Extremely premature infants (born ≤28 weeks of gestation [GA], body weight <1000 g) are a high risk group for abnormal physical and psychomotor development. Among the major problems requiring detailed assessment are: growth, respiratory and vision disorders, motor and cognitive development. This is due to the fact that severe neurodevelopmental complications (cerebral palsy and cognitive developmental delays) that appear between 2 and 3 years of age, affect 12-18% of this population. Neurodevelopmental and psychological disorders, such as behavioral, cognitive and emotional disturbances, appear until school age, and even beyond. Autism spectrum disorder is observed in these children over 10 times more oThen than in the population of term neonates (8% vs. 0. 6%). Vision problems (including myopia, strabismus), especially in children with retinopathy of prematurity, affect between 20 and 50% of this population. This is indicated both by world research and the few studies conducted in Poland [1, 2, 3, 4, 5, 6, 7].
In addition to clinical examination, a significant role is assigned to neuroimaging studies in evaluating abnormalities and predicting prognosis in this population of children. There is no doubt that cranial ultrasound is and will remain the primary diagnostic tool as a non-invasive, easily accessible and inexpensive technique. However, it has significant limitations, so nowadays magnetic resonance imaging (MRI) is the best way to visualize the brain and the necessary technique for evaluating changes within it. MRI allows the visualization of the posterior fossa structures (cerebellum, brainstem), of the cortex and deep structures of cerebral hemispheres, including the thalami and basal ganglia, as well as for the assessment of myelination. Meta-analysis of the first world studies seeking correlations between ultrasound examinations performed during hospitalization, brain MRI in the 40th week of postconceptional age and neurological examination evaluating the long-term development of extremely prematurely-born babies indicates the important role of MRI in determining prognosis [9, 10, 11, 12].
With regard to cognitive development and functioning, extreme prematurity is a risk factor for learning dificulties, even in children with a normal level of intellectual development [13]. Possible and coinciding causes of these dificulties are deficits in executive functions [14], phonological processing [15], but also attention deficit hyperactivity disorder (ADHD), predominantly the inattentive type. Internalizing disorders, especially anxiety disorders, also play an important role [15]. In addition, an eightfold increase in the risk of autism spectrum disorders (ASD) [16] has been observed in extremely prematurely born infants compared to the term infant population. According to the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5), ASD is a complex neurodevelopmental disorder covering a broad range of deficits and symptoms in two major domains: 1) persistent deficits in social interaction and communication across multiple contexts, 2) restricted, repetitive patterns of behavior and activities, as well as unusual interests.
Among potential causes or factors correlating with ASD in this population (apart from genetic predispositions and primary and secondary infections), cerebellar and thalamic abnormalities are generally noted, which are visible on MR imaging only [17]. In addition, synaptogenesis dysfunction is also considered in the etiopathogenesis of various disturbances (such as emotional or cognitive disorders) in extremely preterm infants and is impossible to detect on ultrasound.
Due to the increased risk of mental disorders, including autism spectrum disorders, it seems important to include these infants in early psychological developmental assessment and psychiatric screening, which will allow early diagnosis and implementation of treatment [18, 19]. There are two types of screening tools for the detection of ASD: level 1 and level 2 screening measures. Level 1 screeners are designed to distinguish children with ASD from typically developing peers. In contrast, level 2 screeners serve to differentiate between the risk of ASD and other developmental disorders. A good screening tool for autism spectrum disorders is still sought for, as the most commonly used the M-CHAT test (a level 1 screener) gives many false-positive results in preterm infants [20, 21]. STAT (Screening Tool for Autism in Toddlers and Young Children, Stone et al., 2000; 2004 ), as a level 2 screening tool, has been recommended by the American Academy of Neurology (Filipek et.al., 2000). STAT is an interactive measure, conducted in a playful and joyful manner, and it only takes 20 to 30 minutes to administer. An important feature of this screener is that it can be used by a wide range of community specialists (psychologists, paediatricians, etc.). Among the few longitudinal, follow-up studies of premature infants’ development that have been carried out in Poland, none, to our knowledge, attempts to investigate the possible interconnections between neuroimaging results and neurodevelopmental outcomes, including the assessment of risk for autism spectrum disorder.
The aim of the study was to analyse the relationship between brain imaging results (US and MRI) and neurological, psychological and psychiatric development evaluations of two-year-olds born extremely prematurely.
The study population were infants born ≤28 GA and hospitalised at NICU between June 1, 2013 and December 31, 2015. We present the results of the first 12 children who were admitted to the study at the age of 2 years (corrected age) and had undergone both neuroimaging, and neurological, psychological and psychiatric assessments.
Cranial ultrasound was performed early (3-14 day of life, the number of tests varied, depending on the diagnosed pathology) and late (around 38-40 postconceptional weeks). Sonographic examinations were performed using two ultrasound systems: Aloka ProSound Alpha 7 (linear transducer 4-10 MHz, convex transducer 4-11 KHz) and Philips iU-22 (transducers 5-12 MHz and 5-8 MHz, respectively). Anterior and mastoid fontanelles were used.
Magnetic resonance imaging was performed between 38 and 42 postconceptional weeks with a 1.5T scanner (GE Signa HDxt) using an “adult” sixteen-channel head coil and since the second half of the year 2013 − in the dedicated neonatal eight-channel, phase-array head coil in the MR-compatible incubator (Nomag IC 1.5, Lammers Medical Technology GmbH; INC). MRI protocol included:
Coronal T2-weighted images (T2WI),
Axial T1-weighted images (T1WI),
Axial SWI sequence and GRE/T2*WI,
DWI sequence,
Sagittal T2WI,
Axial T2WI,
Axial FLAIR sequence,
Coronal FSPGR/3D/T1WI.
MR venography, arteriography, and other additional sequences, including post-contrast imaging, were performed according to current needs with an individual approach to every patient.
At 3-6-9-12-18-24 months of corrected age paediatric and neurological evaluation was performed in addition to which psychomotor development assessment was carried out. This was based on the Denver scale and on selected elements of Prechtl’s spontaneous movement assessment at 3 and 6 months. Important medical events that may have influenced CNS development and the time of the introduction and the scope of early stimulation and rehabilitation were recorded. Children were classified into three groups: normal, abnormal, for further observation.
At 2 years of corrected age, psychological assessment of psychomotor development using
At 2 years of corrected age the level 2 screening test STAT (Screening Tool for Autism in Toddlers and Young Children) was administered by a psychiatrist, which was – to our knowledge − the first time that this tool had been used in extremely premature children in Poland. STAT is a direct, play-based assessment of a child’s social-communicative behaviours with a special focus on directing attention, requesting, reciprocal and pretend play, and motor imitation. High specificity, sensitivity, and predictive validity has been reported in 24 to 35-month-old children with developmental delays, later diagnosed with ASD. A score of 2 points and above is considered as a risk for ASD.
A thorough ophthalmological examination was carried out according to the current scheme during hospitalization; laser photocoagulation treatment and/or intraocular Lucentis has been used in case of advanced forms of retinopathy. An ophthalmological assessment was then performed until the age of two years.
A screening hearing test was performed during hospitalization and then an ABR hearing test at the age of a few months.
Data from the pregnancy period and characteristics of patients are presented in Table I. Gestational age ranged from 23 to 27 weeks of gestation, and birth weight from 580 g to 1100g. The high prevalence of prenatal corticosteroid treatment (83%) and delivery by caesarean section (83%) was noted. Corticosteroid treatment was required in 50% of cases during hospitalization, due to the long duration of mechanical ventilation (mean 22 days). 25% of the children were diagnosed with moderate or severe bronchopulmonary dysplasia (BPD). As many as 75% of the children have suffered from retinopathy of prematurity (ROP), requiring treatment (intraocular Lucentis or laser photocoagulation). Periventricular infarction was diagnosed in 3 children (25%), accompanied by a third degree of bleeding in 2 of them (in 1 case with posthaemorrhagic ventricular dilatation).
Characteristics of the study population.
Tabela I. Charakterystyka ogólna badanej populacji.
Population characteristics |
Infantsincluded in analysis |
---|---|
Gestational age (GA) weeks: median (range) |
26(23+6;27+0) |
Birth weight (g): median (range) |
795(580-1100) |
Singleton n (%) |
12 (100) |
Prenatal steroids n (%) |
10 (83) |
Caesarean section n (%) |
10 (83) |
Boys n (%) |
8 (67) |
Days on mechanical ventilation: median (range) |
22 (1;60) |
Postnatal steroids n (%) |
6 (50) |
BPD (O2 at GA of 36 weeks) n (%) |
3(25) |
IVH grade III n (%) |
0 |
PHI n (%) |
2 (17) |
IVH grade III / PHI n (%) |
2 (17) |
NEC n (%) |
1 (8) |
ROP requiring treatment n (%) |
9 (75) |
Definition of abbreviations:
BPD − bronchopulmonary dysplasia; IVH − intraventricular haemorrhage; PHI − periventricular haemorrhagic infarctions; NEC − necrotising enterocolitis
Table II summarizes brain imaging, neurological evaluation, assessment of psychomotor development, and screening for autism spectrum disorders in all individuals. Normal ultrasound scans and brain MR images, normal neurological examination and psychomotor development were found in 4 children (33.3%).
Neuroimaging and neurologic examination and psychomotor development (n=12).
Tabela II. Neuroobrazowanie a badanie neurologiczne i rozwój psychoruchowy populacji badanej (n=12).
Lp | Initials |
Gestational age (GA), weeks |
US at term-adjusted age |
MRI at term-adjusted age |
Neurological exam 2 years corrected age |
Psychomotor development assessment 2 years corrected age |
STAT assessment 2 years corrected age |
---|---|---|---|---|---|---|---|
1 | HB | 26+0 | normal |
normal |
normal |
average |
negative |
2 | KW | 26+0 | IVH II, posthemorragic dilatation |
IVH II, posthemorragic dilatation |
normal |
normal |
no data |
3 | MD | 26+0 | normal |
IVH II cerebellar haemorrhage |
normal |
Iow |
positive |
4 | NL | 26+0 | IVH II | IVH II, posthaemorrhagic cerebellar injury /destruction |
for further observation |
||
5 | 00 | 24+6 | normal |
IVH II, intra- and pericerebellar haemorrhage |
Cerebral palsy (CP) |
general delay |
no data |
6 | PJ | 26+1 | normal |
lll/PHI, cerebellar haemorrhage |
normal |
general delay |
positive |
7 | RO | 24+0 | IVHIII/PHI>5mm |
IVH lll/PHI>5mm, parietal lobe |
ataxia, observation for |
normal |
negative |
8 | SK | 26+0 | negative |
||||
9 | MK | 28+0 | normal |
focal haemorrhage in frontal lobe |
normal |
normal |
positive |
10 | DM | 26+2 | negative |
||||
11 | ZW | 27+0 | negative |
||||
12 | SD | 23+6 | IVH 1, cerebellar |
ataxia, for further |
negative |
Definition of abbreviations:
IVH - intraventricular haemorrhage; PHI - periventricular haemorrhagic infarctions; CP - Cerebral palsy
The ultrasound examination performed between 38 and 42 weeks post-conception showed no abnormalities in 9 children (75%), whereas MRI performed at the same time showed normal results in 4 children (33.3%), only. The brain was normal in both techniques in 4 children (33.3%). Abnormalities in both techniques and similar lesions were found in 2 cases among the remaining eight patients. Additional information found on MRI were cerebellar abnormalities in 5 cases (hemispheric haemorrhage in 4, pericerebellar haemorrhage in 1, including one case with post-haemorrhagic cerebellar hemisphere destruction diagnosed with MRI and not visible on ultrasound), and in one case a small post-haemorrhagic infarct with 3 mm diameter in the frontal lobe.
Neurological examination was performed around 2 years of corrected age (between 1 and 10/12 and 2 and 1/12).
It was normal in 8 out of 12 patients (66.7%). There were 4 children with normal neuroimaging (US, MRI) in this group. In the remaining 4 cases, MRI showed grade II intraventricular haemorrhage, also seen on ultrasound, and additionally cerebellar haemorrhage.
In 4 (33.3%) children the neurological examination was abnormal, including 1 with cerebral palsy diagnosis and 2 with ataxia. MRI revealed haemorrhagic infarction and post-haemorrhagic ventricular dilatation in one, as on US, and grade II IVH with cerebellar haemorrhage, not visible on ultrasound, in three patients.
Even though 9 children were diagnosed with ROP requiring treatment, one child only had severe visual impairment due to vision loss in one eye as a grade 5 ROP consequence. All the other children showed good binocular vision.
Psychomotor development evaluation and autism spectrum disorder screening were performed around 2 years of corrected age (between 1 and 10/12 and 2 and 1/12).
At least average psychomotor development results were found in 7 of 12 children (58%). In four of them, US and MRI showed no abnormalities, while a post-haemorrhagic ventricular dilatation was observed in 3 children (including 1 case in both neuroimaging methods) and haemorrhagic infarct (2 cases, on MRI only). No abnormalities were found in the cerebellum. It should be stressed that in the group of children with at least an average level of psychomotor development, abnormal results of neurological examination (ataxia, observation for CP) were detected in one patient only, while reduced results in at least two of the Children Development Scale sub-tests were present in 3 children (Motor − in all cases, Speech, Vocabulary, Social Behaviour and Scribbling and Drawing − in two, and Comparison – in one case).
Psychomotor development was assessed as lower than average (low or very low / global developmental delay) in 5 children (42%). Cerebellar or pericerebellar bleeding on MRI was reported in all these cases. Neurological examination was normal in 2 children.
Screening for autism spectrum disorder (STAT) was performed in 10 children (83%). It is important to mention that in all the children who presented the risk of autism spectrum disorder (4/10, i.e. 40%) abnormalities on MRI were detected: cerebellar haemorrhage in 3, frontal lobe post-haemorrhagic infarct in 1 child. Positive screening results were associated with abnormal psychomotor development in 3 children, but with abnormal neurological examination in 1 child only.
The study group is small and cannot provide a basis for statistical analysis, but there is a clear tendency to associate abnormal psychomotor development at the age of two years with abnormal brain MRI findings at term. Abnormalities were related to cerebellar pathology in all the cases. The same applies to screening for autism spectrum disorder. The increased risk of ASD has been reported in most cases with cerebellar haemorrhage. However, this result should be treated with great caution and needs to be verified in further diagnostic tests that will show how many of the children in the risk group actually meet the ASD criteria in further development. Such studies are planned as part of the continuation of this project.
US was normal in 4 cases with abnormalities in at least one field: neurological assessment, psychomotor development and STAT, including a case of cerebral palsy, while MRI was not normal in any of them.
In view of the increasing number of papers devoted to the association of cerebellar bleeding with neurodevelopmental problems of prematurely born children, its increased detection rate on MRI is clinically very meaningful. The superiority of MRI over US in this aspect (and not only this one) is not a new finding and has already been described (e.g. 23). In the recent paper by Parodi et al. it has been stated that microhaemorrhages proved to be undetectable by ultrasound even using mastoid fontanelle [24]. MRI is nowadays regarded as the gold-standard technique for the assessment of brain lesions associated with prematurity with emphasis on cerebellar (micro)haemorrhages. There is already quite a large number of studies (e.g. by Limperopoulos et al.) that link cerebellar damage in the premature brain with secondary underdevelopment of cortical projection regions in the contralateral cerebral hemisphere. Decreased prefrontal grey matter volume has been proven to be significantly associated with social and behavioural impairment [25]. In another recent paper Brossard-Racine and colleagues have proved on Diffusion Tensor Imaging (DTI) that preterm infants, even in the absence of overt cerebellar parenchymal injury detected on conventional MRI, have abnormal cerebellar microstructure reflected by fractional anisotropy (FA) and mean diffusivity (MD) alterations as compared with term-born controls [26]. So our focusing on cerebellar damage detected by MRI is clinically and scientifically justified, as the use of MRI increases our ability of prognostication and may be crucial for early therapy.
The preliminary results of our analysis are consistent with other authors suggesting a relatively high prevalence with an increased risk of autism spectrum disorders in children born prematurely [15, 27]. It should be stressed that the risk of ASD is not the same as definitive ASD clinical diagnosis. Furthermore, many authors point to the particularly high prevalence of prematurely-born babies whose positive screening results for ASD are not confirmed in clinical diagnosis [19]. This is especially true for children with various types of health complications, sensory abnormalities and psychomotor development difficulties [28, 29]. Screening tests may produce false-positive results in such circumstances [27]. An in-depth, psychological and psychiatric assessment of a child’s developmental profile, including the specificities of developmental dificulties and competencies, is therefore essential.
On the other hand, children with ASD symptoms may show results below their actual abilities in developmental scales not because of actual cognitive delay, but because of their dificulty in communicating with the diagnostician, refusal to follow instructions, over-absorbance in other activities, etc. These children, due to their adaptive, social and communication dificulties, o&en refuse to engage in a task, and in developmental assessment based on standardized measures they can achieve results that are lower than their actual development potential.
Significantly, early diagnosis of developmental dificulties, especially in extremely premature babies, cannot be regarded as stable in time. Hence, it is important to monitor development and adapt all forms of prevention, early intervention and treatment to the individual developmental needs of the child. This requires cooperation within a team of paediatrician-neurologist-psychologist-psychiatrist. It may cause problems due to significant difierences in the assessment of child development by difierent professionals. One of the reasons is the above-mentioned variable (time-varying) degree of child interaction with a particular examiner. There is no standardized questionnaire / interview tool for parental assessment of child development in Poland. Parents do not always provide the same information on subsequent visits to different professionals. In addition, specialists in different fields use different, not always compatible research tools. For example, a norm for premature infants (especially for extremely preterm ones) in psychomotor development assessment using the Denver scale (which is a screening tool) is the so-called boundary zone, which is very wide. This norm is accepted, because the destructive role of damage is more distinct in disorganizing orderly sequences of physiological processes than in the destruction of individual brain centres. As a consequence, the functional age of children whose brain was prenatally and perinatally significantly exposed to adverse factors remains far behind the chronological age, regardless of possible neurological deficits (30). In the CDS, which is a standardized tool for the general population, the same results will qualify a child below the norm (low or global delay) (Table II).
It should be emphasized that while the results presented here are preliminary, they have already encouraged a deeper reflection on the developmental trajectory complexity of children with extreme prematurity. The observed variety of possible configurations of neurological, psychological and psychiatric findings illustrates the scale of diagnostic and prognostic challenges in this group of children well, and encourages special caution. The results of the analysis indicate the need to take into account the information about development obtained from difierent sources and using difierent methods. They also encourage a prospective, long-term, multidisciplinary and multi-faceted follow-up of the development of children with extreme prematurity, including neurological, psychological and psychiatric assessments. Moreover, it should be stressed that the effectiveness of therapeutic interventions in the group of post-infancy children, related to the improvement of developmental outcomes, especially the level of the development of social and communication competences, has been confirmed by numerous studies [31, 32].
Without neglecting the primary role of transfontanelle ultrasound, we emphasize that magnetic resonance imaging is much more eficient in the evaluation of central nervous system lesions in preterm neonates and better correlates with their further development. The visualization of lesions e.g. in the posterior cranial fossa, remains the domain of MRI, although ultrasound is also performed through the mastoid fontanel. The potential association of these changes with abnormal psychomotor development at the age of two years and with autism spectrum disorder which is raised in the literature and indicated by the results of this preliminary report requires further studies and implies the use of MRI at a term-equivalent age.
The preliminary results of the present study indicate the complex relationships between neuroimaging results, neurological findings, and trajectories of psychological development in extreme preterms, requiring confirmation in further research. The functioning diversity of the children examined indicates the need to place extremely premature babies under well-planned, multi-specialty and long-term care which includes their mental development and functioning evaluation.