Background
The cranial ultrasound (CU) is a cost-effective, portable, non-invasive examination that does not require radiation. These features make it a valuable tool for assessing infantile brain structures [
1]. The CU can detect cerebral pathologies, like hemorrhages or ventricle system disorders, without sedation [
2]. Furthermore, early detection of these lesions ensures appropriate medical management and long-term assessments of neurodevelopmental disabilities [
3].
Currently, the Quality Standards Subcommittee of the American Academy of Neurology and the Practice Committee of the Child Neurology Science recommend routine CU screening for all neonates < 30 weeks of gestational age (weeks GA) [
3]. Beyond those recommendations, Leijser et al. claimed to include neonates that were admitted to the neonatal care unit (NCU) after birth [
4]. The American Institute of Ultrasound in Medicine (AIUM) Practice Guidelines for the Performance of Neurosonography in Neonates and Infants 2014 also recommend CU for all infants with indications (Table
1) [
5]. In Germany, neonatal screening for the auditory system, hip dysplasia, and metabolic diseases, but not CU, are included in regular check-ups (“U2/U3”) to ensure the early detection of treatable diseases with severe outcomes.
Table 1Established AIUM indications for a neonatal CU and corresponding variables as predictors measured in the SNiP study [
5]
Hemorrhage or parenchymal abnormalities in preterm and term infants | birthweight < 1500 g and < 30 WG [ 3], vacuum or forceps delivery [ 8], systemic hypertension or hypotension, or perinatal asphyxia [ 9] |
Hydrocephalus | signs (e.g., macrocephaly, curved fontanel, split cranial sutures) [ 10] |
Vascular abnormalities | prenatal diagnosis (e.g., vascular malformations, aneurysms) |
Possible or suspected hypoxic ischemic encephalopathy | symptoms (e.g., seizures, hypotonia, coma, respiratory distress), UA pH < 7.0, APGAR value of 0–3 at > 5 min [ 3] |
Congenital malformations | prenatal diagnoses (e.g., cysts, cerebellar hypoplasia) [ 11], chromosomal anomalies or malformations, or metabolic diseases [ 12] |
Congenital or acquired brain infections | mycoplasma, toxoplasmosis, cytomegalovirus, streptococcus |
Signs and/or symptoms of central nervous system disorders | symptoms (e.g., facial malformations, macrocephaly, microcephaly, intrauterine growth restriction) [ 5] |
Trauma | cephalo/subgaleal hematoma, subdural hematoma, subarachnoidal hemorrhage |
Craniosynostosis | craniosynostosis |
Previously documented abnormalities, including prenatal abnormalities | prenatal diagnoses (e.g., partial/complete agenesis of corpus callosum) |
Patients treated with hypothermia, ECMO, or other support systems | hypothermia, ECMO etc. |
Additionally, assessments of fetal brain structures and possible deviations could be conducted with a prenatal ultrasound. Like CU, this prenatal examination is not generally performed in Germany, but recently, it has been offered regularly in the study region in western Pomerania.
The population-based birth cohort study, known as the Survey of Neonates in Pomerania (SNiP), included a CU for all participants, irrespective of indication, gestational age, or inpatient admission [
6]. Here, we retrospectively reviewed those data. We aimed to determine the prevalence, severity, and outcome of cerebral conspicuities diagnosed in neonates, when there was no CU indication, according to conventional guidelines; i.e., we addressed the need for an extension of recent CU recommendations.
Discussion
Based on recent international guidelines, a neonatal CU is recommended for the following defined risk neonatal groups: aged < 30 weeks GA or an indication for CU [
3,
5]. Additionally, a proposal was made for neonates with admission to NCU [
4]. In this study, we re-evaluated these recommendations to determine whether an extension was warranted. Specifically, we addressed whether performing a CU was justified for all infants, irrespective of indications, based on the prevalence and severity of CU conspicuities and their neurodevelopmental outcomes. Our population comprised 3696 neonates that had no indication for CU and were not admitted to the NCU. We found that the cumulative incidence of a conspicuous CU finding in neonates without an indication or NCU admission was 3.2%, which was, as expected, lower than the incidences among neonates with indications (7.65%) and neonates with NCU admissions (7.7%). Significant anomalies, like corpus callosum hypoplasia, periventricular leukomalacia (PVL), and IVH II°, were not associated with a conspicuous neurological follow-up; thus, we lacked a correlation between conspicuous CU findings and neurological impairments [
8,
9]. Among the 117 neonates without an indication for CU, but with CU conspicuities, 85.5% had MCs and 14.5% had SCs. Ballardini et al. postulated a prevalence of 1.7% for conspicuous CUs in asymptomatic neonates aged > 37 weeks GA, and the distribution of MCs and SCs was similar to that in the present study [
10]. The prevalence of SCs was 0.33% in group II, which was higher than the prevalence of significant anomalies found by Ballardini et al. (0.19%) or that found by Wang et al. (0.25%) [
10,
11]. The inclusion of bilateral cysts as an SC yielded a higher prevalence in our study. Even higher prevalences were reported by Heibel et al. (9%) and Gover et al. (11.2%). Those values might be explained by differences in operators, populations, and/or techniques [
8,
12]. None of the neonates with bilateral cysts had positive microbiological blood tests. Therefore, this study could not demonstrate that the probability of congenital infections increased with the presence of bilateral parenchymal cysts [
8].
Most MCs (93%) in group II comprised ventricular asymmetry, mild ventricular enlargement, or unilateral singular cysts. Of these MCs, 13% (
n = 13/100) were associated with later neurological impairments, like attention deficit (and hyperactivity) disorder, speaking/language disorders, or epilepsy. This finding did not suggest that the guidelines should be extended for these diagnoses, because, for example, single cysts had a negligible likelihood of leading to neurodevelopmental impairment, and thus, they did not require follow-up [
8].A large proportion of SCs (76%) in group II comprised bilateral or multiple cysts. These cysts can be diagnosed easily with CU. One in 4–5 neonates with bilateral, multiple choroid plexus cysts or subependymal pseudocysts harbor a chromosomal anomaly, like trisomy 18, trisomy 21, Cri du chat syndrome, or Zellweger syndrome, or a congenital infection (cytomegalovirus or rubella virus) with the highest positive likelihood ratio [
8]. Multiple cysts, irrespective of location, can be associated with, for example, IVH and PVL. In addition, multiple cysts can lead to further malformations that have an impact on neurodevelopment [
13]. In this study, among patients in group II with SCs that represented bilateral cysts, 11.8% (
n = 2/17) showed neurodevelopmental delay or speaking disorders; the cumulative incidence was 0.05% (n = 2/3696) in all neonates without a CU indication or NCU admission. None of the named associated diseases or malformations were found.
PVL typically appears in very premature infants, but it is also detected on a CU or magnetic resonance imaging (MRI) at an unknown incidence in late preterm and term infants [
13]. PVL grades 1 to 3 exhibit (1) an abnormally high signal intensity, (2) loss of periventricular white matter, or (3) a necrotic, diffuse component, which can lead to cysts, focal scars, and global delay in myelination. These features were correlated with cerebral palsy and cognitive or behavioral abnormalities [
13,
14]. Among children with PVL grade 1, 56% might also have minor motor problems. An association with West syndrome, which presents with relapsing seizures, increases the probability of disorders in brain development and psychomotor functioning [
15]. In this study, PVL was not associated neurological impairment, and the cumulative incidence of PVL in group II was low (0.03%).
In the general population, the incidence of partial or complete agenesis of the corpus callosum is 0.3 to 0.7%, and it can be detected with CU [
16]. The outcome of partial agenesis of the corpus callosum spans a large range, from mild behavioral problems to severe neurological disorders. Compared to complete agenesis, a partial agenesis is associated with a higher probability of combined genetic syndromes, aneuploidies, further cerebral malformations, inborn errors of metabolism, and extracerebral malformations [
17‐
19]. .Compared to the intelligence of the general population, children with isolated corpus callosum agenesis have a higher probability of borderline intelligence and a lower probability of a Full Scale Intelligence Quotient ≥100 [
20]. In this study, we found a cumulative incidence with isolated corpus callosum agenesis of 0.03% in group II, similar to that reported by Wang et al. (0.08%), but we found no correlation with neurological impairment [
11].
IVH ≥ II° can induce epileptic disorders, perceptual difficulties, cognitive deficiencies, and mental handicaps [
8,
14,
21]. In particular, IVH can induce a worse outcome in full-term neonates than in preterm neonates [
14]. Even a mild IVH alone can independently have an impact on neurological outcome, compared to no IVH [
22]. .We found a 0.05% cumulative incidence of IVH in group II, which was lower than the incidences found in studies by Gover (0.2%) and Wang et al. (0.1%), and we found no association with neurological impairment [
11,
12].
Our results suggested that it is not necessary to extend the recent CU guidelines for neonates without a CU indication. We found an existing, but low cumulative incidence of mild and significant anomalies; thus, we could not demonstrate a correlation between ultrasound findings and neurological follow-up. Neurological development disorders can be detected during the regular, continual pediatric check-ups in Germany. Nevertheless, neonatologists, follow-up-clinicians, parents, and teachers must be cautious in assessing infants with the named pathologies, because additional resources and sustained rehabilitation might be necessary to maximize the child’s potential [
20,
22].
Our study findings emphasized the necessity of conducting CUs in cases specified in the recommended guidelines. Indeed, we found that all neonates < 30 weeks GA had an indication for CU; the cumulative incidence was 34.9% for a CU finding, and the probability was 20.9% for neurodevelopmental disorders. Among neonates with a NCU admission (7.7%) the highest cumulative incidence for SCs was observed (2.8% SCs), and of these, 16.4% displayed neurodevelopmental disorders. Almost every 5th neonate born at 30–34 weeks GA had a CU conspicuity. This finding was consistent with findings by Ballardini et al., who suggested that CU should be performed in all neonates born at ≤34 weeks GA [
9].
The major strength of this study was the large number of neonates (n = 5109) with data available from the population-based prospective cohort study, SNiP, including 3696 neonates without a CU indication The analysis included neonates at ages that ranged between 22 and 43 weeks GA, which exceeded the age ranges studied in previous investigations.
Although the SNiP database comprised 250 variables, data on follow-up were not included; consequently, those data were collected retrospectively. The major limitation of this study was the 25% loss to follow-up in the most interesting group of neonates, which had no indication for CU. The nature of the retrospective study design for the follow up and reevaluation of sonographies impeded a detailed assessment of developmental disorders; therefore, we might have missed some neonates with conspicuities in the CU that later displayed developmental disorders, which is a limitation. The increased diagnostic accuracy of prenatal ultrasound in the last decade can also increase the amount of prenatal diagnosed abnormal cerebral findings in future. Furthermore, any unusual CU findings, including normal variations, were documented as conspicuities in the SNiP database. To eliminate normal variations and detect missed conspicuities, we performed a retrospective re-evaluation of the scanned CU images; this re-evaluation might have resulted in some unrecognized pathologies.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.