Abstract
Objective: To determine the course of vascular changes in childhood post-varicella arteriopathy (PVA) and its relationship to recurrent arterial ischemic stroke or TIA (AIS/TIA).
Methods: Subjects were children with AIS/TIA occurring <1 year after varicella, ischemic localization consistent with unilateral disease affecting the supraclinoid internal carotid artery or proximal anterior or middle cerebral arteries, and no identified AIS/TIA etiology other than PVA. Charts, brain MRI, and sequential cerebral vessel imaging (selective cerebral angiography or MR angiography [SCA/MRA]) were retrospectively reviewed.
Results: Twenty-three children had varicella at age 1.0 to 10.4 years and had single or multiple AIS/TIAs 4 to 47 weeks later. Initial SCA/MRA was performed within 1 month of presentation, and each child had one to five repeat SCA/MRAs during a 4- to 87-month period. There was vascular stenosis in 19 children, maximal on initial studies in 15 of these. Subsequent stenosis regression occurred in 17 children. In 11 of these, one or two additional SCA/MRAs showed further regression as long as 48 months after presentation; there was no restenosis. Eight of 23 children had recurrent AIS/TIA with antithrombotic therapy within 33 weeks of presentation, including 1 of 17 children with documented stenosis regression.
Conclusion: Vascular stenosis of childhood post-varicella arteriopathy takes a monophasic course, generally with subsequent stenosis regression and only occasional stenosis progression after arterial ischemic stroke/TIA. Arterial ischemic stroke/TIA rarely recurs with antithrombotic prophylaxis after stenosis regression occurs.
Primary infection with varicella-zoster virus, commonly referred to as chickenpox or varicella, occurs at a peak age of 1 to 4 years in Canada.1 By age 10 years, 80 to 90% of children are infected. Pediatric studies show an association between a history of varicella and arterial ischemic stroke or TIA (AIS/TIA) during the following year in otherwise healthy children.2–4 Investigation of post-varicella AIS/TIA often reveals typical findings. Cerebral vessel imaging usually reveals unilateral stenosing arteriopathy affecting the distal internal carotid artery (ICA) and proximal segment of the anterior cerebral artery (ACA) and middle cerebral artery (MCA), whereas brain imaging almost always discloses infarcts within the vascular territory of their lenticulostriate branches (i.e., basal ganglia and internal capsule).2 Children with AIS/TIA caused by post-varicella arteriopathy (PVA) are variably treated with corticosteroid or antiviral drugs,5 but it is unclear whether these drugs are effective in preventing recurrent AIS/TIA, and the optimal treatment duration is undetermined. The course of vascular changes in children with PVA is described in only a few small series6–8 and case reports,9–16 and its relationship to recurrent AIS/TIA is unknown although possibly relevant for treatment. We retrospectively reviewed the charts, brain MRI, and sequential cerebral vessel imaging studies of a cohort of children with AIS/TIA and typical PVA to clarify the natural history of vascular changes in childhood PVA and its relationship to recurrent AIS/TIA.
Methods.
Study centers.
At the Hospital for Sick Children (HSC), Toronto, and the Children’s Hospital at the Hamilton Health Science Centre (CHH), Hamilton, children with AIS/TIA are prospectively enrolled in the Canadian Pediatric Ischemic Stroke Registry, investigated and treated according to institutional guidelines, and followed clinically at specialized pediatric stroke clinics for recurrent AIS/TIA. Radiographic follow-up evaluation is done as clinically indicated per institutional guidelines (e.g., with clinical suspicion of AIS/TIA recurrence). In childhood PVA, cerebral vessel and brain imaging studies are often repeated 3 months after initial studies.
Study participants.
We defined AIS/TIA as clinical episodes of focal, rapid-onset neurologic deficit lasting ≥24 hours (AIS) or <24 hours (TIA) and resulting from cerebral ischemia caused by arterial occlusion. We attributed AIS/TIA to typical PVA in children with 1) first-ever AIS/TIA <1 year after varicella with exanthema; 2) no possible AIS/TIA etiology other than PVA after investigation for other cerebral vasculopathies, including angiitis of the CNS, infectious agents other than varicella-zoster virus, cardiac sources of embolism, and metabolic and prothrombotic conditions; and 3) AIS/TIA manifestations and cerebral infarct location consistent with unilateral vascular disease affecting the supraclinoid ICA, A1 or A2 segments of the ACA, or M1 or M2 segments of the MCA. The diagnosis of PVA was supported by cerebral vessel imaging showing vascular stenoses of these arterial segments. In children with normal findings on cerebral vessel imaging, brain imaging had to show cerebral infarct(s) limited to the vascular territory of the lenticulostriate arteries. Because varicella is highly prevalent during childhood,1 we restricted the interval from varicella to first AIS/TIA to <1 year to limit the risk of random association of unrelated events. All children aged <15 years with AIS/TIA caused by typical PVA in the HSC and CHH registry cohorts (1992 to 2000) were eligible for entry in the present study. Children with no repeat cerebral vessel imaging were excluded from the study cohort.
Data collection.
A neurologist (S.L.) reviewed the medical charts of children with AIS/TIA caused by typical PVA to collect clinical and demographic data (age at varicella, age at AIS/TIA presentation, sex, clinical manifestations, antithrombotic treatments, and recurrent AIS/TIA) and investigation results (EKG, echocardiography, and blood tests for metabolic and prothrombotic conditions, infections, and vasculitis). The timing of varicella was established by parental reporting, which was proven to be accurate in identifying children with varicella in the previous year.17 Children were not differentiated according to vaccination before varicella infection because of inconsistent reports of immunization status in the medical charts. We presume most children were not vaccinated because immunization became available in Canada only at the end of the study period. A neuroradiologist (D.A.), blinded to the study dates, patients’ identity, and clinical outcome, reviewed brain MRI and sequential cerebral vessel imaging studies by selective cerebral angiography (SCA) or MR angiography (MRA) in a random order. We classified abnormal cerebral vessel imaging findings as single gradual stenosis of the vessel lumen, single ring-shaped stenosis, or multifocal stenoses or beading. We quantified the degree of maximal stenosis by comparing lumen diameters of the diseased vessel and the closest healthy vessel segments. We regarded an absolute change in the degree of stenosis of ≥20% on successive cerebral vessel imaging studies as a reflection of progression or regression of stenosis and a change of <20% as an indication of stabilization of stenosis. We correlated recurrent AIS/TIA with the progression of stenosis.
Results.
Clinical presentation and brain MRI.
Twenty-seven children were eligible for entry in our study. Four children were excluded because repeat cerebral vessel imaging studies were not available. Our study cohort comprised 17 boys and 6 girls. Most children of our cohort were previously enrolled in an association study of varicella and AIS/TIA.2 Children in our cohort had varicella at age 1.0 to 10.4 years (median, 4.4 years) and first AIS/TIA 4 to 47 weeks later (median, 17 weeks). Twelve (52%) children had a single episode of AIS; two (9%) had a single episode of TIA; and nine (39%) had multiple episodes of AIS/TIA. Brain MRI revealed infarcts of the basal ganglia or internal capsule in 22 (96%) children, cerebral hemispheres in 10 (43%), and thalamus in 3 (13%), or normal findings in 1 (4%).
Sequential cerebral vessel imaging studies.
Seventy-seven cerebral vessel imaging studies were analyzed. Initial cerebral vessel imaging studies were performed <1 month after presentation (median, 3 days) and consisted of SCA in 19 (83%) children and MRA in 4 (17%). One to five follow-up cerebral vessel imaging studies per child (median, 3) were conducted and comprised MRA in all children, with repeat SCA in five. The median number of studies was four for children with recurrent AIS/TIA vs three without (two-sided Mann–Whitney test, p = 0.10) and three for children with and without residual neurologic deficit or epilepsy (two-sided Mann–Whitney test, p = 0.97). Final follow-up studies were undertaken 4 to 87 months (median = 27) after the initial studies.
Sequential cerebral vessel imaging revealed normal findings on all studies done in 4 (17%) children and unilateral vascular changes in 19 (83%). Vascular changes were located at the M1 segment of the MCA (15 children), the M2 segment of the MCA (9 children), the A1 segment of the ACA (9 children), and the supraclinoid segment of the ICA (7 children) and consisted of a single gradual stenosis of the vessel lumen (9 children), a single ring-shaped stenosis (3 children), and multifocal stenoses or beading (7 children). In all children, no collateral circulation was detected. Stenosis was maximal on initial studies in 15 of 19 children, present on initial studies and increased on follow-up studies with stenosis on arterial segments that were initially normal in 2 children, or absent on initial studies and present on follow-up studies in 2 children (figure 1). Stenosis progression was established by comparing MRA with SCA (two children), SCA with SCA (one child), and SCA with MRA (one child), and maximal stenosis was documented ≤26 weeks after presentation. Subsequent stenosis regression was documented in 17 of 19 (89%) children. During 6 to 79 months after stenosis regression was documented, 11 of 17 (65%) children underwent one to two additional cerebral vessel imaging studies, which detected further regression as long as 48 months after presentation and no restenosis.
Figure 1. Sequential selective cerebral angiography in a child with post-varicella arteriopathy. This child had varicella at age 3 years and acute right-sided hemiparesis 17 weeks later. Brain MRI 5 hours after symptom onset disclosed an acute infarct of the left putamen. His hemiparesis deteriorated acutely on day 2 despite IV treatment with heparin. Selective cerebral angiography disclosed normal findings on day 3 (A), multiple stenoses involving the distal M1 (40%) and proximal M2 (70%) segments of the left middle cerebral artery in week 10 (B), and normalization of the M1 segment and regression of proximal M2 segment stenoses in week 26 (C).
Clinical outcome.
Clinical follow-up evaluation consisted of daily evaluations during hospitalizations and 1 to 15 visits per child (median, 9) with the last visit occurring 10 to 120 months after discharge (median, 75 months). Figure 2 shows the proportion of children with clinical follow-up evaluation over time. All children received antithrombotic therapy according to institutional guidelines. Unfractionated heparin or low molecular weight heparin was given for ≥5 days after AIS/TIA, followed by aspirin. If AIS/TIA recurred with heparin or aspirin alone, the children were treated with a combination of both drugs. No child received corticosteroid or antiviral therapy. During the follow-up period, AIS/TIA recurred in 8 of 23 (35%) children. Of these, three children experienced recurrent AIS/TIA 2 to 4 days after presentation while they were taking anticoagulant therapy, and five had recurrent AIS/TIA 1 to 33 weeks after presentation while they were treated with aspirin. Only one of these recurrences occurred after regression of vascular stenosis, corresponding to an AIS/TIA recurrence ratio of 1:17 (6%; 95% CI, 0 to 29%) in this subgroup. No additional recurrent AIS/TIA was diagnosed in these eight children after they were placed on combined anticoagulant and aspirin therapy. The overall neurologic outcome in this cohort consisted of hemiparesis in 13 of 23 (57%) children, hemidystonia in 6 (26%), hemisensory deficit in 3 (13%), speech problem in 3 (13%), symptomatic epilepsy in 1 (4%), and no neurologic deficit or epilepsy in 9 (31%).
Figure 2. Proportion of children with clinical follow-up evaluation over time after clinical presentation.
Discussion.
Our study suggests that in most children with typical PVA, unilateral cerebral vascular stenosis, which is generally detectable at clinical presentation, follows a monophasic course with occasional progression for up to 6 months, succeeded by spontaneous regression for as long as 48 months after presentation and no restenosis. This is consistent with published case reports of stenosis progression documented 2 to 6 months after the first AIS,6,7,13,15 maximal vascular stenosis (or occlusion) unchanged for up to 18 months,6,8 and eventual regression of vascular stenosis, occurring either spontaneously7–9,12,13 or after antiviral or corticosteroid therapy,10,11,14 within 1 to 14 months after the first AIS/TIA. Normal findings on initial cerebral vessel imaging do not exclude PVA, as shown in two children in our cohort who developed typical vascular stenosis only during the clinical follow-up period. We suggest that cerebral vessel imaging must be repeated in children with presumed PVA and normal findings on initial studies because progression to stenosis may increase the risk of recurrent AIS/TIA.
Collateral vessels were not found in our cohort. The absence of collateral circulation likely reflects the rapid progression and reversibility of vascular stenosis or the short intervals between varicella and AIS/TIA in children with PVA. Collateral circulation was described in reports of atypical PVA with bilateral ICA occlusion. These include a 16-year-old girl with AIS as long as 48 months after varicella and leptomeningeal collateral vessels6 and a 3.5-year-old boy with AIS 1 week after varicella, recurrent TIAs in the next 16 months, and bilateral moyamoya on brain MRA 6 months after the initial AIS.8 In rare instances, moyamoya changes can be present in typical childhood PVA as early as 4 months after the first AIS/TIA.16
Our study demonstrates that approximately one-third of children with PVA experience recurrent AIS/TIA up to 33 weeks after presentation despite antithrombotic prophylaxis with aspirin or anticoagulant drugs. Our study reveals also that AIS/TIA rarely recurs with antithrombotic prophylaxis after regression of vascular stenosis. This may suggest that AIS/TIA recurrence relates to acute vascular injury and thrombosis associated with the progression of vascular stenosis. Although repeat vascular cerebral imaging was not systematic in our cohort after stenosis regression was documented, we found no late restenosis during angiographic follow-up of 6 to 79 months, suggesting that childhood PVA is a monophasic vascular disease. Early AIS/TIA recurrence exhibiting worsening13 or absence of improvement of vascular stenosis8 and no AIS/TIA recurrence or restenosis after improvement of stenosis7,9,10 have also been described in a few case reports of childhood PVA. Conversely, other authors have reported a child with aphasia and seizures that recurred 7 months after an initial stroke, with cerebral MRI showing an old posterior extension of the initial infarct and the SCA revealing regression of vascular stenosis.14 However, the timing of the extension of the silent infarct was not specified; the child apparently received no antithrombotic prophylaxis, and it is unclear whether the recurrent aphasia, which resolved when seizures were controlled, resulted from epileptic activity.
We sought to limit potential biases by conducting blind and random analysis of cerebral vessel imaging studies. Although the number of studies per child varied, no significant difference was found between groups with different clinical outcomes. One potential limitation of our study relates to our interchangeable use of SCA and cerebral MRA. MRA compares reliably with SCA in detecting lesions of major cerebral arteries,18 which are typically affected in PVA, but has not been rigorously validated against SCA in measuring intracranial stenosis. MRA may overestimate the degree of stenosis in as many as 25% of children.18 By using a criterion of ≥20% of absolute change to identify progression or regression of stenosis, we may have minimized the risk of overdiagnosing disease progression. However, we believe MRA can be used in children to establish disease regression. Another limitation is the strict criteria we adopted to diagnose “typical” PVA. Our findings should be verified in future studies of children with “atypical” PVA affecting the posterior cerebral arteries8,14,19 or vertebrobasilar system,3,15,20 with bilateral involvement,6–8,18 or with AIS/TIA occurring either before the exanthema of chickenpox20 or >1 year after varicella.6
Practical implications emanate from our study. Evidence of varicella-zoster virus within the vessel wall of diseased arteries is found at autopsy in children with PVA21 and in adults with cerebral arteritis associated with herpes zoster ophthalmicus,22 supporting that childhood PVA results from viral migration from the trigeminal ganglion and nerve to the major cerebral arteries. A recent literature review of varicella-associated AIS/TIA showed no obvious benefit of antiviral drugs or corticosteroid therapy because most children recovered nearly completely regardless of therapy, although the authors advocated antiviral therapy because of suspected recent viral replication in such cases.5 Our study suggests that these medications, if considered in future clinical trials or for treatment of individual children with typical PVA, should be relatively brief because AIS/TIA rarely recurs once vascular stenosis begins to regress. The relationship we noted between regression of vascular stenosis and the low risk of recurrent AIS/TIA in our retrospective cohort suggests that sequential cerebral vessel imaging can establish the prognosis in typical childhood PVA. This should be confirmed in a larger prospective cohort with standardized sequential cerebral vessel imaging. If migration of varicella-zoster virus along the trigeminal nerve is the exact mechanism of PVA, repeated virus migration, PVA reactivation, and late AIS/TIA recurrence are theoretically possible, although probably rare.14 We believe an alternative diagnosis, such as cerebral vasculitis, must be reconsidered if restenosis is documented during the follow-up period.
Acknowledgment
The authors thank Mr. Ovid Da Silva, Bureau d’aide à la recherche, Centre hospitalier de l’Université de Montréal for editorial assistance and Dr. Mylène Guèvremont for comments on the manuscript.
References
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.