Basal ganglia stroke has been reported to occur in infants in association with minor falls. Over the last decade, this association has been described in several case reports and small case series.[1-4] Affected infants present with facial paresis and hemiparesis soon after the injury and small infarcts in the basal ganglia are visible on neuroimaging. This condition has been reported to have a favourable prognosis. Investigations for the cause of stroke in these patients failed to provide an explanation. Although various aetiological factors have been postulated to explain this characteristic presentation and anatomical distribution of the lesions, there is no definitive evidence to prove any pathogenic mechanism. In a recent study, Yang et al. described a series of 16 infants, less than 18 months of age, who developed cerebral infarction after mild head trauma. The authors reported that basal ganglia calcifications and cytomegalovirus (CMV) infection were the only potential risk factors in these infants. In 2008, we noticed the presence of mineralization of the lenticulostriate arteries in infants who had suffered basal ganglia strokes, with the onset of stroke occurring, in most cases, after minor trauma. In this paper, we report the demographic profile, clinical characteristics, imaging features, laboratory findings, and follow-up of these infants.
- Top of page
- Supporting Information
Our centre is a tertiary-level paediatric hospital in the state of Andhra Pradesh, India. This study was approved by the hospital ethics committee and consent was obtained from parents of all the infants. All infants with subcortical stroke who presented at our centre between July 2008 and July 2012 were evaluated. They were categorized into one of several groups according to the type of stroke: arterial ischaemic infarct, venous thrombosis, and haemorrhagic strokes. The cohort of infants with mineralization in the distribution of the lenticulostriate arteries and basal ganglia infarcts formed the study group and was followed up prospectively. Age, sex, demography, history of onset, history of trauma/fall, time lag between fall and weakness, fever, vomiting, and brief recurrent dystonia were recorded. History of developmental milestones as well as medical and family history of similar episodes, was also documented. The results of anthropometry, systemic examination for organomegaly, cardiac evaluation and eye examination, and details of the neurological examination were recorded.
Plain computed tomography (CT) of the brain along with reconstructions in the sagittal and coronal planes were performed in all infants. The following investigations were performed in all infants: complete blood counts, sickling test, lipid profile, serum homocysteine test, serum lactate test, cardiac evaluation, retroviral serology, serum calcium, phosphorus and alkaline phosphatase levels were analysed. The first 11 patients in the series underwent a full battery of investigations to identify hypercoagulable states (protein C, protein S, and antithrombin III assays), sources of emboli or vasculopathy (echocardiography, magnetic resonance imaging [MRI], and magnetic resonance angiography [MRA]) and disorders reported to occur in association with basal ganglia mineralization (serum parathyroid hormone levels, and serum and cerebrospinal fluid lactate levels). Eye examination was performed in all infants to test for the presence of cataracts and any features of intrauterine infection. After an interim review of the initial group of 11 patients, in whom MRI, MRA, and procoagulant work-up were normal, the protocol was modified to omit these expensive investigations in subsequent infants.
These infants were treated with aspirin only, at a dose of 3 to 5 mg/kg body weight once daily, during the initial part of the study. When one infant had a recurrent stroke while on treatment with aspirin, cilostazol was added as a second agent in the next 17 infants. All infants underwent rehabilitation therapy. Recovery from the stroke, further neurological development, and occurrence of recurrent strokes was documented on follow-up. CT examinations were performed if symptoms recurred.
The statistical analyses were carried out using spss version 17 software (IBM Corporation, New York, USA). The linear regression model was created in which recurrence of stroke was a dependent variable. Independent predictors were entered into the model on the basis of a correlation coefficient of less than 0.4. The beta coefficients of the predictors that showed a significant correlation were reported as predictors for recurrent stroke (see Table SI, online supporting information). Results were considered significant only if the p-value was less than 0.05.
- Top of page
- Supporting Information
A total of 76 children suffered arterial ischaemic strokes during the study period. Among the 53 children with cortical infarcts, there was no evidence of parenchymal mineralization. There were 23 infants with subcortical stroke. One infant with tubercular meningitis and basal ganglia stroke did not show mineralization of the lenticulostriate arteries. The 22 infants with basal ganglia ischaemic stroke and mineralization of the lenticulostriate arteries detected using CT formed the study group.
The demographic data, clinical and radiological characteristics, and follow-up details of the study group are summarized in Tables 1 and 2. In total, 15 of the 22 infants were less than 1 year old. Of the 18 infants with a history of trauma, the trauma were caused by rolling over from a bed to the ground (n=6), a fall while attempting to stand or walk (n=10), or a fall while standing on a chair (n=2). In six children, onset of neurological deficits occurred immediately, while in two the deficits were reported 48 hours after the fall. In 10 children onset of neurological symptoms occurred between 10 minutes and 6 hours (mean 74.5min, SD 105.2min) after the fall. Four children had similar characteristic findings and courses despite no definite history of a fall. None experienced loss of consciousness or seizures after the fall. Hemidystonia on the affected side was noted in 16 infants, with onset between the second and fourth day after the fall, subsiding within 24 hours. Hemidystonia was typically brief, with each episode lasting a few seconds without loss of consciousness, and electroencephalography (EEG) carried out during these events was normal. Two infants presented 48 hours after the fall with recurrent episodes of hemidystonia, but had very minimal weakness which went unnoticed by the parents (see Video S1, online supporting information).
Table 1. Demographic and clinical characteristics of participants (n=22)
|Mean age, SD (mo)||11.0, 4.8|
|History of fall||18/22|
|Time lag between the fall and onset of symptoms (n=18)|
Table 2. Clinical and radiological features of participants
|Patient no.||Sex||Age (mo)||History of fall?||Time lag (min)a||Acute dystonia||Mineralizationb (right, left)||Neurodevelopmental outcome||Follow-up (mo)|
|1||Male||14||Yes||30||Yes||2, 2||Distal upper limb weakness||24|
|3||Female||14||Yes||10||No||2, 2||Mild hemiparesis||15|
|4||Male||18||No||–||Yes||1, 1||Mild hemiparesis with speech delay||28|
|5||Female||7||No||–||Yes||2, 2||Distal upper limb weakness||18|
|6||Female||15||Yes||Immediate||No||2, 2||Mild distal weakness||16|
|7||Male||12||Yes||60||Yes||5, 3||Mild distal upper limb weakness||10|
|8||Female||11||No||–||Yes||3, 2||Mild distal upper limb weakness||6|
|12||Male||8||Yes||15||Yes|| ||Mild distal weakness||5|
|16||Male||6||Yes||30||Yes||3, 2||Speech delay with mild motor delay||24|
|18||Male||11||Yes||360||Yes||2, 1||Fine motor difficulty||4|
|19||Female||8||Yes||48h||Yes||1, 1||Motor delay||4|
|20||Male||12||Yes||30||No||2, 2||Mild hemiparesis||3|
|21||Female||8||Yes||Immediate||Yes||3, 4||Mild hemiparesis||1|
|22||Male||13||Yes||Immediate||Yes||3, 3||Mild hemiparesis||2|
Brain CT demonstrated characteristic, sharply marginated, linear hyperdense lesions coursing vertically through the inferior aspect of the lentiform nuclei in all infants. On coronal and sagittal reconstructions, the distribution of these linear lesions was similar to that described for lenticulostriate arteries in anatomical literature. These lesions had attenuation values between 58 and 90 HU, representing mineralization. The number of such mineralized vessels ranged from one to five on each side. There were no other foci of calcification in the basal ganglia or elsewhere in the brain. The infarcts were centred on one of the mineralized lenticulostriate arteries, starting approximately at the mid-putaminal level, with variable superior extent (Figs. 1 and 2).
Figure 1. Bilateral lenticulostriate artery mineralization with infarct on left side (patient no. 11). Axial CT scan of the brain (a) and oblique coronal reformats (b) show the bilateral symmetric mineralization of the lenticulostriate arteries. Wedge shaped hypodensity on the left side (arrows in b) represents the infarct that starts at the mid-putaminal level and extends superiorly into the corona radiata and the body of the caudate nucleus.
Download figure to PowerPoint
Figure 2. Range of appearances of lenticulostriate vessels mineralization. Representative CT from two infants showing (a) multiple mineralized vessels on either side (patient no. 7) and (b) single mineralized vessel on either side (patient no. 4).
Download figure to PowerPoint
On follow-up CT, there was a reduction in the intensity of mineralization, but complete resolution was not seen in any case, even by the age of 3 years 6 months. Magnetic resonance imaging with MRA were normal in the 11 infants in whom it was performed (Fig. 3).
Figure 3. Neuroimaging studies in a child with recurrence of stroke (patient no. 20). Axial CT scan (a) reveals right putaminal infarct and bilateral lenticulostriate artery mineralization, which is not apparent on (b) gradient echo-weighted MRI. Coronal T2-weighted MRI (c) shows subacute infarct in the lenticulostriate artery distribution on the right side and chronic infarct on left side. Magnetic resonance angiogram (d) did not reveal any abnormality.
Download figure to PowerPoint
Investigations including the procoagulant profile, echocardiography, MRA of the brain, and analysis of cerebrospinal fluid lactate levels did not reveal any abnormality in this cohort. Factors predisposing an individual to basal ganglia mineralization, including ophthalmological evaluation for evidence of intrauterine infection, serum calcium levels, phosphorus levels, alkaline phosphatase levels, lactate levels, as well as HIV antibody testing, were normal in all infants.
Eight infants with abnormal limb movements during the course of the illness were evaluated by EEG. This was normal in six children and showed slow-wave discharges in two children. Normal EEG during the events in three infants confirmed the non-epileptic nature of these events. None experienced recurrence of these events during the follow-up period.
Five of the 22 infants experienced a recurrence of stroke (Table 3). In two infants, CT performed at presentation revealed a gliotic lesion in the basal ganglia in addition to the symptomatic acute infarct, indicating a prior episode of stroke that went unnoticed. Three other infants experienced stroke recurrence after the initial diagnoses. All five recurrent strokes were precipitated by a fall. Imaging studies in these five revealed that the second infarct occurred contralateral to the older (chronic) infarct. Individuals 10, 16, and 22 developed the second stroke while on antiplatelet therapy (individual 10 was on aspirin alone, while individuals 16 and 22 were on both aspirin and cilostazol). The interval between initial and second stroke in these three individuals was 2, 3, and 3 months respectively. In two individuals (patients 19 and 20), the first stroke probably occurred early in infancy and went unnoticed. Individual 19 (8mo), presented with recurrent episodes of dystonia and mild hemiparesis. This infant had prior documentation of lenticulostriate vasculopathy on neurosonography during hospitalization for fever and irritability at the age of 1 month.
Table 3. Clinical characteristics of patients with recurrent strokes
|Patient no.||Age at first presentation (mo)||Interval between first and second strokea||Antiplatelet agents before second stroke||Follow-up after second stroke|
|16||6||2mo||Aspirin||Residual dysarthria and hemiparesis at 2y|
|10||6||3mo||Cilostazol and aspirin||Residual hemiparesis at 6mo|
|19||8||Noted old and acute stroke on first presentation with recurrent dystonias||None||Mild residual hemiparesis at 4mo|
|20||12||Noted old and acute stroke at first presentation with hemiparesis||None||Mild hemiparesis with distal weakness at 3mo|
|22||18||3mo||Aspirin and cilostazol||Speech delay with fine motor difficulty at 4mo|
The mean duration of follow-up was 11.36 months (SD 9.4mo; range 1–33mo; median 8mo, SD 9.0mo). There was no mortality in the study group. Four of the five infants in whom stroke recurred had residual weakness and dysarthria, with delay in walking and speech. In the fifth infant with recurrent stroke, who was less than 1 year of age, delayed motor development was accompanied by mild residual hemiparesis. Eight of the 17 infants who did not experience recurrent stroke were neurologically normal on follow-up, while nine infants showed persistent mild hemiparesis, with two also exhibiting speech delay. On further analysis, we observed that hypotonia during first stroke was a negative predictor for recurrent stroke (β −0.490; p<0.05), whereas in acute dystonia at first stroke, low haemoglobin levels were significant predictors of recurrence of stroke.
- Top of page
- Supporting Information
In this study, we described a cohort of infants with mineralizing angiopathy of lenticulostriate arteries, who had basal ganglia stroke following minor trauma. To the best of our knowledge, this is the most comprehensive series describing this distinct clinicoradiological entity. These infants had a characteristic mode of presentation and evolution of symptoms. The salient features of this entity are (1) occurrence in previously healthy infants aged 6 to 24 months; (2) rapid onset of neurological deficits (usually hemiparesis) minutes to hours after a minor fall; (3) appearance of hemidystonia on the affected side in a significant proportion of individuals, typically between day 2 and 4 after the onset of illness, subsiding within 24 hours; (4) CT demonstrating linear mineralization along bilateral lenticulostriate arteries; (5) recurrence of stroke precipitated by another episode of minor trauma; and (6) good short-term and long-term neurodevelopmental outcomes, except in infants with recurrent stroke.
There are several reports of the occurrence of acute ischaemic stroke (basal ganglia infarcts) after minor head trauma.[1-11] These events have been reported in infants below the age of 18 months and constitute less than 2% of all childhood ischaemic strokes. Although this phenomenon has been reported for over a decade, the evolution of clinical manifestations, radiological features, and the mechanism of stroke is still unclear. A review of imaging features in these reports revealed several instances of documented hyperdense foci within the infarcts; similar to that seen in our participants. These foci of hyperdensity within the infarcts have been variably interpreted as thrombus or tiny foci of haemorrhage.[8, 9] Similar demonstration of development of an infarct surrounding the hyperdense focus is evident in one of the individuals illustrated in the report by Sobani et al. In the study by Yang et al., 10 of the 16 infants with stroke after minor trauma had basal ganglia calcification on axial CT images. The vascular nature of mineralization (linear mineralization) has not been documented in these series. We feel that these represent previously unrecognized cases of the same entity. With the widespread use of multislice CT, it is now possible to perform multiplanar reconstructions of these hyperdense foci and their linear distribution along the course of lenticulostriate arteries can be appreciated.
The specific cause of mineralization of the lenticulostriate vessels is still not clear. Yang et al. reported an association between post-traumatic basal ganglia infarction and positive serological testing for CMV, echovirus, Epstein–Barr virus, and mycoplasma. However, review of their cases shows coexistence of basal ganglia mineralization with only CMV seropositivity. Echovirus, Epstein–Barr virus, and mycoplasma were detected as co-infection with CMV and were not independently associated with basal ganglia mineralization. It is not known whether acute CMV infection (as demonstrated by the presence of CMV DNA in urine) can cause mineralization, which is generally a result of long-standing pathology. Further research is needed to explore this association.
The characteristic distribution of these linear areas of mineralization is similar to that described in sonographic lenticulostriate vasculopathy (SLV). This is known to occur in 0.4% of all live-born neonates and in 1.9% to 5.8% of ill neonates.[12-16] It has been reported to occur in association with a variety of congenital and acquired disorders and is known to regress over time.
Mineralization of bilateral basal ganglia vessels has been well documented across the species. A detailed pathological study in younger cynomolgus monkeys aged 3 to 4 years demonstrated a higher incidence of severe lesions than in older monkeys. The type A mineralization noted in the vessel wall of arterioles and venules found in that study could be similar to mineralizing angiopathy in humans. Radiograph microanalysis showed calcium, phosphorus, iron, and chlorine peaks and small amounts of zinc, aluminium, and potassium. This model probably explains many features of mineralization of basal ganglia vessels and its pathological substrate, but what initiates this mineralization in some, and why it is not initiated in others, is still not clear.
Pathologically, lenticulostriate vasculopathy is characterized by thickened hypercellular vessel walls with intramural and perivascular mineralization. Although lenticulostriate vasculopathy is easily visualized on neurosonography, the vascular changes are generally not picked up on CT or MRI.[14, 16, 18-21] We believe that the pathology in our series represents a more severe and persistent form of lenticulostriate mineralization, which is extensive enough to be demonstrated using CT. We propose that this CT demonstrable form of mineralization, along the lenticulostriate arteries, be termed as ‘mineralizing angiopathy’ to differentiate it from SLV. Further longitudinal studies are required to identify the factors associated with persistence of the lenticulostriate artery mineralization in these infants.
The possible association between SLV and mineralizing angiopathy was demonstrated in one individual from our series. This infant had sonographically demonstrated lenticulostriate vasculopathy at the age of 1 month, and CT demonstrated lenticulostriate artery mineralization following a post-traumatic infarct 7 months later. Similarly, Ivanov et al. reported post-traumatic brain infarction in an 8-month-old infant with pre-existing SLV. They postulated that the underlying lenticulostriate vasculopathy predisposed the infant to, or worsened, the vascular obstruction caused by the head trauma. We believe that the stresses across the mineralized lenticulostriate arteries during minor trauma might predispose individuals to thrombosis, precipitating stroke. This appears to be a very age-specific predisposition at particular stage of mineralization. The vessels appear prone to blockage after minor trauma, while remaining asymptomatic at later ages (beyond 2–3y) in spite of persistence of mineralization.
The 2007 National Institute for Health and Clinical Excellence guidelines for the assessment and management of head injury recommend the use of CT as the primary investigation in children with focal deficits after head injury. However, the trivial nature of trauma in this cohort of infants could have been responsible for most of these infants being investigated for stroke, rather than for head injury. Use of MRI as the preferred imaging modality may have contributed to the failure to recognize the pathological entity (which is best demonstrated on CT) in prior reports of infants with post-traumatic basal ganglia strokes. We have retrospectively reviewed the available MR studies in our patients, and could not identify the mineralization on gradient recalled echo or vascular abnormalities on MRA. Even when standard CT for head injury with 5mm axial sections is performed, the true nature of the vascular mineralization may not be appreciated. It may thus be prudent to use thin-section spiral CT with multiplanar reconstructions as the preferred imaging tool in selected individuals where neurological deficits appear after to minor falls, especially infants.
We reviewed over 400 cases of brain CT performed in infants of the same age group for unrelated symptoms, and identified incidental lenticulostriate artery mineralization in only four individuals (1%). These infants were 7 months, 8 months, 11 months, and 1 year old and did not have clinical or imaging evidence of prior stroke. These infants are under close follow-up for any changes in neurological status.
Although the oldest child was 26 months old at presentation, the majority of infants in our series were below 18 months of age, similar to other studies in the literature for post-traumatic basal ganglia strokes. As a tertiary care centre for paediatric neurological disorders, our institute caters to three states in south India. The infants in this series did not demonstrate any predilection for any particular geographical, socio-economic, or cultural backgrounds or any particular ethnic group. With reports of similar cases described from across the world, this condition appears likely to be pan-ethnic.[1-11] All investigations for causes of childhood stroke and procoagulant states were non-contributory in the 11 infants in whom they were performed. There was no ophthalmological evidence of long-standing viral infections, any organomegaly, or any failure to thrive. Similarly, MRA studies were non-contributory for infants. Demonstration of lenticulostriate artery mineralization on CT may thus help in avoiding expensive laboratory investigations and invasive techniques for neurovascular imaging. The cause of mineralization of lenticulostriate arteries needs further research. Interestingly, there was no family history of similar problems in this highly consanguineous population, making it less likely to be genetic in origin.
Recurrent strokes and outcome
In this study five infants experienced recurrent stroke, in all cases after minor trauma. The neurological outcome in the infants in whom stroke did not recur was good. All recurrent strokes in our series occurred before the age of 15 months. Three infants experienced recurrence of stroke while on antiplatelet therapy; however, Yang et al. reported one recurrent stroke in their series though antiplatelet agents were not used. Our observation of stroke recurrence with minor trauma despite the use of antiplatelet agents raises the issue of the utility of these agents in the management of this entity.
Limitations of the study
Procoagulant workup and vascular imaging were not performed in half of the infants in our study. Since our patient population was homogeneous and these investigations were non-contributory in the first 11 individuals, these were omitted after an interim review.