Introduction

Ventriculomegaly (VM) is one of the most frequently diagnosed antenatal brain anomalies with a prevalence from 1.48 up to 22 per 1000 births [1, 2]. It is defined by a dilatation of the posterior foetal lateral ventricles and is commonly evaluated during the assessment of the foetal anatomy between 20 and 22 weeks of pregnancy. Even though there is no determined description of VM, the widely accepted definition is a dilatation of > 10 mm [1]. The leading reasons for a foetal VM are spina bifida, agenesis of the corpus callosum, Dandy-Walker malformation, stenosis of the aqueduct and Chiari II malformation [2]. Furthermore, it is associated with foetal infections, as well as with chromosomal and structural anomalies [3]. Severe VM is related with a poor prognosis [4, 5]. Moreover, a positive correlation was found between the degree of neurodevelopmental delay and the width of the lateral ventricles [6]. Some studies indicate that the prognosis mainly depends on the aetiology and the existence of associated anomalies [4, 7, 8]. If the foetus has no other abnormalities detected by prenatal ultrasound, the VM is generally considered as “isolated” ventriculomegly. The mortality rate is lower, prognosis is better and neurodevelopmental outcome is superior in isolated VM than in presence of associated malformations [6, 9, 10].

There are many foetuses diagnosed with an extension of the cerebral posterior lateral ventricle of 8–10 mm, which is often referred to as prominent or borderline VM. The significance of this finding, however, is not well known. Borderline VM can be interpreted as an abnormal finding and can be associated with chromosomal or structural cerebral anomalies or can just be interpreted as a powerless discovery without any pathological significance [11]. Counselling of the parents therefore is very challenging. The aim of this study was to assess cases with VM between 8 and 15 mm, with a main focus on cases with borderline VM between 8 and 10 mm that were isolated on prenatal ultrasound. We aimed to describe the prenatal course and management and assess short- and long-term development at the age of 2 years.

Methods

This is a 5-year retrospective, single-centre data analysis. The study was evaluated and approved by the local ethics committee (EKNZ 2015-447). We performed a Viewpoint data base search (GE healthcare, Version 5.6.22.30). According to local regulations, informed consent was not required for this retrospective study. All children were included, who were prenatally diagnosed by ultrasound with a ventriculomegaly (VM) at the University Hospital in Basel and were born between June 2008 and August 2013. The lateral ventricles were evaluated with an axial perspective on the foetal head, on the transventricular level, at the position of the glomus of the atrium and by measuring the ventricular cavity between the echoes produced by the lateral walls. The ultrasounds were conducted by experienced certified senior ultrasound specialists, with Voluson E8 and Voluson Expert, GE US machines. Borderline VM was defined as 8.0–10 mm, mild VM 10.1–12.0 mm, moderate VM 12.1–15.0 mm and severe VM from 15.1 mm. The prenatal management as well as data after birth was collected from the patient’s files. In the time period from June 2008 to August 2013, 159 children with apparently isolated borderline VM were included, from which the following information was collected: foetal MRI conducted (yes/no), a test for toxoplasmosis, rubella, cytomegalovirus and herpes simplex virus (TORCH)—testing (yes/no), progression of the VM during pregnancy (yes/no), delivery outcome and karyotyping (yes/no). A progression of the VM was defined as an increase of the ventricle width of ≥ 5 mm during the pregnancy and a regression was determined ≤ 5 mm.

From 159 children diagnosed with isolated borderline VM, 106 were born at our institution.

From all children born at our institution, we assessed the following parameters: week of pregnancy, birthweight, APGAR scores, sex, neurologic findings, transfer to the neonatal unit, postnatal cerebral ultrasound or MRI and neuropediatric consultation. In a second step, a questionnaire on the neurologic development at the age of 2 was sent out to the paediatricians in charge. The paediatrician was known in 84.9% (n = 90/106). Motor skills, cognitive behaviour, language development, social behaviour, growth charts and relevant neurologic interventions or diagnoses (cerebral palsy, shunt insertion) were evaluated.

Results

Two-hundred-fifty children with VM were included (Fig. 1). Among these, 78.8% (n = 197/250) had borderline VM.

Fig. 1
figure 1

Overview of the study population

Full size image

Out of these, 19.3% (n = 38/197) showed associated anomalies and two of them (5.2%; n = 2/38) had chromosomal abnormalities (one case of trisomy 18 and one case of monosomy x).

A total of 159/197 (77.1%) were apparently isolated cases (Fig. 1). Outcome at delivery was available for 153 cases. A total of 11.1% (n = 17/153) underwent prenatal invasive diagnostic testing. All of these showed normal karyotypes. A total of 3.3% (n = 5/153) underwent a prenatal MRI confirming VM without additional findings, 15.7% (n = 24/153) had TORCH (Toxoplasmosis, rubella, cytomegalovirus, herpes simplex virus 2 and other infections) testing, 4.6% (n = 7/153) had a positive IgG and IgM for Borrelia (n = 3), HSV (n = 3) or Parvovirus B19 (n = 1) (Table 1). Two cases were diagnosed with trisomy 21 postpartum (1.3%; n = 2/153). Both of these children showed an increased risk in first trimester screening but the patient denied further testing and the foetuses showed no additional anomalies on follow-up ultrasound exams apart from borderline VM.

Table 1 Prenatal findings (n = 153) of cases with isolated borderline VM. This table shows the prenatal course of these children until birth. Data is presented as median ± range or as number (n) and percentage (%)
Full size table

A total of 21.6% (n = 33/153) women presented for follow-up ultrasound exams during pregnancy. Of these, 33.3% (n = 11/33) had a progression of the VM. Actually, 12.1% (n = 4/33) developed mild VM. The lateral ventricles width decreased in 21.2% (n = 7/33) but all of the cases remained above 8 mm. In 45.4% (n = 15/33), the width of the lateral ventricles was stable (Fig. 2).

Fig. 2
figure 2

Development of VM during pregnancy (n = 33) in cases of isolated VM. A progression of the VM was defined as an increase of the ventricle width of ≥ 5 mm and a regression was determined as a decrease of ≤ 5 mm

Full size image

All but one of the 153 children with an isolated borderline VM were born alive. There was one early induction in 23 gestational weeks because of maternal complications (preeclampsia). A total of 20.9% (n = 32/153) had a premature delivery (< 37 weeks of pregnancy), from which 40.6% (n = 13/32) were twins and 12.5% (n = 4/32) had intrauterine growth restriction (IUGR).

Hundred-six children had a follow-up after delivery (Table 2). Sixty-nine percent (n = 74/106) were boys.

Table 2 Perinatal outcome (n = 106) of cases with isolated borderline VM. Data is presented as median ± range or as number (n) and percentage (%). The child with abnormal findings on postnatal ultrasound also is mentioned among the irregular neurological exam postpartal
Full size table

The median gestational week at delivery was 38 (27–41 weeks). Median birth weight was 3270 g (860–4660 g). The median APGAR after 1, 5 and 10 min was 8 (1–10), 9 (4–10) and 10 (6–10) respectively. 1.9% (n = 2/106) were transferred to the neonatal unit due to VM. A total of 21.7% (n = 23/106) had a postnatal brain ultrasound and one (0.9%) showed abnormal findings, i.e. haemorrhagic hydrocephalus after a delivery at 29 weeks.

A total of 2.8% (n = 3/106) had a postnatal MRI. A total of 1.9% (n = 2/106) showed mild neurologic abnormalities such as a reduced neonatal muscular tonus with the haemorrhagic hydrocephalus and another child showed a pseudo microcephaly most likely due to a forced intrauterine position. None of them was referred to or initially evaluated by a neuropediatric specialist.

Follow-up at 2 years was available for 40 children (Table 3). The questionnaire response rate was 44.4% (n = 40/90). A total of 27.5% (n = 11/40) had abnormal findings in the four categories of motoric skills, cognitive behaviour, language development and social behaviour. In our study, 15.0% (n = 6/40) of the children had slight motoric delays and 15% (n = 6/40) of them had social behavioural problems, such as, for example aggressions in nursery school. A total of 2.5% (n = 1/40) had cognitive behaviour difficulties such as attention deficit disorders and 20% (n = 8/40) of them had speaking troubles in form of delays.

Table 3 Two-year follow-up (n = 40). Results of the questionnaire at ≥ 2 years of age. Abnormal findings in questionnaire were evaluated in the categories: motor skills, cognitive behaviour, language development and social behaviour. Data is presented as median ± range or as number (n), percentage (%) and p. = percentile
Full size table

Considering the height, 30.0% (n = 12/40) were above the 90 percentile and 5.0% (n = 2/40) were below the 10 percentile. Infant weight was above the 90 percentile in 30.0% (n = 12/40). A total of 27.5% (n = 11/40) had head circumferences above the 90 percentile and 5.0% (n = 2/40) had head circumferences under the 10 percentile. There was no case of cerebral palsy, but one child had an insertion of a shunt (2.5%). In the prenatal period, this child had no associated anomalies, the borderline VM increased, but stayed below 10 mm. After delivery at 29 weeks, it showed neurological abnormalities and had cerebral haemorrhage, most likely due to premature delivery. After 2 years, the child showed a retardation in language development.

A total of 7.5% (n = 3/40) of the children had other neurological diagnoses (familial megalocephalia and seizures with status post-meningitis) and 7.5% (n = 3/40) had additional different diagnoses (exotropia and hyperopia, recurrent bronchitis and recurrent otitis, supraventricular extra systoles with an atrioventricular block).

Discussion

This is a large retrospective population-based study of borderline VM over a 5-year period. Borderline VM was defined as a width of the lateral ventricles of 8.0–10.0 mm. Although the common definition of mild VM is considered from a width of 10 mm, we commonly have referrals with borderline cases below 10 mm. Data in the literature regarding this situation is scarce. We here present data on the progress of this type of VM during pregnancy, data directly after birth and unique long-term information on the outcome of the children at the age of 2 years.

A total of 250 children were diagnosed by prenatal ultrasound with VM during a 5-year period. Of these, 78.8% were identified with borderline VM. Almost one fifth of the children diagnosed with borderline VM had associated malformations and around 5% were diagnosed with chromosomal abnormalities. This is in line with data from recent meta-analyses on mild VM 10–12 mm which reported 4.7% abnormal karyotypes [12]. The rate of associated anomalies was higher at 33.5% [12] which would correlate with the findings that moderate VM has a significantly higher association with anomalies than mild VM [7, 8, 13]. Prenatal therapies have been described for the in utero-treatment of menigomyelocele and have shown that prenatal surgery can be of benefit for selected patients [14]. In utero-treatment for hypertensive-hydrocephalus by ventriculo-amniotic, shunting has been described in the literature with a benefit in selected cases, but the technique is challenging and prospective clinical trials are lacking [15]. From our data, we can support referral for a detailed prenatal ultrasound for ruling out other abnormalities and to find the cause of VM, which could have therapeutic consequences for selected foetuses.

Foetal infections such as TORCH may cause ventricular dilatation. Numerous studies demonstrate that a single manifestation of prenatal cytomegalovirus (CMV) infection may be a mild VM [16, 17].

In our study, 16% had TORCH testing, whereas 4.7% showed positive IgMs for different types of infections. There was no case of CMV infection. None of the mothers had clinical signs of infection and IgG or IgM levels did not increase on follow-up. So, most likely, these were not cases with an acute infection. In our study, TORCH testing leads to follow-up serologic exams and no proof of maternal infection. Our data does not support TORCH testing for cases of isolated borderline VM in the absence of clinical symptoms.

This is supported by other data recommending targeted CMV testing instead of TORCH screening [18, 19]. However, it might be discussed in individual cases in patients which are at risk for these infections or in foetuses with associated anomalies [18, 19].

Concerning singleton pregnancies with normal intrauterine growth, 10% of our cohort had a premature delivery, which is in line with data from the World Health Organisation (WHO) showing premature delivery rate of 9.0% in higher-income countries [20].

Prenatal MRI was not conducted in the majority of the cases as it was not used systematically. Depending on the study, the rate of newly discovered diagnoses through prenatal MRI is varying between 5 and 66% [19,20,21,22,23,24,25,26]. The increased amount of other findings after a specialist sonography, however, has been shown to be low and does not recommended MRI as a standard [25,26,27,28,29].

Of the women presenting for follow-up exams, 21.1% had an increase of the width of the lateral ventricles and developed mild VM (> 10.1 mm). None of them developed severe VM. which has been also shown in other studies [30, 31]. The exclusion of progression by periodical follow-up seems reasonable.

Two cases were diagnosed with trisomy 21 postpartum (1.3%; n = 2/153). Both of these children showed an increased risk in the first trimester screening but denied further testing and the foetuses showed no additional anomalies on follow-up ultrasound exams apart from borderline VM. This shows the importance of offering a karyotyping even to isolated cases of borderline VM at presentation if first trimester screening was not performed.

After delivery, two children were transferred to the neonatal unit due to the VM. One child had asymmetric ventricles and the other one was diagnosed with hydrocephalus after premature delivery. Only one MRI showed abnormal findings: a child with a haemorrhagic hydrocephalus. Two showed mild neurologic abnormalities after birth, which included a child with pseudo microcephaly due to false position intrauterine and again the child with the increasing haemorrhagic hydrocephalus who showed reduced tonus. The increasing hydrocephalus was caused by the bleeding due to premature delivery rather than the VM being the underlying cause. None of the children was referred to a neuropediatrician. So, apart from one child, none of the children had clinically relevant findings.

Long-term follow-up after 2 years was available for 40 children. A total of 27.5% had head circumferences above the 90 percentile and 5.0% had head circumferences under the 10 percentile, explaining that large ventricles correlate with a large head.

A total of 27.5% had findings in the four categories of motor skills, cognitive behaviour, language development and social behaviour as rated by the paediatrician in charge.

According to the KiGGS study (“Studie zur Gesundheit von Kindern und Jugendlichen in Deutschland”), 19.3% of the children at the age of 3–6 showed values above the normal range in the categories: emotionality, behaviour, hyperactivity, contacting and prosocial behaviour [32, 33]. There are additional studies investigating of early childhood development.

In Great Britain, the prevalence of psychological disorders was illustrated to be 22.3% at the age of 3 and in Germany 18.4% at the age of 3–14 years [32,33,34,35,36,37,38]. Since the number of patients in our study populations are comparatively small and the findings described by the paediatricians are not considered as diagnoses that need interventions, we assume that our findings are the reflection of the normal limits of early development and rather not findings that are linked to borderline VM. Here, there are some limitations to this study, which would be necessary for a definite interpretation such as the retrospective study design and the limited number of patients in the follow-up at 2 years.

Mostly important, there was only one case of shunt insertion and no case of cerebral palsy. This concludes that only 0.6% of the children had evident consequences at the age of 2 in cases of isolated borderline VM and the reason for that most likely was premature delivery. Our data would support the data from Signorelli who propose that an atrial width of 10–12 mm without abnormalities can be interpreted as a normal finding, since all children in their study (n = 38) with a width of 10–12 mm had a normal neurodevelopment up until 10 years after birth [11]. This leads us to the hypothesis that in the absence of anomalies, a width of the lateral ventricles of 8.0–10.0 mm is a variation of the norm.

Parental reassurance seems to be advisable in cases of borderline VM when a detailed ultrasound reveals no additional anomalies and first trimester screening showed low risk for chromosomal abnormalities. However, due to the two cases of trisomy 21 in our cohort, an isolated VM at 20 weeks in the absence of first trimester screening should prompt counselling on invasive testing or NIPT.

Based on our analysis, we suggest to set a cut-off for ventriculomegaly at the width of 10 mm or more and to describe a posterior lateral ventricle under 10 mm as a variation of the norm in the absence of associated anomalies. We therefore suggest that ventricles between 8 and 10 mm are evaluated by an ultrasound specialist to check for associated findings. Posterior lateral ventricles between 8 and 10 mm without associated abnormalities in the second or third trimester should lead to parental reassurance. Additional work up such as MRI and TORCH testing did not lead to clinical relevant additional information in our cohort of borderline VM but should be discussed with the parents as an option.

Even though isolated borderline VM showed favourable outcome in most of the patients, we nevertheless would recommend follow-up with ultrasound on a monthly basis for in utero progression to exclude the development of severe hypertensive hydrocephalus (≥ 15 mm) which would require interdisciplinary evaluation with prenatal specialists, obstetricians, neonatologist and paediatric neurosurgeons to develop a management plan.