Does diffusion restriction changes in magnetic resonance imaging predict neurological outcome in neonatal seizures?

BACKGROUND: Neonatal seizures are a common manifestation of brain dysfunction. Neonatal magnetic resonance imaging (MRI) has rapidly become the study of choice for the evaluation of central nervous systems disorders in newborns. According to a study conducted in Wilhelmina Children's Hospital, University Medical Center Utrecht, diffusion Restriction (DR) changes in the MRI is a good indicator of cell dysfunction (reversible or irreversible) within one week of insult.
OBJECTIVES: The main aim of this study was to find the association of DR changes in MRI of brain for neonatal seizures with long term neurodevelopment outcome.
METHODS: This is a retrospective observational study conducted in Sri Ramachandra University. Retrospective data was collected for the time period of January 2010 to December 2011 from medical records department (MRD) for patient data, neonatal intensive care unit and reports from PACS for MRI images and the Karthikeyan child development unit for their developmental follow up reports.
RESULTS: Comparison of composite score for various domains with DR changes was done with a t-test and comparison of babies with developmental delay and DR changes with Chi-square test. MRI DR changes with developmental outcome in different domains namely cognition, language-receptive/expressive, fine and gross motor was studied. There is no statistical significance among those who have DR changes and with those who do not have DR changes.
CONCLUSION: Though diffusion restriction changes in MRI may not predict adverse long term neuro developmental outcome, they can be of use with regards to individual etiological profile as in stroke. Larger group study and long term follow up is required to substantiate these findings.

Introduction

Neonatal seizures are a common manifestation of brain dysfunction. The newborn brain is vulnerable for seizures. It is also associated with poor neurodevelopmental outcome.[1] The occurrence of neonatal seizures may be the first, and sometimes the only clinical sign of a central nervous system (CNS) disorder in the newborn infant. Neonatal magnetic resonance imaging (MRI) is a relatively new technique which has rapidly become the study of choice for the evaluation of CNS disorders in newborns. MRI provides an excellent anatomical depiction of the brain which far surpasses cranial ultrasound and computed tomography. The advantage of MRI is that it is the only technique which can identify myelination and differentiation of grey matter and white matter in the neonatal brain hence neonatal MRI is becoming an important tool in predicting neurodevelopmental outcomes,[2] and the future of MRI is directed at understanding the prognostic implications of CNS disease within newborns. After the introduction of MR compatible monitors and neonatal coils, safety and image quality has improved. According to Alderliesten et al., diffusion restriction (DR) changes in the MRI is a good indicator of cell dysfunction (reversible or irreversible) within 1 week of insult.[3] In a case series of perinatal stroke, DR changes in MRI were strongly associated with motor sequelae.[4] The main aim of this study was to find the association of DR changes in MRI of the brain for neonatal seizures with long-term neurodevelopment outcome.

Materials and Methods

This is a retrospective observational study conducted at Sri Ramachandra University. Retrospective data was collected for the time period from January 2010 to December 2011 from Medical Records Department (MRD) for patient data, Neonatal Intensive Care Unit (NICU) and reports from picture archiving and communication system for MRI images and the Karthikeyan Child Development Unit (KCDU) for their developmental follow-up reports. MRD in the hospital has the facility to retrieve case sheets with the diagnosis using International statistical classification of diseases and related health problems. With the search word of “neonatal seizures-code no P-90” we found out 98 babies who had seizures within 28 days of life as shown in Figure 1. Newborn babies who had seizures within 28 days of life and had undergone MRI of the brain within 7 days of onset of seizures and who had undergone a developmental assessment at least at 6 months of corrected age were included in the study. Babies with prematurity, congenital anomalies, and those who died were excluded.

Figure 1

Flowchart of study group

The standard unit protocol for neonatal seizures in our NICU is to do an MRI of the brain within 7 days of onset of seizures. The MRI used is GE Signa HDX 1.5 Tesla. The various sequences taken were T1- and T2-weighted images and DR, mean corpuscular volume, MR angiogram for every baby. After explaining the procedure and obtaining parental consent, the babies were cuddled and well wrapped and during natural sleep usually postfeed if the baby is on feeding, the MRI was done. The babies are sedated if needed with intranasal midazolam and monitored during the procedure with MRI compatible monitors. All the MRI was interpreted by a single radiologist. The various MRI abnormalities recorded included parenchymal infarction, DR in the ventrolateral thalamus and posterior putamen, DR in bilateral occipitoparietal cortex and splenium of corpus callosum, T1-flair hyperintensity, dysplastic splenium and corpus callosal thinning, hemorrhage in the lateral ventricle as shown in Figure 2. Etiology was defined based on the clinical presentation and MRI abnormality.

Figure 2

Magnetic resonance imaging images of the study group. (a) Bilateral basal ganglia diffusion restriction changes suggestive of hypoxic ischemic encephalopathy. (b) Left middle cerebral artery infarct. (c) Diffuse subcortical diffusion restriction changes suggestive of encephalitis. (d) Occipital and splenium of corpus callosum changes suggestive of hypoglycemic encephalopathy

All the babies in our institution with neonatal seizures were followed up by the KCDU for their developmental assessment at six and 12 months and then yearly for 6 years. As per the KCDU protocol, there would be a reminder letter sent before their scheduled follow-up, for those who do not report on that day, the second reminder through phone calls would be done, and the appointment would be rescheduled. For the study purpose, follow-up data of babies between 6 and 12 months of age were collected by matching with babies' identity. After all the data were collected, the babies' identity was concealed.

Fifty-seven out of the 98 babies who had undergone MRI were matched with KCDU database; there were 34 babies who had come for follow-up hence included in the study. Complete data for 34 babies were collected, a wide range of parameters including maternal, perinatal, neonatal parameters such as gestational diabetes mellitus (GDM), pregnancy-induced hypertension (PIH), seizure disorders, abnormal antenatal scan, mode of delivery, various resuscitative measures, gestational age, birth weight, head circumference, onset and type of seizure, and etiology of seizures. We looked for DR changes in various areas of the brain.

The developmental assessment was done using Bayley's scale of infant developmental tool III at 6 and 12 months in the KCDU. If the child had come for both follow-up, then 12 months assessment was considered for the study analysis. Composite for each domain is calculated. In this study, we considered delay in development when the neurological deficit is <85% in any of the domain namely cognitive, language, motor, social, and adaptive [Table 1].

Table 1

Bayley scale of infant and toddler development tool (II)

Statistical analysis

Comparison of a composite score for various domains with DR changes was analyzed with t-test and comparison of babies with developmental delay, and DR changes were analyzed with chi-square test.

Results

In the present study, among the 98 babies who had neonatal seizures in the study period, 34 had both MRI and developmental follow-up. Twenty-nine had developmental follow-up at 1 year of age, 5 babies had come for follow-up at 6 months and lost to follow-up thereafter.

Among the study group, 23 (67%) were male babies. The mean gestational age was 37 ± 1.5 weeks; birth weight of 2615 ± 471.6 g and head circumference of 33 ± 1.7 cm. GDM was found among 4 (11%) and PIH in 2 (6%) of mothers. There was no family history of epilepsy. One (3%) of them had abnormal umbilical Doppler flow in the antenatal scan. Six (17%) had undergone emergency lower segment cesarean section (LSCS) while 2 (6%) had elective LSCS. Nine (26%) received bag and mask ventilation and 3 (9%) babies were intubated. The birth injury was found among 3 (8%) babies. On examination, none of them had neurocutaneous markers. The most common type of seizure is focal clonic in 18 (52%) followed by multifocal in 9 (26%), subtle in 8 (23%), and generalized in 5 (14%) and myoclonic seizures in 4 (11%). Multiple types of seizures were present in 14 (41%) babies, and none had tonic seizures. Seizures occurred within 24 h in 12 (35%), between 24 and 72 h in 13 (38%) and >72 h in 9 (26%) babies. The etiology of neonatal seizures with decreasing order of frequency based on clinico-radiological findings were hypoxic-ischemic encephalopathy (HIE) 11 (32%), hypoglycemic encephalopathy 9 (26%), encephalitis 8 (23%), stroke 4 (11%) and for hypocalcemia 2 (5.8%). Table 2 depicts the etiological profile based on the clinic-radiological findings and DR changes. Table 3 depicts the comparison of MRI DR changes with the developmental outcome. Developmental delay was similar in both groups (babies with DR changes and without DR changes). Comparison of MRI DR changes with developmental outcome in different domains namely cognition, language-receptive/expressive, fine and Gross motor is shown in Table 4 and there was no statistical significance among those who have DR changes and with those who do not have DR changes.

Table 2

Etiological profile based on clinico-radiological findings and DR changes

Table 3

Comparison of MRI DR changes with developmental outcome

Table 4

Comparison of MRI DR changes with developmental outcome in different domains

The topographic pattern of injury in the study series were parietal lobe changes in 14, occipital lobe changes in 14, frontal lobe changes in 5, temporal lobe changes in 6, internal capsule changes in 8, corpus callosal changes in 9, thalamus changes in 5, cerebral hemispheres involvement in 7, caudate nuclei in 2, lentiform nuclei in 2, right middle cerebral artery (MCA) infarct in 1, left MCA infarct in 1, hemorrhage in lateral ventricle in 2.

Discussion

Since the introduction of the diffusion-weighted image (DWI), it has become an indispensible part of MRI examination of the brain, as it is more likely to show the changes than the conventional MRI, especially when done between the 2nd and 4th day of seizure onset. DWI is a special sequence in MRI of the brain which measures the alteration of the diffusion of water within the extracellular space and between intracellular and extracellular spaces.[56] Diffusion imaging complements conventional imaging in the evaluation of perinatal brain injury, as diffusion abnormalities better illustrate the extent of perinatal brain injury when compared with conventional MRI, particularly when performed early.[5] Similarly, in our study, we obtained good delineation of the extent of perinatal brain injury. Furthermore, because of the high water content of the neonatal brain, these changes were well marked on DWI. In our study, DR imaging yielded well as it was performed within 7 days, and it is known in literature that DR changes in certain etiology may disappear later as in hypoglycemia.[78] However, there is insufficient evidence on the percentage of reversibility of DR changes in MRI and its co-relation with the developmental outcome. A DR change is a marker of acute ischemia and is only one of the prognostic indicators.[9] MRI facilitates to define the various topographic pattern of injury, and we obtained DR changes mainly in the parietal lobe, occipital lobe, corpus callosum and internal capsule. Due to nonavailability of bedside electroencephalography (EEG), seizures were diagnosed clinically whenever there was a paroxysmal event otherwise unexplained. The most common type of seizures in our study is the clonic type (52%), and this type is known to have better EEG co-relation.[10]

Babies' neurodevelopmental outcome was assessed using Bayley's scale of infant and toddler development,[11] This tool is an individually administered instrument which is one of the good tools to assess the developmental outcome between 1 and 42 months. It also provides information for interventional planning. The Bayley-III assesses infant and toddler development across five domains which are cognitive, language, motor, social, and adaptive.[11] Though the mean composite score is slightly less in the babies who had DR changes but it is not statistically significant.

The seizure outcome would depend on the underlying etiology and the conditions with good prognosis (90–100% normal) are hypocalcemia and subarachnoid hemorrhage. Conditions with poor prognosis (<50% normal) are developmental defects, meningitis, hypoglycemia, HIE.[10] In our study, 11 out of 34 (32%) were due to HIE and 8 out of 34 (23%) due to encephalitis. Twenty-seven out of 34 (79%) babies had seizures which were due to a condition which can lead to the potentially poor neurodevelopmental outcome.

In HIE, among 8 babies who had DR changes, 6 (75%) had DD and out of the total 11 babies, 8 (73%) had DD in one or more domain with a composite score of <85%. In babies with encephalitis among 6 babies who had DR changes 4 (66%) had DD. Among the 2 babies who had hypocalcemia, no 1 had DR changes in MRI but 1 (50%) baby had DD. Usually, hypocalcemic seizures are associated with good prognosis. In a study by Ali et al. 26 out of 81 cases were found to have a developmental delay with normal MRI of the brain, just as few cases of this indexed study.[12] Similar findings were also reported by Faerber et al. and Momen et al.[1314]

De Vries et al. showed that increased signal intensity on DWI at the level of the posterior limb of internal capsule and cerebral peduncle in babies with stroke had subsequent hemiplegia.[15] Similarly, in our study, for stroke DR changes in MRI is correlating well with neurodevelopmental outcome because all the babies with stroke had DR changes and developmental delay in gross motor, fine motor, and cognition. Though DR changes may not be helpful to predict adverse long-term neurodevelopment outcome for these babies, it is found it has a better co-relation for babies with stroke. HIE is known to affect the central gray nuclei and perirolandic cortex along with basal ganglia and brainstem as hyperechogenic and edematous changes; a similar depiction was found in our study.[1617] In hypoglycemic encephalopathy sub group, 1 out of 4 babies with DR changes had DD. They commonly have occipital lobe abnormalities, and, therefore, regular ophthalmological follow-up and visual assessment is very important among these babies as we can anticipate a problem with vision. Tam et al. showed that DR changes in babies with symptomatic hypoglycemia correlated better with visual outcome.[7] CNS infections usually impact the subcortical area, and site of the lesion may also vary depending on the organism causing the infection. Group B streptococcus infection can represent ischemic changes in basal ganglia and thalamus while hydrocephalus and isolated dilatation of the fourth ventricle is seen with Escherichia coli infection. Streptococcus pneumonia and hemophilic influenza cause a subdural effusion.[15]

In expressive and receptive language domain, the mean composite score in both groups were above 88% which could be because there were done at 6 and 12 months. The follow-up data may give a clear picture. Regarding fine motor domain, group with DR changes and without DR changes had low composite score in contrary to gross motor domain which had a composite score of >85% in both groups.

Our study has certain limitations. First, it is a retrospective, and the possibility of selection bias exists. Second, the sample size was small and few babies were lost to follow-up. We also did not have a follow-up MRI of the study group to look for the persistence of the DR changes and development of T1- and T2-weighted changes in MRI.

Conclusion

Though DR changes in MRI may not predict the adverse long term neurodevelopmental outcome, but can be of use with regards to individual etiological profile as in stroke. Larger group study and long-term follow-up are required to substantiate these findings.

That DWI is a recommended imaging in babies with neonatal seizures is already known. Our study adds that DR changes may not influence the adverse long-term outcome for these babies, but it can be of use with regards to individual etiological profile as in stroke.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest