Paediatric arterial ischaemic stroke (PAIS) is relatively rare, with an estimated incidence of 3 to 8 per 100 000 for children older than 1 month.1 Many risk factors leading to PAIS have been identified.1–3 The long-term outcome in children with PAIS is frequently poor, and a wide spectrum of impairments influencing physical and cognitive abilities as well as quality of life (QOL) has been reported.4–6 Most studies focus on neurological impairments after PAIS,7 and only a few studies have systematically addressed functional outcome.4,6–8 According to the World Health Organization’s International Classification of Impairments, Disabilities and Handicaps, in clinical research functional outcome can be assessed on the following levels: disease process and impairment (the general effect of the disease on the child), disability (the restriction of the ability to perform tasks within the physical and social environment), handicap (the social consequences of these impairments and disabilities in the domains of relationships, school, and leisure activities), and QOL (the sense of the child’s well-being and life satisfaction in physical, social, and emotional domains).9 The first two levels are physician-oriented outcomes, whereas those at the third and fourth levels consist mainly of patient-based outcomes. The fourth level also contains information from the patient’s or parent’s perspective. Neurological impairments after PAIS may affect up to 90% of children and include hemiparesis, epilepsy, visual deficits, and language, educational, cognitive, and behavioural problems.7,10–14 The percentage of disabilities after PAIS can be as high as 60%.14 Most restrictions are found in the domains of education, motor function, self-care skills, communication, and socialisation.8,12 When measured on the modified Rankin Scale (mRS), 12/100 of children have no disability, 63% have mild disabilities, and 25/100 have severe disabilities.7,15 Handicaps in children with childhood stroke include a necessity for higher levels of educational support and special education, psychiatric symptoms and disorders, sleeping problems, fatigue, emotional lability, and aggressive outbursts and social problems.7,11,12,16 The QOL in this patient group is significantly lower in the domains of autonomy, parents’ relations, social, physical, emotional, school, and general functioning.8,10,11,13 In the present study, functional outcome in children with PAIS was evaluated and an attempt was made to identify possible prognostic risk factors.
Aim To study functional outcome in children aged 1 month to 18 years after paediatric arterial ischaemic stroke (PAIS) and to identify risk factors influencing their quality of life.
Method In a consecutive series of 76 children (35 males 41 females, median age at diagnosis 2y 6mo, range 1mo–17y 2mo; median length of follow-up 2y 4mo, range [7mo–10y 6mo]) with PAIS diagnosed at the Erasmus Medical Centre Sophia Children's Hospital between 1997 and 2006, we collected clinical, biochemical, and radiological data prospectively. In 66 children surviving at least 1 year after PAIS, functional outcome could be evaluated with the World Health Organization’s International Classification of Impairments, Disabilities and Handicaps.
Results Significant risk factors at presentation for a poor neurological outcome were young age, infarction in the right middle cerebral artery territory, and fever at presentation. Fifty-four % of children had severe neurological impairments at 12 months after PAIS, and at last follow-up more than half needed remedial teaching, special education, or institutionalization. Health-related quality of life (HRQOL) questionnaires showed a significantly lower HRQOL in all age groups. Children with a longer follow-up had a lower HRQOL in the cognitive functioning domain.
Interpretation Our study shows significant morbidity and mortality and a reduced HRQOL after PAIS depending on age, fever at presentation, and infarction in the right middle cerebral artery territory.
Health-related quality of life
Middle cerebral artery
Modified Rankin Scale
Paediatric arterial ischaemic stroke
Quality of life
QOL questionnaire for parents of children aged 1–6y
QOL questionnaire for children aged 6–15y, parent form
QOL questionnaire for children aged 8–15y, child form
QOL questionnaire for adolescents aged ≥16y
Netherlands Organisation for Applied Scientific Research Academic Medical Center Leiden
All 76 children (35 males, 41 females), aged between 1 month and 18 years who were diagnosed with PAIS at the Erasmus Medical Centre/Sophia Children’s Hospital between 1997 and 2006 were evaluated in accordance with a standard evaluation protocol. Data were collected prospectively in the Rotterdam Paediatric Stroke Registry (approved by the medical ethics board of Erasmus University Hospital) and included information on mode of onset, clinical data, neurological impairments at onset, affected arteries, vascular distribution areas involved, and possible causes and risk factors for stroke.1–3 Patient characteristics are summarized in Table I. In the 67 children with a follow-up time of 1 year or more, neurological impairments, disability, and handicap were assessed 12 months after onset of PAIS. In 2007, QOL assessment became part of standard outpatient care in our department for children with acquired brain injury. For 66 children we were able to assess QOL after a follow-up time of 1 year or more.
|At PAIS onset||At 12 months follow-up||At last assessment date|
|N of children||76||67||66|
|Median age at diagnosis, y:m (range)||2:5 (0.1–17.2)|
|Median follow-up period, y:m (range)||2:4 (0.6–10.5)|
|Visual field deficit||17||22.4||7||10.5||6||9.1|
|Nr of risk factors identified|
Outcome in children with PAIS was assessed at the levels of disease process and impairment, disability, handicap, and QOL.9
Disease process and impairment
Magnetic resonance imaging (MRI) or computed tomography (CT), with and without contrast enhancement, were re-evaluated with regard to site and extent of area of infarction and involvement of cerebral arteries. On the basis of cranial CT or MRI images or both, ischaemic infarctions were subdivided into infarctions in the territories of the basilar or carotid arteries. Infarctions in the carotid vascular territory were further subdivided into infarctions of the anterior cerebral artery (ACA), of the middle cerebral artery (MCA), and of the lenticulo-striatal arteries. MCA infarctions were classified into four categories: MCA complete (infarction in the cortical and subcortical distribution areas of MCA), MCA cortical (infarction involving the cortical MCA distribution area), MCA subcortical (infarction involving the subcortical part of the MCA distribution area), and infarctions limited to the MCA distribution area in the thalamus, basal ganglia (MCA central), or both. Cerebral images were not available for re-evaluation in two children, necessitating the use of radiological reports.
Clinical data included sex, age at onset, presenting symptoms, pre-existing or concurrent disease, and possible risk factors for PAIS. These risk factors were divided into seven groups: cardiac, infectious, arteriopathy, prothrombotic disorders, malignancy, metabolic diseases, and miscellaneous.
Children were examined to establish impairments at onset of the of PAIS, 12 months after onset, and at the moment of the last follow-up by a paediatric neurologist. Neurological symptoms were grouped into the following categories: altered consciousness, visual field deficits, aphasia, motor symptoms (pyramidal symptoms or ataxia), epilepsy, behavioural, and cognitive.
The degree of disability was scored with the mRS at 12 months after the onset of PAIS and at the last follow-up in accordance with the following four categories:14 mRS 1: no residual disability; the child attends regular education and does not need remedial teaching; mRS 2: mild residual disability; the child is able to attend regular education but needs remedial teaching because of mild motor disturbances, mild learning disability, or both; mRS 3: severe residual disability; the child has a severe motor deficit (needs braces or wheelchair), severe learning disability, or both, attends a school for special education or is confined to a daily care centre.; mRS 4: death.
Information concerning the effect of disabilities and impairments on social participation, learning, and on the educational setting was obtained from parents at the last follow-up. Behaviour was evaluated at the last follow-up by the consulting paediatric psychiatrist or neuropsychologist (FKA) by means of behavioural observation of the child and interviews with the parents, following the DSM-IV criteria.17
Quality of life
At the last outpatient follow-up visit, HRQOL was assessed using the following measures: the Netherlands Organisation for Applied Scientific Research Academic Medical Center Leiden (TNO-AZL) Preschool Children Quality of Life Questionnaire (TAPQOL) for children aged 1 to 6 years, the TNO-AZL Children’s Quality of Life Questionnaire Parent Form (TACQOL-PF) for children aged 6 to 15 years, the TNO-AZL Children’s Quality of Life Questionnaire Child Form (TACQOL-CF) for children aged 8 to 15 years, and the TNO-AZL Adolescents’ Quality of Life Questionnaire for adolescents aged 16 years and older (TAAQOL).18
Modified Rankin scale data were analysed with the Mann–Whitney U test and with a linear regression analysis. We used a stepwise procedure with Akaike’s Information Criterion to select the most important predictors of outcome.19 Proportional odds analyses were performed for the four mRS groups by using R software.20 We estimated 95% confidence intervals for the odds ratios non-parametrically with bootstrapping.21 A total of 2000 bootstrap samples were drawn. The data of the HRQOL questionnaires were compared with validated reference values from the Dutch population, similarly to the cultural background of children and parents included in our study. HRQOL scores were converted into z-scores. The z-score reflects the extent to which a score deviates from the mean of a normative population.18 The χ2 test was used to compare the variance of our study group with that of the normative population. HRQOL data were analysed with the Mann–Whitney U test and Spearman’s rank correlation. HRQOL data fulfilling requirements of parametric testing and n≥20 were analysed with two-sided independent samples t-tests and two-sided paired samples t-tests.
Disease process and impairments
Sixteen children presented with fever. In five there was no clear infectious cause, four had sepsis, four had meningitis, one had meningo-encephalitis, one had influenza, and one presented with haemolytic uremic syndrome (Table I). Thirteen (17.1/100) children had two or more infarctions, of whom eight (10.5/100) had bilateral infarctions. Affected arteries and distribution areas are shown in Table II. Eight children died, of whom five were in the acute phase within 2 weeks of stroke onset. Three died as a result of massive oedema after a complete stroke in the area of the right (one child) or left (two children) carotid artery. In these children, PAIS was associated with sepsis. Two children died as a result of massive oedema after bilateral complete MCA infarctions associated with cardiac disease. One child with moyamoya disease died after a recurrent stroke 7 months after first stroke onset. One child with mitochondrial encephalopathy, lactate acidosis and stroke-like episodes and bilateral complete MCA infarctions died 11 months after PAIS. One child died 6 years after the cerebral thromboembolic event as a result of renal failure and sepsis. Risk factors were identified in 92% of the children (Table I). The most important risk factors were infection, arteriopathy, and cardiac disease. Ten children had prothrombotic abnormalities such as protein S deficiency, Factor V Leiden mutation, or prothrombin mutation. Four children had cancer at the time of PAIS: acute pro-B lymphoblastic leukaemia, rhabdomyosarcoma, cerebellar ependymoma, and craniopharyngeoma. In five children metabolic disease was diagnosed, namely glucose-phosphate isomerase deficiency, hyperhomocysteinemia, methylene tetrahydrofolate reductase deficiency, mitochondrial encephalopathy lactate acidosis and stroke syndrome, and cobalamin biosynthesis deficit. In one child, PAIS occurred after an accidental intravenous administration of spironolactone. One child had trisomy 21 as a possible risk factor. Impairments at presentation and follow-up are shown in Table I.
|According to artery||n||%||Location of infarction||n||%|
|Anterior cerebral||6||7.9||Posterior area||16||21.1|
Of the 67 children with a follow-up of 1 year or more, 17 children were classified in mRS 1, 17 in mRS 2, and 33 in mRS 3. Younger age at onset of PAIS was correlated with a higher mRS at 12 months after onset (Spearman’s r=−0.27, p=0.02) and had thus a poorer outcome. Proportional odds regression analysis revealed that younger age at onset, fever, and stroke in the right MCA territory (including all four subcategories) were the strongest predictors of a poor neurological outcome at 12 months after onset of PAIS (Table III). Six of eight children with multiple infarctions in either the left cerebral hemisphere (one child) or both cerebral hemispheres (seven children) survived. Surprisingly, no significant difference in mRS classification was found in this group of children in comparison with children with a single infarction (p=0.10). At last assessment, only the child who died 6 years after the onset of PAIS had converted to another mRS classification.
|Characteristic||OR for dichotomy (95% CI)||OR for four categories (95% CI)|
|Involvement of the right MCA||6.0 (1.6–43)a||4.2 (1.2–24)a|
|Fever at presentation||6.1 (1.2–27)a||3.7 (1.2–19)a|
|Age at onset||0.87 (0.65–1.0)a||0.89 (0.74–1.0)a|
At the last follow-up, 10 children had behavioural disturbances, which we classified in accordance with DSM-IV17 (Table I). DSM-IV diagnoses were as follows: overanxious disorder of childhood (n=2); attention-deficit–hyperactivity disorder, predominantly inattentive type (n=6); autistic disorder (n=1); and mental retardation (n=1). Seventy-three % of children received a normal education, of which 24/100 needed remedial teaching. Twenty-seven per cent attended special education schools or were institutionalized.
Health-related quality of life
The means and standard deviations of z-scores and p values of the QOL questionnaires are summarised in Table IV. All parents filled in the TAPQOL, and 92% of parents completed the TACQOL-PF. Sixty-five % of children completed the TACQOL-CF and 75% the TAAQOL. The other children could not respond to the questionnaires because of the severity of their impairments. On the TAPQOL, parents indicated significantly lower motor functioning, more anxiety, a less positive mood, and more stomach and lung problems; on the TACQOL-PF significantly more problems in all domains except for positive emotions were reported in comparison with parents of healthy children. Children indicated more social problems on the TACQOL-CF and more fine motor and cognitive problems on the TAAQOL than healthy children. Parents of children who filled in the TACQOL-PF 6 years or more after the onset of PAIS reported significantly more cognitive problems than parents of children with an interval of less than 6 years (z=−2.12, p=0.03). Children younger than 2 years at PAIS onset scored significantly lower on physical functioning on the TACQOL-CF (z=–1.98, p=0.05) than children who were older than 2 years at stroke onset. No significant difference was present between the TACQOL-PF and the TACQOL-CF. Children in mRS 3 with a severe disability scored lower on motor functioning (F=6.71, p≤0.01, df=43), cognitive functioning (F=4.00, p=0.03, df=37), and autonomy (F=3.32, p=0.05, df=39) of the TACQOL-PF than children within mRS 1 with no disability. Children with a mild disability in mRS 2 scored lower on cognitive functioning than children in mRS 1 without disabilities (F=4.00, p=0.03, df=37).
|Questionnaire and category||z-score||p|
|TAPQOL (n=11, response rate 100%)|
|Abdominal pain||−6.52 (0.8)||<0.001|
|Lung problems||−2.39 (1.59)||0.017|
|Positive mood||−2.16 (2.64)||0.031|
|Motor functioning||−6.80 (4.61)||<0.001|
|TACQOL-PF (n=44, response rate 92%)|
|Positive emotions||−0.26 (1.04)||0.078|
|Negative emotions||−0.37 (1.13)||0.018|
|TACQOL-CF (n=24, response rate 65%)|
|Positive emotions||0.09 (0.86)||0.661|
|Negative emotions||−0.03 (1.07)||0.885|
The relatively small number of children included in our study can be seen as a limitation and may also explain the fact that we did not find male predominance in our study group.22 However, findings concerning age at onset, risk factors, recurrence risk, and mortality are in accordance with earlier findings.4,7
To assess functional outcome in children with PAIS, the WHO’s International Classification of Impairments, Disabilities and Handicaps was applied.9 We preferred this classification over the recently published revised version of the International Classification of Functioning, Disability and Health, Children and Youth23 because measurements of personal factors such as QOL are included in the classification used. In our opinion these factors are important parameters in quantifying the impact of disease on the child, the parents, and the family. They are included in the International Classification of Impairments, Disabilities and Handicaps, but regrettably not in the International Classification of Functioning, Disability and Health, Children and Youth version.9,23
Several childhood stroke outcome measures have been evaluated in earlier studies. The Paediatric Stroke Activity Limitation Measure shows good reliability, but its predictive utility and validation against other measures of outcome has not been fully evaluated.8 The Paediatric Stroke Outcome Measure provides a composite score of radiographic information and deficits in several modalities of function. It has a high interrater reliability and good concurrent validity when compared with standard tests of intelligence, speech, and upper extremity function.24 However, the consequences of the impairments (the disability) in the ability of a child to perform tasks within the physical and social environment are not taken into account.15 The mRS disability outcome score has been used in previous PAIS studies and has been used in children with other causes of acquired brain injury.7,10,15,25 In agreement with results in earlier studies, a favourable neurological outcome (mRS 1 and 2) was found in about 5/10 of the children.25,26 However, the 45/100 of surviving children requiring special education or remedial teaching after PAIS in our cohort is still nine times the average of 5/100 in the Netherlands.25 A clear limitation of the mRS is the rather wide definition of mRS 3, leading to a discrepancy between severity of disability and school type. This is further illustrated by our finding that only 54% of the children classified as mRS 3 attended special education facilities or a daycare centre. In the Netherlands, this is because of the current policy of keeping children with even severe motor disabilities but relatively preserved cognitive abilities within the regular schooling system with supportive measures, financially facilitated by the government (in the ‘Going to school together’ project).
Our finding that right MCA infarction is a risk factor for poor outcome in PAIS, to our knowledge, has not previously been described. In one study, 38 children with unilateral ischaemic stroke in the MCA territory acquired after the neonatal period were described.5 In these children, mean Performance IQ but not verbal IQ was decreased in both the left and right lesion groups. Unfortunately, functional recovery was not reported in the two groups. In adult patients after right hemisphere stroke, symptoms of neglect proved an important predictor for poor functional recovery.27 Because of the young age of our study group, we were not able to assess the presence of neglect systematically, but a similar underlying deficit to that in adults can be hypothesized. Fever as a risk factor for severe outcome after PAIS in our study was often related to severe infections. This corresponds to results in adult studies, in which fever in the early phase of stroke has been found to have a negative influence on outcome at the 3-month follow-up.28 Our finding that younger age at onset of PAIS is a significant risk factor for a poor neurological outcome supports findings by the Hetherington group that children of 6 months to 4 years of age at stroke onset have a worse cognitive outcome than older children.26
Comparison of studies on QOL after PAIS is difficult because different methods have been used to analyse patient groups.4,6–8,10 In two earlier studies, the validated Child Health Questionnaire was used, which evaluates functioning in different domains but not the subjective impact of a deficit and has been validated only in children older than 5 years.6,8 In a Canadian study, the modified Euroqual questionnaire was used, and another study used a non-validated questionnaire.4,10 All but one of these studies concluded that children experience a good long-term QOL after PAIS.4,6,8,10 The TACQOL was chosen as a QOL measuring instrument because of its unique feature that the child’s or the parent’s opinion on the child’s health status is considered, rather than that of a health professional. Simultaneously, this is also a limitation of our study because children with severe impairments were not able to complete the child form of the questionnaires. However, parents gave these children lower scores on motor functioning, cognitive abilities, and autonomy than parents of children with no disabilities. Strikingly, we found impaired QOL in all age groups, both in children with a relatively mild disability and in children with severe impairments. This confirms data that QOL may be independent of the severity of neurological outcome.8 Parents of children with a follow-up period of more than 6 years reported more cognitive problems than those with a shorter follow-up. These long-term impairments may be caused by a failure to develop adequate new neurological pathways as a result of the previous ischaemic lesions. This phenomenon has been named ‘growing into deficit’ and has been reported in children with benign brain tumours but not in children with PAIS.25
PAIS is a severe childhood disease. Our study shows significant morbidity and a reduced HRQOL in survivors of PAIS, depending on age at onset, fever at presentation and infarction in the right MCA territory. Children surviving more than 6 years after stroke onset may show the phenomenon of ‘growing into deficit’ because they experience more cognitive problems than children with a shorter follow-up.
What this paper adds
- • This is one of the largest series of children with paediatric arterial ischaemic stroke (PAIS) prospectively studied.
- • Half of survivors of PAIS have severe neurological impairments 1 year after stroke onset.
- • This is the first study in which quality of life of children with PAIS was studied systematically.
- • Three risk factors for poor outcome of PAIS were found; young age at onset, fever at presentation, and infarction in right middle cerebral artery territory.
We are indebted to Dr Y Vergouwe, biostatistician, for her help with the final statistical analysis.