Cerebral Infarction Associated with Acute Subdural Hematoma in Infants: A Rare Entity from Northeastern India

1. Treatment protocol

The treatment of childhood convulsive status epilepticus (SE) aligned with the Indian consensus guidelines [15]. Intravenous midazolam (0.2 mg/kg) was administered, sequentially followed by loading and maintenance doses of levetiracetam and phenytoin and subsequently by midazolam infusion (0.05 to 2 mg/kg/hr). Elevated intracranial pressure (ICP) was managed with boluses (1 to 3 mL/kg) of 3% hypertonic saline, followed by continuous infusion (0.5 to 1 mL/kg/hr). In normotensive children, intravenous mannitol (2.5 mL/kg over 20 minutes) was also administered. Mechanical ventilation targeted arterial partial pressures of oxygen and carbon dioxide in the ranges of 75–100 and 35–40 mm Hg, respectively. Circulatory shock was initially treated with dopamine at 0.5 mcg/kg/min and escalated as necessary. Infants with hemoglobin levels below 8 g/L received packed erythrocyte transfusions (packed red blood cell [PRBC]; 10 mL/kg). All infants exhibiting coagulopathy (elevated prothrombin time [PT], high activated partial thromboplastin time [APTT], or international normalized ratio ≥1.5) received multiple transfusions of fresh frozen plasma (FFP; 10 mL/kg) until normalization of the coagulation profile. Infants in group C (those with coagulopathy suspected to stem from vitamin K deficiency) additionally received 2 mg intravenous vitamin K. Infections were managed with intravenous antibiotics according to the institutional protocol. The infants underwent physical rehabilitation for their motor deficits.

2. Statistical analysis

Categorical variables were presented as numbers (percentages), and intergroup comparisons were made using the Pearson chi-square test. Continuous variables were reported as mean±standard deviation, and comparisons between groups were made using one-way analysis of variance. The Kolmogorov-Smirnov test was used to assess the normality of continuous data. Spearman correlation analysis was conducted to evaluate correlations of age, Glasgow Coma Scale (GCS) score at admission, the interval between seizure onset and hospital admission, total seizure duration, hematoma thickness, and midline shift with pedASPECTS, pc-ASPECTS, and modified Rankin scale (mRS) scores at the 12-month follow-up. P values less than 0.05 were considered to indicate statistical significance.

1. Etiology

All previous articles described pediatric CIASDH following head trauma [11,18,19], except the work of Steinbok et al. [12], who reported CIASDH in six term neonates within 72 hours of uneventful delivery, postulating that ASDH was likely secondary to cerebral infarction. Sepsis-associated coagulopathy and late-onset HDN represent new etiologies for infantile CIASDH. Minor head trauma in infancy is known to cause ASDH [18,19]. Post-traumatic CIASDH extensively involves the cerebral hemispheres [1,11,18], whereas infantile head trauma without ASDH is associated with bilateral occipital cerebral infarction, comparatively sparing the frontal and temporal lobes [20-22]. Momose et al. [11], in a retrospective study of 12 children with CIASDH following major trauma (from abuse, falls, and road traffic accidents), reported a mean age of 14.1±15.0 months. Of their cohort, 58% had bilateral global infarcts, 25% unilateral CIASDH corresponding to the side of ASDH, 8.5% unilateral CIASDH despite bilateral ASDH, and 8.5% asymmetrical bilateral CIASDH despite unilateral ASDH [11]. However, our study predominantly showed asymmetrical bilateral lesions (59%), compared with 17% bilateral global, 12% unilateral global, and 12% unilateral focal CIASDH. Beyond the effects of ASDH, major trauma can cause vascular injuries, systemic hypoperfusion, vasospasm, and embolization; this may explain the higher incidence of bilateral global infarcts noted by Momose et al. [11] relative to our study. Larger studies are needed to clarify differences between CIASDH following minor versus major trauma.

2. Pathogenesis

Mechanical attenuation of the intracerebral arteries due to pressure from the subdural hematoma (SDH) and endothelin-mediated vasospasm can exacerbate the mismatch between increased metabolic demands and decreased perfusion beneath the hematoma, resulting in cerebral infarction [6,23]. Transcranial Doppler and single-photon emission CT have demonstrated elevated vascular resistance and hypoperfusion on the affected side [9,11]. Blood products in the subdural space trigger nociceptive chemical stimulation of dural trigeminal afferents, resulting in cerebral microcirculatory hypoperfusion [11,24]. The profuse trigeminal innervation of proximal intracranial arteries versus the minimal innervation of lenticulostriate and perforator arteries explains the cortical infarct distribution sparing the basal ganglia and brainstem observed in prior pediatric studies [11,25]. Following ASDH, subfalcine and transtentorial herniations cause ACA and PCA compression, resulting in either paramedian or posterior temporal and occipital infarcts, respectively. However, we observed extensive infarcts in multiple large arterial and small vessel territories without clinical or radiological evidence of herniation. Therefore, herniation syndromes were unlikely to have caused the CIASDH present in our cohort, with a multifactorial etiology instead suggested. Another study attributed extensive hypodensities to SE-induced brain damage and hyperperfusion injury [26]. Incidentally, we also noted gyral luxury hyperperfusion on repeat NCCTs within the subacute infarcted cortex—a novel finding in CIASDH—likely due to loss of cerebral microvascular autoregulation following CI, resulting in blood flow exceeding metabolic requirements [27]. SE-induced T2/FLAIR abnormalities exhibit five distinct patterns of affection: cortical, hippocampal, claustral, subcortical, and splenial patterns [27,28]. While we discerned no definite pattern, we noted basal ganglia and brainstem involvement not previously described following SE. In a prior systematic review, most abnormalities reversed completely within a median of 96.5 days, without any long-term sequelae [28]. In contrast, our study demonstrated persistent T2/FLAIR abnormalities among all infants, which did not revert and resulted in moderate developmental delay.

3. Radiological and EEG findings

Although previous pediatric reports have described only cortical infarcts [11,12], adult patients with CIASDH can exhibit cortical, basal ganglionic, and corpus callosum infarcts, both ipsilateral and contralateral to the side of the SDH, in the ACA [9,21], MCA [5,6,9,29], PCA [6,30,31], and lenticulostriate [29] arterial territories. Ours is the first report describing combinations of cortical, corpus callosum, and basal ganglionic infarcts in infants. Notably, our finding of brainstem infarcts has not been previously reported in pediatric or adult CIASDH. Interestingly, we saw no cases of CIASDH contralateral to the ASDH. Falx interhemispheric subdural hematoma (FISH) accounts for less than 0.5% of all ASDH [1]. We observed novel findings of unilateral and bilateral CIASDH in four (23%) infants with FISH. Adult cases have displayed MRA abnormalities such as MCA spasm [5], displacement of the M1 segment of the MCA [29], and thrombosis of long perforating vessels [7]; these were absent from our cohort and previous pediatric studies [11,12]. FVH adjacent to the ASDH indicates sluggish blood flow through leptomeningeal collaterals due to disrupted flow from the cerebral arteries to the microcirculation. FVH was reported in 25% of cases in previous research [11], compared to 29% of infants in our study. Although MRV was normal in our cohort, increased ICP ipsilateral to ASDH has caused slow venous blood flow and the absence of cortical venous visualization on MRV in a few previously reported patients [11]. However, venous infarction is not responsible for CIASDH, as venous vasogenic edema is associated with increased apparent diffusion coefficient (ADC) values. This contrasts with the substantially diminished ADC values that have been consistently noted in CIASDH [11,12], which are indicative of cytotoxic edema.

Although previous studies have described EEG findings in ASDH, ours is the first to describe serial EEG recordings in CIASDH. Initial generalized or asymmetrical unilateral slowing (ipsilateral to hemispheric infarction) suggested gross dysfunction of the infarcted cortex. However, with normalization of ICP, cerebral function improved. This was reflected by normalization of EEG background activity, although interictal epileptiform discharges persisted in areas of residual gliosis.

4. Treatment, outcomes, and outcome predictors

We observed no clear advantage of surgical over conservative management, as the ASDH thickness was less than 10 mm, the midline shift was under 5 mm in all cases, and GCS scores improved with treatment [32]. Siahaan et al. [19] described hinge craniotomy as an effective method to reduce ICP in low-resource settings. Momose et al. [11] performed craniotomies in five children (42% of their patients with CIASDH), including hematoma evacuation, and reported one death (8%). Steinbok et al. [12] reported craniotomy in four neonates (67%), needle aspiration in one (17%), and conservative management in another (17%). Good outcomes were noted in the three infants who received follow-up for at least 1 year (50%), with mild neurological deficits but no seizures [12]. This contrasts with our finding of moderate neurocognitive delay, also reported by Park [26]. Although early outcomes can be favorable, subsequent brain atrophy in infarcted areas impairs age-appropriate long-term neurocognitive development [30].

A prior study identified hematomas thicker than 5 mm with a midline shift of ≥3 mm, disturbance of consciousness at admission, seizures, and absence of skull bone fracture on CT as predictors of CIASDH [11]. Interestingly, we noted that infarcts often disproportionately exceeded the severity of ASDH, even in patients with minimal midline shift. Moreover, despite significantly thicker hematomas, the infants in group C did not differ significantly from the other groups regarding infarct laterality or size. Hematoma thickness and midline shift were not significantly correlated with pedASPECTS or pc-ASPECTS values, and hence infarct size. Thus, we inferred that ASDH severity (as represented by hematoma thickness and midline shift) does not predict infarct location or ultimate size, although it does predict the occurrence of CIASDH.

In pediatric arterial ischemic stroke, Blackburn et al. [33] noted that ASPECTS was not a useful predictor of outcomes following MCA territory stroke, while Lu et al. [14] and Mackay et al. [34] reported fair to good predictive efficacy of pc-ASPECTS and pedASPECTS for unfavorable neurological outcome, cerebral palsy, and epilepsy. Within the MCA territory, individual analysis of the 10 regions indicated that M1, M2, and M3 did not predict 3-month mRS scores; however, odds ratios for infarcts in other regions ranged from 2.6 to 3.8 [35]. Since each ASPECTS region represents a different proportion of brain tissue, and since DWI lesion volume strongly correlates with functional outcomes, individual regions likely contribute unequally to outcomes [35]. Despite this, ASPECTS assigns equal weight to all regions [36], diminishing its predictive utility in the presence of limited ischemia. In extensive infarctions, such as in our cohort, eloquent area involvement is more probable, resulting in significant ASPECTS-mRS correlation and making ASPECTS a strong predictor of poor outcomes [36]. Our novel use of the CASCADE classification and the ASPECTS scoring approach in infantile CIASDH facilitates objective comparison between studies and helps demystify the pathophysiology of this rare entity.

5. Strengths and limitations

Strengths of this study include its use of prospective data collection and uniform evaluation time points for neuroimaging and EEG during follow-up. However, the absence of a comparator group (infants with ASDH but without cerebral infarction) limited our capacity to interpret causality and risk factors. Due to the lack of serial MRIs, we could not determine whether early pseudonormalization or true normalization of DWI/ADC maps over time, and thus changing ASPECTS, correlated with outcomes. Additionally, our study was constrained by the small sample size in each group, affecting the representativeness and generalizability of the findings. Nevertheless, publishing novel clinical and radiological manifestations of rare diseases increases clinician awareness and encourages further research. Collaborative institutional efforts would expand the dataset, enabling the identification of significant prognostic factors and differences among patients with varied etiologies.

6. Conclusion

We describe infants with CIASDH who presented with RSE and predominantly asymmetrical bilateral infarcts, resulting in moderate global developmental delay at the 12-month follow-up, with no mortality reported. Our findings contribute novel etiologies (sepsis-associated coagulopathy and late-onset HDN) and novel radiological presentations (CIASDH associated with FISH, combined cortical and subcortical infarcts, brainstem infarcts, and luxury hyperperfusion during the subacute stage) to the extremely limited existing literature on infantile CIASDH. Infants with HDN were significantly younger and had thicker hematomas, although infarct sizes did not differ significantly among the three groups. While the severity of ASDH did not correlate significantly with infarct size or mRS at 12 months, pedASPECTS and pc-ASPECTS values demonstrated significant moderate correlations with 12-month mRS.