Abstract
Introduction:
Perinatal arterial ischemic stroke (PAIS) underlies approximately 10% of infantile spasms (IS). We aim to identify patterns of brain injury in ischemic stroke that may predispose infants to infantile spasms.
Methods:
Sixty-four perinatal arterial ischemic stroke patients were identified meeting the following inclusion criteria: term birth, magnetic resonance imaging (MRI) showing ischemic stroke or encephalomalacia in an arterial distribution, and follow-up records. Patients who developed infantile spasms (PAIS-IS) were analyzed descriptively for ischemic stroke injury patterns and were compared to a seizure-free control group (PAIS-only). Stroke injury was scored using the modified pediatric ASPECTS (modASPECTS).
Results:
The PAIS-IS (n = 9) group had significantly higher modASPECTS than the PAIS-only (n = 16) group (P = .002, Mann-Whitney). A greater proportion of PAIS-IS patients had injury to deep cerebral structures (67%) than PAIS-only (25%).
Conclusion:
Infarct size was significantly associated with infantile spasms development. Results support theories implicating deep cerebral structures in infantile spasms pathogenesis. This may help identify perinatal arterial ischemic stroke patients at risk of infantile spasms, facilitating more timely diagnosis.
Keywords
Perinatal arterial ischemic stroke (PAIS) is a heterogenous group of disorders involving a focal disruption of blood flow to a specific vascular territory of the brain between 20 weeks’ gestation to 28 days of life. 1 In newborns, multiple areas of the brain can be involved simultaneously depending on the origin of the thrombus or embolus. At an estimated incidence of 1 in 4000 live term births, stroke in the perinatal period is more common than at any other time in childhood. 1 Perinatal arterial ischemic stroke may be diagnosed acutely around the time of birth as neonatal arterial ischemic stroke or retrospectively in older infants as presumed perinatal arterial ischemic stroke. 1 Neonatal arterial ischemic stroke often presents with focal clonic seizures and/or encephalopathy in the neonatal period. Diagnosis is confirmed with neuroimaging (computed tomography / magnetic resonance imaging [MRI]). 2 Presumed perinatal arterial ischemic stroke presents with early handedness, hemiplegia, or seizures in later infancy or early childhood. Diagnosis is made retrospectively with confirmatory neuroimaging, 3 with the knowledge that no other acute event had occurred during the child’s life. Consequences of perinatal arterial ischemic stroke include cognitive, motor, and language deficits as well as a spectrum of seizure disorders. The incidence of epilepsy after perinatal arterial ischemic stroke ranges from 29% to 67%, with a recent study reporting up to 94%. 2,4 -6 Even in cases where initial symptom severity seems mild, evolving deficits often appear as children “grow” into their lesion and disparities between these patients and similarly aged peers become more evident. 7,8
Infantile spasms (IS) is an epileptic encephalopathy consisting of brief, clustered flexion-extension seizures that appear in the first 2 years of life. 9 Subtle variations of eye-rolling or head drop also exist, posing a challenge to diagnosis. These seizures are often associated with developmental regression or stagnation and a hypsarrhythmic pattern on EEG, historically characterized as West syndrome. Cases of infantile spasms also occur without identifiable hypsarrhythmia. 9,10 The International League Against Epilepsy (ILAE) has moved toward classifying these seizures as “epileptic spasms” rather than “infantile spasms” to highlight their development and/or persistence beyond the period of infancy. For the purposes of this study and to remain consistent with existing literature, we will use the term “infantile spasms” or IS.
Infantile spasms can be difficult to treat, and although the spasms may resolve spontaneously within the first few years of life, there are enduring associated deficits that can have a devastating impact on developmental outcome. 9,11 As many as 60% of children develop new seizure types and refractory epilepsy. 12,13 Good neurodevelopmental and intellectual outcomes are rare in children with infantile spasms, and mortality rates as high as 33% have been reported. 12,14 -16
The United Kingdom Infantile Spasms Study (UKISS) investigated 207 infants with infantile spasms. 17,18 Twenty-one (∼10%) had hypoxic-ischemic encephalopathy and 16 (∼8%) had stroke or infarct as the underlying etiology. Of these ischemic events, 54% occurred in the perinatal period. These findings suggest that nearly 10% of infantile spasms are due to perinatal ischemic injury.
Despite the relative rarity of infantile spasms with an incidence of 2-3 per 10,000 live births, this epilepsy syndrome occurs disproportionately frequently in cohorts of perinatal arterial ischemic stroke patients. 4,8,19 -22 Few studies have attempted to describe a relationship between perinatal arterial ischemic stroke and infantile spasms. Alvarez et al 21 described 2 term babies as having infantile spasms secondary to presumed perinatal ischemic infarcts. Their study supported the concept that unilateral, focal insults may lead to infantile spasms despite its generalized or multifocal nature. Golomb et al 23 observed large branch territory infarcts to have occurred in all perinatal stroke patients who developed infantile spasms. What remains unknown is whether there are features of arterial ischemic stroke in the newborn that have both predisposing features and prognostic value. Vascular patterns or territories of arterial ischemic injury that may predispose infants to infantile spasms and predict seizure outcome have not yet been identified.
Through the Pediatric Stroke Program at the Stollery Children’s Hospital in Edmonton, Alberta, we describe a cohort of patients with perinatal arterial ischemic stroke, including neonatal arterial ischemic stroke and presumed perinatal arterial ischemic stroke, who went on to develop infantile spasms. Using a control group of perinatal arterial ischemic stroke patients who did not develop seizures, we sought to explore associations between stroke size, anatomic regions of ischemic injury, and seizure development. We hypothesized that (1) stroke size would be associated with the development of infantile spasms and (2) anatomic regions of ischemic stroke injury (ie, deep gray matter cerebral structures) would predispose infants to infantile spasms.
Methods
We present a descriptive case series of PAIS-IS patients and employed retrospective case-control observational methods to study a cohort of perinatal arterial ischemic stroke patients with infantile spasms against a group of perinatal arterial ischemic stroke patients who remained seizure-free.
Participants
Patients were identified retrospectively through the Alberta Perinatal Stroke Program Edmonton database. In founding this database, informed consent was obtained from each patient’s legal guardian. As part of a larger project assessing brain injury and seizure outcomes after perinatal arterial ischemic stroke, patients were included based on the following criteria: (1) birth between January 2007 and December 2017 at ≥37 weeks’ gestational age, (2) MRI evidence of ischemic stroke or encephalomalacia in an arterial distribution, and (3) follow-up records greater than 1 year from diagnosis documenting seizure occurrence and treatment. Patients presenting with a watershed pattern of infarct or found to have a comorbid medical condition (eg, cardiac disease, genetic anomaly, or meningitis) were excluded. Patients without documented brain imaging or appropriate informed consent were excluded as well. A total of 64 perinatal arterial ischemic stroke patients (36 neonatal arterial ischemic stroke, 28 presumed perinatal arterial ischemic stroke) were identified (Figure 1) and categorized into mutually exclusive groups based on their seizure outcome. Within our cohort of perinatal arterial ischemic stroke patients, we identified 9 (14%) who developed infantile spasms. This patient group (PAIS-IS) served as our population of interest for this study. Sixteen perinatal arterial ischemic stroke patients (25%) who had not developed seizures served as the control group (PAIS-only).
Study design. Patients with perinatal arterial ischemic stroke were identified from the Alberta Perinatal Stroke Program based on inclusion and exclusion criteria, consent, and screening verification by the primary investigator (J.Y.). This retrospective case-control study compared the infantile spasms group (PAIS-IS, n = 9) to the no seizures group (PAIS-only, n = 16).
Comparisons were made between infants who went on to develop infantile spasms after perinatal arterial ischemic stroke and those remaining seizure-free after perinatal arterial ischemic stroke.
Data Collection
Demographics including timing of stroke, birth history, age at infantile spasms presentation, and epilepsy course were gathered through chart review of electronic medical records.
Neuroimaging Analyses
For all patients, images were reviewed on regional neuroimaging software (PACS) at the Stollery Children’s Hospital. Areas of injury, including encephalomalacia, were identified on diagnostic brain MRI and scored based on anatomic involvement. The modified pediatric ASPECTS (Alberta Stroke Program Early CT Score) was used to qualify and quantify the type and arterial distribution of stroke and as a measure of stroke severity. This modified version (modASPECTS) has been shown to be a reliable method of estimating infarct volume with a high degree of validity on acute MRI in both neonates and children. 2,24 Regions supplied by large branches of the anterior cerebral artery, middle cerebral artery, posterior cerebral artery, and the insula were considered cortical. Deep cerebral structures included the caudate, lentiform, internal capsule, and thalamus. Score ranges from 0-15 per hemisphere, with a maximum total score of 30. An area was scored as positive even if a small portion of that region was involved. Based on the timing of imaging within the patient’s clinical course, MRI scans showed either acute injury or were consistent with prior injury. If MRI had been obtained acutely, as for the neonatal arterial ischemic stroke patients, diffusion-weighted imaging (DWI) sequences were used for scoring. If no acute stroke imaging was available, as for presumed perinatal arterial ischemic stroke patients, T2 sequences were used to score prior infarct.
Data Analysis
Results from the PAIS-IS cohort and patterns of injury for all patients were visually analyzed by one of the investigators (R.S.) and reported descriptively. Neuroimaging scores were quantified by modASPECTS. The Mann-Whitney U test was used to compare continuous variables (modASPECTS) between the PAIS-IS and PAIS-only groups.
Results
Nine patients with perinatal arterial ischemic stroke and infantile spasms were identified (89% male, mean age at stroke diagnosis = 4.9 months). PAIS-IS patient characteristics are presented in Table 1 and include a detailed description of ischemic distribution and seizure presentation. Four patients were identified with stroke in the neonatal period (neonatal arterial ischemic stroke), and 5 were identified retrospectively (presumed perinatal arterial ischemic stroke) between 8 and 10 months of age. All 9 PAIS-IS patients had ischemic stroke injury of the middle cerebral artery, with 2 of 9 (patient ID 4 and 5) involving an additional arterial territory and bi-hemispheric injury. Three (33.3%) patients (IDs 1, 3, and 8) had solely cortical involvement of their strokes, whereas 6 (66.7%) had stroke injury to both cortical and deep gray matter structures. None of the patients had exclusively basal ganglia or thalamic stroke injury. The mean modASPECTS in the PAIS-IS group was 11.11 (range 5-22). Figure 2 shows a representative MRI of a presumed perinatal arterial ischemic stroke patient with associated modASPECTS based on the validated scoring system. 25 At the time of data collection, the mean age at last follow-up in the PAIS-IS group was 5.3 years (range 2-10).
Case Series of Patients with Perinatal Arterial Ischemic Stroke and Infantile Spasms (PAIS-IS): Patient Characteristics.
Abbreviations: ACA, anterior cerebral artery; C, caudate; IC, internal capsule; IS, infantile spasms; L, lentiform; MCA, middle cerebral artery; modASPECTS, modified pediatric ASPECTS; MRI, magnetic resonance imaging; NAIS, neonatal arterial ischemic stroke; PCA, posterior cerebral artery; PPAIS, presumed perinatal arterial ischemic stroke; T, thalamus.
Brain magnetic resonance imaging (MRI) of a representative presumed perinatal arterial ischemic stroke patient. Patient 2 presented at 6 months of age with hemiparesis and gaze deviation. He was found to have a remote left middle cerebral artery infarct with injury to cortical and deep cerebral structures, including the thalamus, lentiform, and internal capsule (modASPECTS = 10). He developed infantile spasms at 10 months. (Scoring based on Beslow et al. 25 )
The PAIS-only group consisted of 16 patients who were diagnosed retrospectively after presenting with hemiparesis or developmental delay and did not develop seizures (56% male, mean age at stroke diagnosis = 14.4 months). Mean age at follow-up was 8.6 years (range 6-11). All 16 patients had unilateral middle cerebral artery territory stroke injury, with 1/16 having additional stroke injury to the ipsilateral anterior cerebral artery territory. Mean modASPECTS in this group was 4.69 (range 2-7). Of the PAIS-only patients, 12 of 16 (75%) had exclusively cortical stroke injury, whereas 4 of 16 (25%) had stroke injury involving the deep gray structures. Using the PAIS-only group as a seizure-free control, Figure 3 shows that the development of infantile spasms was significantly associated with stroke size as determined by modASPECTS (P = .002, Mann-Whitney).
Stroke size as determined by modASPECTS and the development of infantile spasms. In children with perinatal arterial ischemic stroke, infantile spasms was associated with a higher modASPECTS compared with the control group (P = .002, Mann-Whitney).
In terms of location of injury, the proportion of patients with deep gray ischemic stroke injury was greater in the PAIS-IS group as compared to the PAIS-only patients (Table 2). Three patients in the PAIS-IS group (patients 5, 6, and 9) were found to have injury to the right deep gray structures, whereas none of the control group patients had injury to the right basal ganglia or thalamus. These were the only 3 patients who presented with neonatal seizures before developing infantile spasms.
Discussion
This study identifies a unique and understudied population of perinatal ischemic stroke patients with infantile spasms. Although these conditions are rare and co-occurrence even rarer, patients often have significant neurologic disability. We found that in a cohort of 64 perinatal arterial ischemic stroke patients identified over an 11-year period, 14% developed infantile spasms. Our data reinforce that infantile spasms occur disproportionately in perinatal stroke patients, particularly in the absence of additional clinical factors (eg, genetic anomaly and history of infection) that may contribute to increased infantile spasms risk. Results suggest that the size of the ischemic injury in perinatal arterial ischemic stroke was associated with development of seizures, as has been described in acute neonatal arterial ischemic stroke. 2 Larger infarct size and arterial stroke injury involving the deep cerebral gray matter structures in our analysis were associated with increased risk for developing infantile spasms after perinatal stroke.
Given the findings of underlying mechanisms for the development of infantile spasms may relate to cortico-thalamic interactions, we hypothesized that perinatal stroke patients with involvement of the deep gray matter structures may be predictive for the development of infantile spasms. Results indicate that deep cerebral ischemic stroke injury may be important in the development of infantile spasms, and support theories implicating deep cerebral structures in the pathogenesis of infantile spasms. 26 Studies of infantile spasms following diffuse hypoxic-ischemic injury and more focal structural and developmental etiologies show association with injury to the basal ganglia and thalami. 25,27,28 Harini et al 28 have suggested that spasms may be expressed when ictal discharges spread through affected corticosubcortical anatomic and metabolic pathways, but true pathogenesis remains uncertain. Gano et al 25 found that the development of infantile spasms in 8 of 176 (4.5%) patients with hypoxic-ischemic encephalopathy was associated with basal ganglia and thalamic injury, as well as total brain injury. Our study similarly found stroke size, as a quantification of brain injury, to be associated with the development of infantile spasms.
Results of this study will help inform physicians of which patients may be most at risk of infantile spasms after perinatal arterial ischemic stroke. These patients may be monitored more closely for the development of spasms in infancy, which could improve detection of this epileptic encephalopathy and possibly lead to more timely treatment. Common concerns that arise from parents of children with perinatal stroke include whether their child will have epilepsy. 5 This study may also contribute to family discussions surrounding post-stroke prognosis and help reduce uncertainty that accompanies a child’s development after perinatal stroke.
Limitations
The objective of this pilot study was to evaluate potential avenues for further investigations in perinatal stroke patients who may be particularly at risk for infantile spasms. Because of the rarity of both perinatal arterial ischemic stroke and infantile spasms, statistical analyses were limited by our sample size. Broad patterns of injury were assessed, that is, cortical strokes vs stroke injury to deep cerebral structures, rather than associations with specific basal ganglia nuclei. Nonetheless, a relative strength is the number of PAIS-IS patients examined in this cohort. As none of our patients had exclusive injury to deep cerebral structures, it is possible that the involvement of deep structures in our study may be a marker of larger infarcts as a risk factor for the development of infantile spasms and epilepsy. Though existing literature suggests deep injury is important for the development of infantile spasms, as highlighted above, our data are unable to verify this. Although ASPECTS was developed for use in cases of acute stroke, neuroimaging analyses were performed with modASPECTS as an objective measure of both infarct size and vascular territory. This scale may have less applicability to chronic scans due to cystic encephalomalacia and remote injury, making it difficult to determine acute infarct from Wallerian degeneration. In order to exclude potential confounding etiologies for seizure risk in perinatal arterial ischemic stroke patients, those underlying conditions or comorbid injury were excluded. Family history has also been shown to predict the development of seizures over time and was not explored in this study. 5 Additionally, we would have liked to include pattern of injury associated with epilepsy outcome after the development of infantile spasms, but available data do not adequately address seizure control.
Conclusion
Results of our study support existing literature identifying stroke size as a significant risk factor in the development of infantile spasms after perinatal arterial ischemic stroke. Similarly, injury to deep cerebral structures may serve as a secondary significant risk factor in infantile spasms pathogenesis. Future studies will aim to identify and analyze specific stroke location as a risk factor for the development of infantile spasms and for seizure outcomes in childhood epilepsy following infantile spasms. This will include identifying specific areas of deep cerebral injury in perinatal stroke that may confer increased risk of infantile spasms compared to infants who remain seizure-free. Larger sample sizes are needed to confirm our current findings.
Footnotes
Author Contributions
RS completed the data analyses and intepretations and drafted and revised the manuscript. OS contributed to manuscript writing and revision of figures and tables. EA contributed to data collection and organization. FMJ and JY contributed with study implementation and manuscript revision. All authors were involved in initial study design and final review of this project.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Ethical Approval
All research was done in accordance with ethics.