malformations according to the neuropathological findings, we show that the timing of TOP is distributed in a manner that is consistent with what is currently known on the development of brain. For defects occurring during the two first steps of CNS development (neurulation and differentiation of cerebral vesicles), terminations were performed early, as these defects gave rise to major malformations that were easy to detect. Conversely, for other developmental anomalies (microcephaly vera), although resulting from early defects (neuronogenesis anomalies), late terminations of pregnancy were performed, as their identification depends on the growth of the cerebral hemispheres during the second half of gestation. Late terminations (after 24 GW) were for anomalies that could not be identified earlier (LISI) or needed to be followed in time in order to better delineate their prognosis (corpus callosum agenesis). These finding could help indicate when certain diagnoses might be considered. Finally, these data show the complementarity between prenatal imaging and neuropathological studies in the diagnosis of brain malformations. The neuropathological studies not only confirmed the presence of the anomaly detected by prenatal imaging, but also revealed additional lesions which could not have been prenatally detected and allowed the identification of the exact phenotype.

Acknowledgments

We are indebted to E Andermarcher, for editing the manuscript. We thank F Seguret, for helpful discussions.

References

Footnotes

  • Competing interests None.

  • Ethics approval Ethics approval was provided by the Agence de la Biomédecine.

  • Provenance and peer review Not commissioned; externally peer reviewed.