Childhood arterial ischaemic stroke incidence, presenting features, and risk factors: a prospective population-based study
Andrew A Mallick, Vijeya Ganesan, Fenella J Kirkham, Penny Fallon, Tammy Hedderly, Tony McShane, Alasdair P Parker, Evangeline Wassmer, Elizabeth Wraige, Samir Amin, Hannah B Edwards, Kate Tilling, Finbar J O’Callaghan
Background Arterial ischaemic stroke is an important cause of acquired brain injury in children. Few prospective population-based studies of childhood arterial ischaemic stroke have been undertaken. We aimed to investigate the epidemiology and clinical features of childhood arterial ischaemic stroke in a population-based cohort.
Methods Children aged 29 days to less than 16 years with radiologically con? rmed arterial ischaemic stroke occurring over a 1-year period (July 1, 2008, to June 30, 2009) residing in southern England (population denominator 5·99 million children) were eligible for inclusion. Cases were identi? ed using several sources (paediatric neurologists and trainees, the British Paediatric Neurology Surveillance Unit, paediatricians, radiologists, physiotherapists, neurosurgeons, parents, and the Paediatric Intensive Care Audit Network). Cases were con? rmed by personal examination of cases and case notes. Details of presenting features, risk factors, and investigations for risk factors were recorded by analysis of case notes. Capture–recapture analysis was used to estimate completeness of ascertainment.
Lancet Neurol 2014; 13: 35–43Published OnlineDecember 2, 2013
Department of Paediatric Neurology, Bristol Royal Hospital for Children, Bristol, UK (A A Mallick MB BCh); Neurosciences Unit, UCL Institute of Child Health, London, UK (V Ganesan MD, Prof F J Kirkham MD, F J O’Callaghan PhD);
Findings We identi? ed 96 cases of arterial ischaemic stroke. The crude incidence of childhood arterial ischaemic Department of Child Health,
stroke was 1·60 per 100 000 per year (95% CI 1·30–1·96). Capture–recapture analysis suggested that case ascertainment Hospitals NHS Trust, was 89% (95% CI 77–97) complete. The incidence of arterial ischaemic stroke was highest in children aged under Southampton, UK 1 year (4·14 per 100 000 per year, 95% CI 2·36–6·72). There was no di? erence in the risk of arterial ischaemic stroke (Prof F J Kirkham); Department between sexes (crude incidence 1·60 per 100 000 per year [95% CI 1·18–2·12] for boys and 1·61 per 100 000 per year of Paediatric Neurology,
St George’s Hospital, London,
[1·18–2·14] for girls). Asian (relative risk 2·14, 95% CI 1·11–3·85; p=0·017) and black (2·28, 1·00–4·60; p=0·034) UK (P Fallon MBBS);
children were at higher risk of arterial ischaemic stroke than were white children. 82 (85%) children had focal features Department of Paediatric (most commonly hemiparesis) at presentation. Seizures were more common in younger children (≤1 year) and Neurology, King’s College headache was more common in older children (>5 years; p<0·0001). At least one risk factor for childhood arterial Hospital NHS Foundation
Trust, London, UK
ischaemic stroke was identi? ed in 80 (83%) cases.(T Hedderly MBBS); Department
of Paediatric Neurosciences,
ected the risk of arterial ischaemic stroke in children. Investigation Evelina Children’s Hospital, Interpretation Age and racial group, but not sex, a?
London, UK (T Hedderly,
of such di? erences might provide causative insights.
E Wraige MBBS); Department of Paediatric Neurology, John
Funding The Stroke Association, UK.Radcli? e Hospital, Oxford, UK
(T McShane MD); Department
population-based study in the south of England using of Paediatric Neurology,
cation by Cambridge, UK (A P Parker MD); Arterial ischaemic stroke is an important cause of acquired several ascertainment sources, with veri?
Department of Paediatric brain injury in children, with long-term morbidity and personal examination of cases and case notes.
1–3Neurology, Birmingham substantial health, economic, and personal costs. Risk
factors in children are markedly di? erent from those Birmingham, UK associated with arterial ischaemic stroke in adults, and Study design(E Wassmer MBChB); and School limited understanding of these risk factors has meant that Arterial ischaemic stroke was de? ned as acute of Clinical Sciences (S Amin MSc,
there is no evidence base to guide management.neurological symptoms secondary to acute focal cerebral H B Edwards MA, F J O’Callaghan)
and School of Social and
Most previous epidemiological studies of childhood infarction in an arterial distribution on brain imaging. Community Medicine
arterial ischaemic stroke have used retrospective database Separate episodes of arterial ischaemic stroke were (Prof K Tilling PhD), University
cit with further of Bristol, Bristol, UK searches, usually based in one institution or in settings recorded if a new acute neurological de?
without universal health-care coverage, to identify
cases.5 Also, in most previous studies assessment of racial di? erences was not possible because the studies were done in racially homogeneous populations. There are few data from these studies on completeness of ascertainment of cases or diagnostic accuracy. We investigated the epidemiology and characteristics of childhood arterial ischaemic stroke in a prospective
www.thelancet.com/neurology Vol 13 January 2014
cerebral infarction occurred after the index arterial Correspondence to:ischaemic stroke; these were deemed recurrent arterial Dr Finbar J O’Callaghan,
Neurosciences Unit, Institute of
ischaemic strokes.Child Health, University College All children were treated in the UK’s National Health London, London WC1N 3LU, UKService (NHS). Because the NHS is a comprehensive, f.o’email@example.com funded health service with universal coverage that is free at the point of use, it is an ideal system for undertaking epidemiological and population-based
studies. Children aged 29 days to less than 16 years at the summarised in panel 1. The IPSS also reports the onset of arterial ischaemic stroke were eligible for proportion of children in whom infection is a risk factor, inclusion. Perinatal and neonatal stroke are typically but these infections are also listed under acute systemic regarded as distinct from later childhood stroke, with disorders, acute head and neck disorders, or chronic di? erent causative factors, presenting features, outcomes, head and neck disorders. Because there is increasing and recurrence rates.6 Health care in the UK for people evidence that simple, common childhood infections can aged 16 years and older is often provided by adult services increase the risk of arterial ischaemic stroke, a further
36 rather than within paediatric services and, therefore, age up to 16 years is typically used as the upper limit in epidemiological studies of uncommon paediatric disorders.7 The study area encompassed all counties and metropolitan regions in the southern part of England and the British Crown Dependencies of Jersey and Guernsey. The northern borders of the study area were Here-fordshire, Worcestershire, West Midlands Metropolitan County, Warwickshire, Northamptonshire, Cambridge-shire, and Norfolk (? gure 1). The study area included 5·99 million children, representing about 50% of all children within the UK.Presenting features and risk factors were categorised according to the scheme previously published by the International Paediatric Stroke Study (IPSS),4 as Figure 1: Map of Great Britain showing the study areaAll acute National Health Service Trusts within the study area (dark shading) were included. Channel Islands are not shown in the ? gure but were also included in the study area.
risk factor category of recent (not present at stroke onset) infection (within 4 weeks preceding the onset of arterial ischaemic stroke) was also used.8 Categories were not mutually exclusive, although a single infection was not included in the recent infection category if it met the de? nition of acute systemic disorders, acute head and neck disorders, or chronic head and neck disorders. Individual patients were included in multiple categories if several risk factors were present. Laboratory and radiological investigations were at the discretion of the treating clinicians, although national guidelines were available and encourage a uniform approach.9 Details of investigations were recorded.The study was granted multicentre ethical approval from Southampton and South West Hampshire Research Ethics Committee (B) and site-speci? c approval at all NHS Trusts within the study area.ProceduresChildren with onset of arterial ischaemic stroke between July 1, 2008, and June 30, 2009, were noti? ed to the study researchers. Cases were identi? ed by three groups of sources: paediatric neurology, paediatric, and other. The paediatric neurology sources consisted of a lead paediatric neurologist at each of the tertiary paediatric neurology centres in the study area, all other paediatric neurologists working in the study area, trainee members of the British Paediatric Neurology Association, and parallel surveillance with the British Paediatric Neurology Surveillance Unit (a rare disease surveillance system that targets all consultant members of the British Paediatric Neurology Association). The paediatric sources consisted of a lead paediatrician at each hospital within the study area and all consultant members of the Royal College of Paediatrics and Child Health working within the study area. The other sources group consisted of a radiologist (typically with an interest in paediatrics or neuroradiology) at each hospital, a physiotherapist at each hospital, at least one neurosurgeon at each paediatric neurosurgical centre, direct noti? cation from parents or carers, and surveillance with the Paediatric Intensive Care Audit Network (a system that collects data on all children admitted to paediatric intensive care units in the UK).All the health professionals involved were sent a monthly electronic reminder requesting reports of any cases seen in the preceding month or to state that no cases had been seen. The researchers reviewed the clinical records and radiological investigations of noti? ed cases to verify eligibility for inclusion and to categorise the presenting features and risk factors. Recurrence of www.thelancet.com/neurology Vol 13 January 2014
Panel 1: International Paediatric Stroke Study scheme for categorisation of childhood arterial ischaemic stroke risk factors Arteriopathy? Protein S de? ciency
? Focal cerebral arteriopathy? Protein C de? ciency
? Moyamoya disease? Prothrombin 20210A mutation
? Arterial dissection? Antithrombin III de? ciency
? Vasculitis? Hyperhomocysteinaemia
? Sickle-cell arteriopathy
? Post-varicella arteriopathyAcute systemic disorders
? Other speci? ed arteriopathy? Fever >48 h
? Unspeci? ed arteriopathy? Sepsis (with positive blood, urine, or CSF cultures)
Cardiac disorders? Dehydration
? Congenital heart disease? Acidosis
? Acquired heart disease? Hypoxia
? Isolated patent foramen ovale? Viral gastroenteritis
? <72 h after cardiac surgery
? Previous cardiac surgeryChronic head and neck disorders
? Cardiac catheterisation? Migraine
? Extracorporeal membrane oxygenation? Brain tumour
? Left ventricular assist device? Other cranial or neck tumour
? Arrhythmia? Ventricular shunt
? Other cardiac? Cerebral aneurysm
? Intracranial arteriovenous malformation
Chronic systemic disorders? PHACES syndrome
? Sickle-cell disease
? Indwelling catheterAcute head and neck disorders
? Trisomy 21? Head or neck trauma
? Other genetic disorders? Pharyngitis
? Haematological malignancy? Meningitis
? Iron de? ciency? Recent intracranial surgery
? Oral contraceptive pill? Otitis media
? Connective tissue disorder? Sinusitis
? Solid extracranial tumours? Mastoiditis
? L-asparaginaseRisk factors for atherosclerosis in adulthood
Prothrombotic states? Hypertension
? Methylenetetrahydrofolate reductase de? ciency? Hyperlipidaemia
? Hyperlipoproteinaemia (alpha)? Type 1 diabetes mellitus
? Factor V LeidenReproduced from Mackay and colleagues4 with permission from John Wiley and Sons. ? Other genetic thrombophiliaPHACES=Posterior fossa malformations, Hemangiomas, Arterial anomalies, Cardiac ? Acquired thrombophiliadefects, Eye abnormalities, Sternal cleft.
arterial ischaemic stroke was also assessed by case note incidence rates were also age adjusted to standard review and by parental interview 12 months after the European20 and WHO populations.21
index stroke.We used capture–recapture techniques to estimate the
completeness of ascertainment22,23 using a three-source
Statistical analysis(paediatric neurology, paediatrics, and other sources) We used data from the O? ce for National Statistics,10–15 the analysis.24 Goodness-of-? t-based CIs were calculated and States of Jersey Statistics Unit,16,17 and the States of dependency between sources was investigated and Guernsey Policy and Research Unit18 to estimate the accounted for by log-linear modelling.25 We used a population denominator. We calculated crude, age-standardised approach with Akaike and Bayesian speci? c, sex-speci? c, and racial-group-speci? c incidence information criteria to select the best ? tting model.26 We rates. 95% CIs were calculated using the Poisson calculated an adjusted incidence rate using the estimated distribution.19 To compare incidence rates, we used the total number of cases from the best ? tting model.Poisson model to calculate incidence rate ratios We described the presenting features and risk factors (expressed as relative risk [RR]), 95% CIs, and two-tailed of the whole cohort of children with arterial ischaemic p values. To facilitate comparisons with other studies, stroke. We tested di? erences in the frequency of www.thelancet.com/neurology Vol 13 January 2014 37
presenting features according to sex and age group, and di? erences in the frequency of risk factors according to racial group, using Fisher’s exact test. Logistic regression, using a penalised maximum likelihood estimation approach,27 was used to identify relations between risk factors and age at stroke and presenting features. Age at stroke onset in decimal years was classed as a continuous study period. The total at-risk population was 5 987 400 and therefore the crude incidence rate of childhood arterial ischaemic stroke was 1·60 per 100 000 per year (95% CI 1·30–1·96). The age-standardised incidence rates were 1·53 per 100 000 per year (95% CI 1·22–1·83) using the European 2013 standard population20 and 1·58 per 100 000 per year 38 variable in logistic regression analyses.
Statistical analyses were done with Stata IC 11.2 (Statacorp, College Station, TX, USA).
Role of the funding source
The sponsor of the study had no role in the study design, data collection, data analysis, data interpretation, writing of the report, or decision to submit for publication. The corresponding author had full access to all the data in the study and had ? nal responsibility for the decision to submit for publication.
905 (76%) of 1185 health professionals invited to respond did so at least once during the identi? cation period. The mean response rate to the monthly electronic reminders was 48·3% (range 45·5–52·8).186 separate noti? cations met the case de? nition. After accounting for duplicates, we identi? ed 96 unique incident cases of childhood arterial ischaemic stroke. One case had had a previous arterial ischaemic stroke 2·8 years before the event within the study period. All other cases were ? rst-ever strokes and no child had another or recurrent arterial ischaemic stroke within the
Figure 2: Age distribution of cases of arterial ischaemic stroke (n=96)
(95% CI 1·26–1·89) using the WHO 2000–2025 standard population.21
Figure 2 shows the distribution of cases by age at onset of arterial ischaemic stroke. The incidence of arterial ischaemic stroke was numerically highest in children aged under 1 year (4·14 per 100 000 per year, 95% CI 2·36–6·72) compared with those aged 1–5 years (2·42 per 100 000 per year, 1·78–3·22), 6–10 years (0·56 per 100 000 per year, 0·27–1·03), or 11–15 years (1·22 per 100 000 per year, 0·78–1·84).
49 children were boys and 47 girls. The crude incidence rate for boys was 1·60 per 100 000 per year (95% CI 1·18–2·12) and for girls was 1·61 per 100 000 per year (1·18–2·14). The age-standardised rate for boys was 1·56 per 100 000 per year (95% CI 1·13–2·00) and for girls was 1·59 per 100 000 per year (1·14–2·04) using the WHO 2000–2025 standard population. The RR of arterial ischaemic stroke for boys compared with girls was 1·00 (95% CI 0·65–1·52; p=0·99). The mean age at presentation for boys was 6·2 years (SD 0·8) and for girls 5·0 years (SD 0·6). Age at presentation did not vary signi? cantly by sex (p=0·22, two-tailed t test).Incidence was signi? cantly a? ected by race, with Asian (p=0·017) and black (p=0·034) children at higher risk of arterial ischaemic stroke than white children (table 1). The incidence rate for white boys was 1·49 per 100 000 per year (95% CI 1·05–2·06) and for girls was 1·23 per 100 000 per year (0·82–1·76), resulting in a male:female RR of 1·22 (0·73–2·05; p=0·43).
Figure 3: Number of cases of arterial ischaemic stroke identi? ed according to source of ascertainment
www.thelancet.com/neurology Vol 13 January 2014
58 (60%) cases were identi? ed by the paediatric systemic disorders, acute head and neck disorders, neurology group, 51 (53%) by the paediatrics group, and chronic head and neck disorders, or recent infection) was 46 (48%) by the other sources group (? gure 3). 47 (49%) present in 27 (28%) cases.
cases were identi? ed by one source, 39 (41%) by two Cerebrovascular imaging was done in 69 (72%) cases, sources, and ten (10%) by all three sources.13 (19%) of whom had more than one modality of The best ? tting model in the three-source capture–cerebrovascular imaging. MR angiography was done in recapture analysis, which included the interaction 68 (71%) cases, conventional angiography in 11 (11%) between the paediatric neurology and other sources, cases, and CT angiography in six (6%) cases. estimated 108 (95% CI 99–124) cases of arterial ischaemic Echocardiography was done in 75 of 94 (80%) cases, stroke, suggesting that case ascertainment was 89% 34 (36%) of whom had bubble contrast. Echocardiography (95% CI 77–97) complete (table 2). Using the point was done transthoracically in 71 (76%) cases and via the estimate of 108 cases, the adjusted incidence rate of transoesophageal route in seven (7%) cases. Testing for at childhood arterial ischaemic stroke was 1·80 per least one prothrombotic risk factor was done in 52 (61%) 100 000 per year (95% CI 1·48–2·18).of 85 cases. Children with apparently idiopathic arterial Focal features, in particular hemiparesis, were the ischaemic stroke (no identi? ed risk factor; n=16) were most common presenting features of arterial ischaemic investigated at least as extensively as the group as a stroke (table 3). 31 (32%) children only had focal features whole: 14 (88%) had vascular imaging, 14 (88%) had at presentation, but di? use features and seizures seldom echocardiography, eight (50%) had bubble contrast occurred in isolation (? ve [5%] children and one [1%] echocardiography, and 13 (81%) had testing for at least child, respectively). 32 (33%) children had both focal and one prothrombotic risk factor.
di? use presenting features, ? ve (5%) had focal features Older children (adjusted for sex and race) were less and seizures, eight (8%) had di? use features and likely than younger children to have more than one risk seizures, and 14 (15%) had focal features, di? use features, factor (odds ratio [OR] 0·91, 95% CI 0·84–1·00; p=0·043), and seizures.acute systemic disorders (0·89, 0·81–0·99; p=0·030), or The frequency of presenting features (including infection (0·89, 0·80–0·99; p=0·036).
subcomponents of the main categories) did not vary Risk factors varied with race. Chronic systemic signi? cantly according to sex (data not shown), but there disorders were found at higher rates in Asian (OR 5·64, were di? erences in the frequency of presenting features 95% CI 1·68–19·0; p=0·005) and black children (36·0, according to age group (table 4). Seizures were most 5·32–243·8; p<0·0001) than in white children. Black common in very young children (<1 year) whereas di? use children also had higher rates of arteriopathy (OR 4·64, features in general and headache in particular were more 95% CI 1·13–19·0; p=0·033) than white children. All common in older children (≥6 years). Diagnosis was analyses were controlled for age and sex. Further often delayed, with the median time to diagnosis from examination of these risk factor categories showed that onset of symptoms being 24·3 h (IQR 6·5–76·0).Asian children had a signi? cantly greater prevalence of No risk factor was identi? ed in 16 (17%) cases. One risk iron de? ciency anaemia than white children (four [29%] factor category was identi? ed in 35 (36%) cases, two in of 14 cases vs three [5%] of 66 cases, p=0·016). 33 (34%) cases, and three in 12 (13%) cases. Figure 4 shows Haemoglobin concentrations at presentation for Asian the frequency of risk factor categories. Infection as a risk children with iron de? ciency anaemia ranged from 25 g/L factor (infection risk factors categorised under acute to 96 g/L and mean corpuscular volumes ranged from
www.thelancet.com/neurology Vol 13 January 2014 39
In this epidemiological study of arterial ischaemic stroke
in children, we found a crude incidence rate of 1·6 per 100 000 per year, higher than previously reported in the UK.28 We show a striking e? ect of age on incidence, with children under age 1 year being particularly at risk (4·14 per 100 000 per year). The risk of arterial ischaemic stroke was not a? ected by sex, but both black and Asian children had signi? cantly higher rates of stroke than did white children. Children with arterial ischaemic stroke present similarly to adults, with focal neurological de? cits such as hemiparesis, but their diagnosis is often delayed. However, the risk factor pro? le for children with arterial ischaemic stroke is very di? erent to that for adults, with few of them having conventional adult atherosclerostic risk factors. A substantial proportion of children with stroke have an underlying arteriopathy and evidence of coincident infection.
We believe that this incidence study has produced one of the most accurate estimates in the published work so far (panel 2). However, a small number of cases who were resident in the study population but treated for their stroke outside the study area will have been missed. Although this is the largest prospective population-based study of childhood stroke so far, the actual numbers of stroke cases is still small in statistical terms. We acknowledge that when no di? erence in incidence was found by sex, we might have missed a marginal real di? erence in the population because the numbers of cases were too small to detect one. Similarly, the di? erences reported in RR for racial groups might be susceptible to a type 1 error in view of the small numbers and the fact that the CIs for the incidence rates by racial group overlap. The analysis of risk factors was limited by the extent to which the individual clinicians pursued their investigations. This study was observational and we did not provide clinicians with an investigation protocol, although guidelines for investigation of childhood stroke exist.9 Therefore, there is probably under-reporting of speci? c risk factors. For example, only 72% of cases had cerebrovascular imaging and therefore cases of arteriopathy will have probably gone undetected.
In this study, we used several sources of ascertainment and veri? ed each incident case by case note and radiology review. We also attempted to assess the extent of ascertainment using capture–recapture techniques. The crude and adjusted incidence rates of 1·6 and 1·8 per 100 000 per year are higher than many previous estimates and this is probably a result of more complete ascertainment. Most previous epidemiological studies of childhood arterial ischaemic stroke used retrospective searches of databases, typically using International Classi? cation of Diseases (ICD) codes, to identify
cases. This method probably led to substantial inaccuracies in previous estimates because both the speci? city and sensitivity of ICD codes for arterial ischaemic stroke are low.29–31 Four studies31–34 have
www.thelancet.com/neurology Vol 13 January 2014
Figure 4: Arterial ischaemic stroke risk factor categories identi? ed
59·6 fL to 69·0 fL. Unsurprisingly, black children had a signi? cantly greater prevalence of sickle-cell disease than white children (eight [89%] of nine vs none [0%] of 66 children) and a greater prevalence of sickle-cell arteriopathy (six [67%] of nine vs none [0%] of 66 children).
reported higher incidence rates (from 1·7 to 2·4 per Panel 2: Research in context100 000 per year) than those in this study, but these all included neonates, accounting for between 8% and 46% Systematic reviewof their total cases. Neonates have a high incidence of We searched PubMed and Embase from 1955 until Nov 6, stroke and the inclusion of neonates in these studies will 2013, for studies published in English for the following have substantially in? ated the overall incidence rate for terms: “stroke”, “cerebrovascular accident”, “infant”, “child”, childhood arterial ischaemic stroke.
The high incidence rate of arterial ischaemic stroke in infancy reported in this study is consistent with those from other studies.33,35,36 In this study, the nadir of the age-speci? c incidence rates was in mid-childhood followed by a small rise in adolescence. A higher rate in adolescence has also been reported in other studies.35,36 Why the incidence rate of childhood arterial ischaemic stroke changes with age is not clear, but is presumably a result of changes in the underlying risk factors with age.
We did not ? nd a signi? cant sex di? erence in the risk of stroke. This contrasts with previous studies in California35 and the IPSS,37 which both found a small but signi? cantly increased risk of stroke in boys. However, the California study relied on information extracted from a retrospective database search using ICD codes, and the IPSS is not a population-based study. In the present study, the incidence of stroke in boys was numerically, but not signi? cantly, higher than in girls for white children. Racial di? erences between cohorts might a? ect di? erences between sex-speci? c rates.
In this European study, we show racial di? erences in the risk of childhood arterial ischaemic stroke. Both Asian and black children had higher rates of arterial ischaemic stroke than white children. In their study of childhood stroke in California, Fullerton and colleagues35 also showed that black children had a higher risk of ischaemic stroke (arterial ischaemic stroke and cerebral venous thrombosis) than white children, but Asian children did not.35 The Asian populations in England and California are very di? erent. Asian children in England are predominantly from an Indian subcontinent background (India, Pakistan, Bangladesh, Sri Lanka, Nepal, and Bhutan). In California, about 90% of the Asian population have an eastern Asian background (China, Philippines, Vietnam, Japan, and Korea).38 In the present study, the increased risk of arterial ischaemic stroke for black children was largely explained by sickle-cell disease, which confers a high risk of arterial ischaemic stroke.39 The higher risk of arterial ischaemic stroke in Asian children in our study could possibly be explained by the high prevalence of iron de? ciency anaemia in this racial group. Severe iron de? ciency anaemia without any other identi? able risk factors has been described as a cause of childhood arterial ischaemic stroke in other studies.40,41 Asian children in the general population of England have higher rates of iron de? ciency anaemia than white children, which is related to factors such as socioeconomic deprivation, low maternal iron stores, and feeding practices.42,43
www.thelancet.com/neurology Vol 13 January 2014 “adolescent”, “paediatric”, “pediatric”, “incidence”, and “epidemiological studies”. We searched the reference lists of retrieved articles for any further studies that might have been missed in the initial search. We identi? ed 17 studies that had calculated the incidence of childhood arterial ischaemic stroke.InterpretationMost previous studies (15 of 17) were retrospective reviews of populations that have identi? ed patients using International Classi? cation of Diseases codes, which have limited sensitivity and speci? city. We undertook a large, prospective, population-based study of childhood arterial ischaemic stroke in the published work. We accessed detailed clinical and radiological information about each case and veri? ed case ascertainment by capture–recapture techniques. The present study shows how stroke incidence varies by race (more common in black and Asian children than white children) and by age (more common in children <1 year than 1–15 years), but we did not detect any di? erence in sex. We showed that stroke presents di? erently according to age: seizures were present in most children younger than 1 year whereas headache was more common in older children. However, focal neurological de? cits (85%) and hemiparesis (72%) were the most common presenting features in childhood arterial ischaemic stroke. Increased awareness of arterial ischaemic stroke in children is needed to help reduce the substantial time delays in diagnosis that prevent rapid initiation of neuroprotection and possible use of thrombolysis.Focal neurological de? cits (85% of cases), and hemiparesis in particular (72% of cases), were the most common presenting features of childhood arterial ischaemic stroke. However, di? use presenting features and seizures were also common, occurring in 61% and 29% of cases, respectively. The presenting features of childhood arterial ischaemic stroke are not dissimilar to those in adult arterial ischaemic stroke, and stroke recognition instruments developed for use in adults have reasonable sensitivity in childhood arterial ischaemic stroke.44 Nevertheless, there is still a need for increased awareness of childhood arterial ischaemic stroke among health professionals if the diagnostic delays that are frequently reported are to improve.45,46However, presenting features did vary with age. In particular, seizures were more common in infants, occurring in 75% of those aged less than 1 year at stroke onset. Children under the age of 1 year with arterial ischaemic stroke in a North American cohort also had an increased risk of presenting with seizures.47 The 41
42 immature brain might be more excitable because of age-5 Mallick AA, O’Callaghan FJ. The epidemiology of childhood stroke. related di? erences in brain receptor composition, Eur J Paediatr Neurol 2010; 14: 197–205.number, and distribution, and this enhanced excitability 6 Nelson KB, Lynch JK. Stroke in newborn infants. Lancet Neurol might favour seizure propagation in the context of acute 2004; 3: 150–58.injury such as arterial ischaemic stroke.7 Lynn RM, Pebody R, Knowles R. Twenty years of active paediatric 48 The prevalence surveillance in the UK and Republic of Ireland. Euro Surveill 2006; of some presenting features such as speech disturbance 11: E060720 4.and headache might vary with age because of age-related 8 Fullerton HJ, Elkind MS, Barkovich AJ, et al. The vascular e? ects of developmental changes. Headache at presentation was infection in Pediatric Stroke (VIPS) Study. J Child Neurol 2011; 26: 1101–10.associated with increasing age. Headache is largely a self-9 Paediatric Stroke Working Group. Stroke in childhood: clinical reported symptom so relies on the development of guidelines for diagnosis, management and rehabilitation. London: su? cient communication to be able to be identi? ed. The Royal College of Physicians, Clinical E? ectiveness and Evaluation Unit, 2004.youngest child with reported headache was 2·4 years old. 10O? ce for National Statistics, Population Estimates Unit. Much of the increase in headache with increasing age Mid-2008 population estimates: selected age groups for local may be because of an age-related reporting bias.authorities in the United Kingdom; estimated resident population. Fareham: O? ce for National Statistics Centre for Reported risk factors for childhood arterial ischaemic Demography, 2010.stroke are varied and di? erent from those linked with 11 O? ce for National Statistics, Population Estimates Unit. adult stroke. At present, there is a scarcity of robust Mid-2008 population estimates: estimated resident population by quinary age groups for local authorities. Fareham: O? ce for evidence to support the use of the term risk factors in National Statistics Centre for Demography, 2010.childhood arterial ischaemic stroke and many of the risk 12 O? ce for National Statistics, Population Estimates Unit. factors reported must be regarded as putative.49 Nevertheless, Mid-2009 population estimates: selected age groups for local authorities in the United Kingdom; estimated resident population. we have used the term risk factor to ensure consistency Fareham: O? ce for National Statistics (Population Demography with the existing paediatric arterial ischaemic stroke Division), 2011.published work.4 Nearly half of children in the present 13 O? ce for National Statistics, Population Estimates Unit. Mid-2009 population estimates: estimated resident population by study had multiple risk factors. However, 17% of cases quinary age groups for local authorities. Fareham: O? ce for remained idiopathic (although, as described, the battery of National Statistics (Population Demography Division), 2011.investigations was not prescribed and was, therefore, not 14 O? ce for National Statistics, Population Estimates Unit. Mid-2008 population estimates: single year of age and sex for local uniform). Almost a third of cases had an arteriopathy and authorities in the United Kingdom; estimated resident population. there is increasing belief that focal cerebral arteriopathies, Fareham: O? ce for National Statistics (Population Statistics perhaps related to infection, underlie many cases of Division), 2010.childhood stroke.15 O? ce for National Statistics, Population Estimates Unit. 50,51 The discovery of an arteriopathy Mid-2009 population estimates: single year of age and sex for local might have substantial implications both for treatment authorities in the United Kingdom; estimated resident population. and prognosis,34,52 and we echo recent guidelines that Fareham: O? ce for National Statistics (Population Statistics Division), 2010.suggest vascular imaging should be done in all cases of 1 6 States of Jersey Statistics Unit. Report on the 2001 Census. suspected childhood stroke.1 St Helier: States of Jersey Statistics Unit (Policy and Resources Department), 2002.Contributors 17 States of Jersey Statistics Unit. Jersey Census 2011: bulletin 1—The study was conceived by FJO’C, VG, and FJK and designed by FJO’C, total population. St Helier: States of Jersey Statistics Unit, 2012.VG, FJK, and AAM. AAM, SA, and HBE collected the data. AAM and 1 8 States of Guernsey Policy and Research Unit. Guernsey annual FJO’C analysed the data. KT advised on capture–recapture analyses. population bulletin 2010. St Peter Port: States of Guernsey Policy FJO’C, VG, FJK, PF, TH, TMS, APP, EWa, EWr, and AAM were on the Council, 2011.study steering committee and led identi? cation of cases. AAM wrote the 1 9 Ulm K. A simple method to calculate the con? dence interval of a ? rst draft. AAM, FJO’C, and VG revised the report and all authors standardized mortality ratio (SMR). Am J Epidemiol 1990; commented on subsequent drafts and approved the ? nal version. The 131: 373–75.study was supervised by FJO’C.2 0 Pace M, Lanzieri G, Glickman M, et al. Revision of the European Con? icts of intereststandard population. Luxembourg: Publications O? ce of the We declare that we have no con? icts of interest.European Union, 2013.2 1 Ahmad OB, Boschi-Pinto C, Lopez AD, Murray CJL, Lozano R, AcknowledgmentsInoue M. Age standardization of rates: a new WHO standard. The study was funded by a research grant from the Stroke Association, UK.Geneva: World Health Organization, 2001.References2 2 Crowcroft NS, Andrews N, Rooney C, Brisson M, Miller E. Deaths 1 Roach ES, Golomb MR, Adams R, et al. Management of stroke in from pertussis are underestimated in England. Arch Dis Child 2002; infants and children: a scienti? c statement from a Special Writing 86: 336–38.Group of the American Heart Association Stroke Council and the 23 Kroll ME, Murphy MF, Carpenter LM, Stiller CA. Childhood cancer Council on Cardiovascular Disease in the Young. Stroke 2008; registration in Britain: capture–recapture estimates of completeness 39: 2644–91.of ascertainment. Br J Cancer 2011; 104: 1227–33.2 Ganesan V, Hogan A, Shack N, Gordon A, Isaacs E, Kirkham FJ. 24 Hook EB, Regal RR. Capture–recapture methods in epidemiology: Outcome after ischaemic stroke in childhood. Dev Med Child Neurol methods and limitations. Epidemiol Rev 1995; 17: 243–64.2000; 42: 455–61.2 5 Regal RR, Hook EB. Goodness-of-? t based con? dence intervals for 3 Lo W, Zamel K, Ponnappa K, et al. The cost of pediatric stroke care estimates of the size of a closed population. Stat Med 1984; and rehabilitation. Stroke 2008; 39: 161–65.3: 287–91.4 Mackay MT, Wiznitzer M, Benedict SL, Lee KJ, Deveber GA, 2 6 Hook EB, Regal RR. Validity of methods for model selection, Ganesan V. Arterial ischemic stroke risk factors: the International weighting for model uncertainty, and small sample adjustment in Pediatric Stroke Study. Ann Neurol 2011; 69: 130–40.capture-recapture estimation. Am J Epidemiol 1997; 145: 1138–44.www.thelancet.com/neurology Vol 13 January 2014
2 7 Firth D. Bias reduction of maximum-likelihood-estimates.
Biometrika 1993; 80: 27–38.
2 8 Kirkham FJ, Williams AN, Aylett S, Ganesan V. Cerebrovascular
disease/stroke and like illness. In: British Paediatric Surveillance Unit. 17th Annual Report. London: Royal College of Paediatrics and Child Health, 2003: 10–12.
2 9 Golomb MR, Garg BP, Saha C, Williams LS. Accuracy and yield of
ICD-9 codes for identifying children with ischemic stroke.
Neurology 2006; 67: 2053–55.
3 0 Tuckuviene R, Kristensen SR, Helgestad J, Christensen AL,
Johnsen SP. Predictive value of pediatric thrombosis diagnoses in the Danish National Patient Registry. Clin Epidemiol 2010; 2: 107–22.3 1 Agrawal N, Johnston SC, Wu YW, Sidney S, Fullerton HJ. Imaging
data reveal a higher pediatric stroke incidence than prior US estimates. Stroke 2009; 40: 3415–21.
3 2 Steinlin M, P? ster I, Pavlovic J, et al. The ? rst three years of the
Swiss Neuropaediatric Stroke Registry (SNPSR): a population-based study of incidence, symptoms and risk factors. Neuropediatrics 2005; 36: 90–97.
3 3 Barnes C, Newall F, Furmedge J, Mackay M, Monagle P. Arterial
ischaemic stroke in children. J Paediatr Child Health 2004; 40: 384–87.3 4 Fullerton HJ, Wu YW, Sidney S, Johnston SC. Risk of recurrent
childhood arterial ischemic stroke in a population-based cohort: the importance of cerebrovascular imaging. Pediatrics 2007;
3 5 Fullerton HJ, Wu YW, Zhao S, Johnston SC. Risk of stroke in
children: ethnic and gender disparities. Neurology 2003; 61: 189–94.3 6 Tuckuviene R, Christensen AL, Helgestad J, Johnsen SP,
Kristensen SR. Paediatric arterial ischaemic stroke and cerebral sinovenous thrombosis in Denmark 1994–2006: a nationwide population-based study. Acta Paediatr 2011; 100: 543–49.
3 7 Golomb MR, Fullerton HJ, Nowak-Gottl U, Deveber G.
Male predominance in childhood ischemic stroke: ? ndings from the International Pediatric Stroke Study. Stroke 2009; 40: 52–57.3 8 US Census Bureau. Census 2000 data for the State of California.
http://www.census.gov/prod/2002pubs/c2kprof00-ca.pdf (accessed Nov 20, 2013).
3 9 Kirkham FJ. Therapy insight: stroke risk and its management in
patients with sickle cell disease. Nat Clin Pract Neurol 2007;
www.thelancet.com/neurology Vol 13 January 2014 Articles4 0 Basak R, Chowdhury A, Fatmi L, Saha N, Mollah A, Yasmin S. Stroke in the young: relationship with iron de? ciency anemia and thrombocytosis. Mymensingh Med J 2008; 17: 74–77.4 1 Munot P, De Vile C, Hemingway C, Gunny R, Ganesan V. Severe iron de? ciency anaemia and ischaemic stroke in children. Arch Dis Child 2011; 96: 276–79.4 2 Lawson MS, Thomas M, Hardiman A. Iron status of Asian children aged 2 years living in England. Arch Dis Child 1998; 78: 420–26.4 3 Wharton BA. Iron de? ciency in children: detection and prevention. Br J Haematol 1999; 106: 270–80.4 4 Yock-Corrales A, Babl FE, Mosley IT, Mackay MT. Can the FAST and ROSIER adult stroke recognition tools be applied to con? rmed childhood arterial ischemic stroke? BMC Pediatr 2011; 11: 93.4 5 McGlennan C, Ganesan V. Delays in investigation and management of acute arterial ischaemic stroke in children. Dev Med Child Neurol 2008; 50: 537–40.4 6 Srinivasan J, Miller SP, Phan TG, Mackay MT. Delayed recognition of initial stroke in children: need for increased awareness. Pediatrics 2009; 124: e227–34.4 7 Zimmer JA, Garg BP, Williams LS, Golomb MR. Age-related variation in presenting signs of childhood arterial ischemic stroke. Pediatr Neurol 2007; 37: 171–75.4 8 Abend NS, Beslow LA, Smith SE, et al. Seizures as a presenting symptom of acute arterial ischemic stroke in childhood. J Pediatr 2011; 159: 479–83.4 9 Roach ES, Lo WD, Heyer GL. Pediatric stroke and cerebrovascular disorders, 3rd edn. New York: DemosMedical, 2012.5 0 Amlie-Lefond C, Bernard TJ, Sebire G, et al. Predictors of cerebral arteriopathy in children with arterial ischemic stroke: results of the International Pediatric Stroke Study. Circulation 2009; 119: 1417–23.51 Hills NK, Johnston SC, Sidney S, Zielinski BA, Fullerton HJ. Recent trauma and acute infection as risk factors for childhood arterial ischemic stroke. Ann Neurol 2012; 72: 850–58.52 Danchaivijitr N, Cox TC, Saunders DE, Ganesan V. Evolution of cerebral arteriopathies in childhood arterial ischemic stroke. Ann Neurol 2006; 59: 620–26.43