Downloaded from adc.bmj.com on October 11, 2014 - Published by group.bmj.com
ADC Online First, published on October 6, 2014 as 10.1136/archdischild-2014-306382 Original article
Paediatric cerebral sinovenous thrombosis: findings of the International Paediatric Stroke Study R N Ichord,1 S L Benedict,2 A K Chan,3 F J Kirkham,4,5 U Nowak-Göttl,6 for the International Paediatric Stroke Study Group 1
Department Neurology & Pediatrics, Children’s Hospital of Philadelphia, Perelman School of Medicine of the University of Pennsylvania, Philadelphia, Pennsylvania, USA 2 Department Neurology, Primary Children’s Hospital, University of Utah, Salt Lake City, Utah, USA 3 Division of Hematology and Oncology, Department of Pediatrics, McMaster University, Hamilton, Ontario, Canada 4 Neurosciences Unit, University College London Institute of Child Health, London, UK 5 University hospital Southampton NHS Foundation Trust, Southampton, UK 6 Department of Pediatric Hematology/Oncology, Universitätsklinikum SchleswigHolstein, Kiel, Germany Correspondence to Dr Rebecca Ichord, Department Neurology CTRB 10th Fl, Children’s Hospital of Philadelphia, 3501 Civic Ctr Blvd, Philadelphia, PA 19104, USA; [email protected] FJK and UNG shared senior authorship. Received 10 March 2014 Revised 7 September 2014 Accepted 16 September 2014
ABSTRACT Objectives We evaluated clinical features, treatment practices and early outcome in a multicentre cohort of children with cerebral sinovenous thrombosis (CSVT). Methods Children with CSVT from 10 countries were enrolled from January 2003 to July 2007 in the International Paediatric Stroke Study. We analysed clinical symptoms, underlying conditions, antithrombotic treatment and neurological outcome at hospital discharge in 170 children. Results Of 170 children enrolled, 60% were male; median age 7.2 years (IQR 2.9–12.4). Headache, altered consciousness, focal deficits and seizures were common presenting clinical features. Infarction affected 37% and intracranial haemorrhage 31%. Risk factors included chronic disease in 50%; acute systemic illness or head/ neck disorders 41%; prothrombotic state 20% and other haematological abnormality 19%. Discharge neurological status was normal in 48%, abnormal in 43% and unknown in 5%. Antithrombotic therapy was common, most often low molecular weight heparin was common, with significant regional variation in treatment practices. Mortality was low (4%) and was associated with no anticoagulation but not underlying chronic disease, anatomic extent of thrombosis or intracranial haemorrhage. Abnormal neurological status at discharge or death was associated with decreased level of consciousness at presentation and the presence of an identified prothrombotic state. Conclusions Our study extends the observations of previously published smaller studies in children with CSVT that this is a morbid disease with diverse underlying causes and risk factors. Divergent treatment practices among highly specialised centres as well as limited data on treatment efficacy and safety suggest that further study of this condition is warranted.
INTRODUCTION
To cite: Ichord RN, Benedict SL, Chan AK, et al. Arch Dis Child Published Online First: [ please include Day Month Year] doi:10.1136/archdischild2014-306382
Cerebral sinovenous thrombosis (CSVT), defined as blood clot in a vein carrying blood from the brain to the heart, affects 0.34–0.67/100 000 children annually, of which 27%–35% are neonates.1–3 Mortality is 3%–12%, and neurological sequelae affect up to 62% of survivors.4–6 Aetiology in the paediatric population includes a wide spectrum of underlying conditions and symptoms at clinical presentation.1–3 5 7 Recent studies suggest that haematological abnormalities, including prothrombotic disorders and iron deficiency, contribute to development of CSVT.1 5 8–12 Previously published cohort studies have described epidemiology, treatment practices and outcomes, but are limited due to small sample size or single-centre enrolment.
What is already known on this topic? ▸ Cerebral sinovenous thrombosis in childhood is fatal in 3%–12% and causes neurological sequelae in up to 62% of survivors. ▸ Previously published cohort studies have described epidemiology, treatment practices and outcomes, but are limited due to small sample size or single-centre enrolment. ▸ Uncertainty remains regarding risk factors for death and adverse neurological outcome and optimal treatment.
What this study adds? ▸ The major strengths of our study are its large sample size and representation of multiple international centres not included in smaller studies. ▸ This adds valuable new data regarding a rare condition, improving knowledge of risk factors, current treatment practices and outcomes. ▸ Our finding of increased mortality in patients not treated with systemic anticoagulation supports previous observations that have suggested there is probable benefit of anticoagulation.
The International Paediatric Stroke Study (IPSS) established a prospective multicentre registry to characterise epidemiology, treatment and outcome of infants and children with arterial ischaemic stroke (AIS) and CSVT. The objective of this paper is to analyse this large multicentre international data set for clinical presentation, underlying diseases and provoking conditions, treatment and outcome at hospital discharge among children with CSVT and to define risk factors for death and abnormal neurological status at hospital discharge.
METHODS AND DESIGN Ethics This study was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. Research ethics approval was obtained at each site by the local institutional review boards to include data of consecutively enrolled patients in the IPSS database maintained in Toronto, Canada.
Ichord RN, et al. Arch 2014. Dis ChildProduced 2014;0:1–6. doi:10.1136/archdischild-2014-306382 1 Copyright Article author (or their employer) by BMJ Publishing Group Ltd (& RCPCH) under licence.
Downloaded from adc.bmj.com on October 11, 2014 - Published by group.bmj.com
Original article Setting The IPSS is an international paediatric stroke registry established in 2003 with the purpose of collecting data on demographic, clinical and radiological features, risk factors, treatment and outcomes at the time of hospital discharge among children from birth up to the age of 19 years with AIS and CSVT. We used the same methodology for enrolment, data abstraction and database management as previously described.13 This analysis included consecutively enrolled cases submitted to the database between 1 January 2003 and 1 July 2007. Sample size was determined by the number of cases enrolled in that time interval that met inclusion criteria. Local treating physicians determined diagnostic and treatment decisions.
Patients Site investigators identified cases and determined eligibility.14 Patients were included in the current childhood cohort study who met consensus-based, published clinical and radiological criteria for the diagnosis of childhood CSVT,6 as follows: (1) age >28 days up to 19 years at onset of symptoms, (2) acute neurological syndrome with signs and symptoms of CSVT, such as altered mental status, headache, papilloedema, seizures or focal neurological deficits and (3) radiological confirmation of CSVT by one or more modalities (MRI with MR venogram, CT venogram or cerebral catheter venography) showing thrombus or flow interruption within cerebral veins or dural venous sinuses, with or without venous distribution infarction.
Data collection
outcomes at intervals past hospital discharge were beyond the scope of the IPSS registry.
Statistical analysis Characteristics of the patient population were summarised with descriptive statistics, expressed as medians and IQRs. All statistical analyses were performed using the StatView 5 software package (SAS Institute, Cary, North Carolina, USA) and MedCalc (V.11.1.1., Mariakerke, Belgium). Continuous data are presented as median values and IQR. For categorical data, frequency distributions were compared between groups using χ2 test or Fisher’s exact test (F), as appropriate. The criterion for statistical significance was set at p=0.05 based on two-sided test. In order to evaluate (i) clinical symptoms (no clinical symptoms vs. any clinical symptoms), (ii) underlying medical conditions (no underlying medical condition vs. any underlying medical condition), (iii) thrombophilic risk factors (no thrombophilia vs. any thrombophilia) and (iv) initial anticoagulation (UFH or LMWH vs. no anticoagulation), analysis was performed using logistic regression. Adjustment variables for anticoagulation included recruitment areas (Australia, Canada, Europe, South America, the USA). Risk ratios and adjusted risks were expressed as ORs together with 95% CIs. Clustering by centre was not accounted for in the analysis because of small numbers contributed by each centre.
RESULTS Patient demographics Investigators at 30 sites in 10 countries enrolled 1187 evaluable cases of AIS and CSVT between January 2003 and July 2007 (https://app3.ccb.sickkids.ca/cstrokestudy/). Figure 1 shows enrolment characteristics by age group (neonate vs. nonneonate) and diagnostic subtype. A total of 170 patients aged 1 month to 19 years with a diagnosis of childhood CSVT were enrolled, representing 65% of patients in the IPSS cohort with CSVT and 14% of the combined IPSS cohort with either AIS or CSVT of all ages. Of these, 162 had CSVT alone and 8 had combined AIS and CSVT (figure 1). There was a male predominance, with 102 boys (60%) (table 1). The median age was 7.2 (IQR 2.9–12.4) years, with 25% of children under the age of 3 years.
Site investigators abstracted data from medical records onto IPSS case report forms (CRF). Data were entered via a webbased data entry system to the IPSS database, created and maintained in the IPSS data coordinating centre at the Toronto Hospital for Sick Children. Clinical data abstracted included demographics (age, gender, country of origin); clinical symptoms at presentation (altered mental status, headache, seizures, focal neurological deficits); underlying disease conditions (heart disease, genetic disorder, connective tissue disorder, prothrombotic state (including acquired and genetic thrombophilias), haematological disorder, malignancy, pre-existing cerebrovascular disorder); acute illness or provoking condition (systemic disorders, head or neck infections, trauma, treatment with a prothrombotic drug such as L-asparaginase, cranial surgery); neuroimaging findings and treatment modalities (antithrombotic treatment with unfractionated heparin (UFH) or low molecular weight heparin (LMWH), anticonvulsants, antibiotics, immunomodulatory therapy, neurosurgical interventions, transfusion or exchange transfusion). Initial and secondary antithrombotic treatments were reported, where secondary antithrombotic treatment refers to a treatment change from the initial agent to a different agent during admission. Data on prothrombotic states were reported to the registry database only for abnormal results. Practices varied widely across centres regarding testing for prothrombotic states. Data concerning the testing performed on each patient were not available as this was beyond the scope of the IPSS protocol. Data on neuroimaging included site of thrombosis, presence of infarction and presence and type of intracranial haemorrhage. Site of CSVT was classified as involving superficial or deep venous systems or both. Outcome data were collected on discharge disposition (home, inpatient rehabilitation unit, nursing facility), vital status at hospital discharge (alive or dead) and neurological status at discharge classified as normal or abnormal. Data on standardised functional
Table 1 summarises clinical and radiological characteristics, risk factors and discharge status. Common presenting signs and symptoms in children with CSVT included headache, altered consciousness, focal deficits and seizures. The superficial dural sinuses alone were most commonly involved (56%). Cerebral venous infarction affected 37% and intracranial haemorrhage 31% of cases. Clinical neurological status at discharge was normal in 48%, abnormal in 43% and unknown or not reported in 5%. Six children (4%) died before hospital discharge. In total, 84% of children were discharged to home, while 7% went to inpatient rehabilitation and 6% to another hospital or nursing facility. Chronic health disorders were reported in 50% of cases; acute systemic illness or acute head/neck disorders in 41% and neither or unknown in 9%. A prothrombotic state was identified in 34/170 (20%) of cases, 28 with isolated prothrombotic states without other chronic underlying conditions and the remainder associated with other chronic underlying disorders. As diagnostic practices varied widely across centres, all subjects were not tested for all prothrombotic factors. Data collection regarding
2
Ichord RN, et al. Arch Dis Child 2014;0:1–6. doi:10.1136/archdischild-2014-306382
Clinical and radiological features, risk factors, treatment and outcome
Downloaded from adc.bmj.com on October 11, 2014 - Published by group.bmj.com
Original article Figure 1 Enrolment characteristics of International Paediatric Stroke Study (IPSS) registry. DNQ, did not qualify or excluded for insufficient data.
The aim of this study was to describe clinical symptoms, underlying medical conditions, anticoagulant treatment and acute neurological outcome in 170 children with CSVT from a multicentre international registry. This is one of the largest cohort studies with the most diverse geographical representation published to date. Our study extends the observations of previously published smaller studies in children with CSVT that this is a morbid disease with diverse underlying causes and risk factors. The age distribution with median age of 7 years and male predominance of 60% are comparable with previous reports.2–5 As seen in other studies, acute head and neck disorders in otherwise healthy children are dominant causes of childhood CSVT. Certain chronic diseases known to predispose to CSVT are clearly represented in this cohort, including congenital heart disease, inflammatory bowel disease, haematological disorders and malignancy. Presenting signs and symptoms were dominated by non-specific symptoms of headache and altered
consciousness, with focal neurological deficits (43%) more common than reported in prior studies. Cerebral venous infarction or haemorrhage or both were more common than in previously published studies, possibly reflecting widespread use of MRI. The mortality rate of 4% is comparable to that reported in recent prospective cohort studies from Europe and Canada.3 4 Abnormal neurological status at discharge was documented in 43% of children in this study, further illustrating the serious nature of this disease. While long-term outcome data were not available in our study, these observations are consistent with reported rates of long-term neurological impairments and neurocognitive deficits in 40%–60% of survivors.3 5 8 12 15 16 Evaluation of risk factors for poor outcome or death identified coma as a risk factor for death and neurological sequelae, although this is only statistically significant for the latter group, probably because of the small number of deaths. Our finding of increased mortality in patients not treated with systemic anticoagulation is in keeping with other cohort studies showing probable benefit of anticoagulation.7 Further study in a larger cohort or in a clinical trial where confounding factors such as underlying disease, presence of infarction or haemorrhage can be measured and controlled for in the analysis would be helpful in clarifying the effect of anticoagulation treatment on mortality and long-term functional outcome. We found significant regional differences in antithrombotic treatment practices. A significant difference in the mode of initial anticoagulant administration was found with respect to different patient recruitment areas. In Canada, Europe and the USA, children with CSVT onset were treated with LMWH in approximately 60%–69% of cases, while in Australia and South America the use of UFH was more frequent. Our study has a number of strengths and limitations. Important limitations include wide variations in practices concerning evaluation of prothrombotic risk factors, differences across centres in modality and timing of imaging, choice of treatment and the lack of data on long-term outcome. Central review and adjudication of imaging would be ideal for any cohort study of paediatric CSVT, but it was beyond the scope of our study. The major strengths of our study are its large sample size and representation of multiple international centres. Notably, our study includes a large number of patients from the USA and other regions of the world (South America), which have previously been under-represented in cohort studies of paediatric CSVT. This adds valuable new data regarding a rare condition, thereby improving knowledge of risk factors, current treatment
Ichord RN, et al. Arch Dis Child 2014;0:1–6. doi:10.1136/archdischild-2014-306382
3
which patients were tested for each prothrombotic factor was not available as this was beyond the scope of the IPSS protocol. Specific prothrombotic abnormalities were reported by annotations on the CRFs, detailed in table 1. Haematological malignancy and iron deficiency were each reported in around 10% of cases, but only one patient had a haemoglobinopathy (sickle cell anaemia). The anatomical extent of the thrombosis was not associated with presence of prothrombotic factors. Table 2 summarises initial and secondary antithrombotic treatment. Antithrombotic treatment was given in 141 of the 170 (83%) children, of whom 33 (28%) had intracranial haemorrhage present at the time of treatment. Antithrombotic treatment was withheld in 29/170 children, of whom 10 had intracranial haemorrhage. Among the 6 children who died, 3 received antithrombotic treatment.
Clinical–radiological associations and outcome predictors Table 3 shows the associations of clinical and radiological features with outcome. There was an association of death and nontreatment with anticoagulation, with an OR of 5.2. Abnormal neurological status at discharge was associated with decreased level of consciousness at presentation, but was not associated with anticoagulation treatment. There was no association of death or abnormal neurological status with presence of an underlying chronic disease, extent of thrombosis or presence of intracranial haemorrhage at initial presentation.
DISCUSSION
Downloaded from adc.bmj.com on October 11, 2014 - Published by group.bmj.com
Original article Table 1 status
Demographic, clinical and radiological characteristics and discharge
7.2 years (2.9–12.4 years) 102 (60) 2 (1) 9 (5) 46 (27) 19 (11) 9 (5) 85 (50) 105/137 (76) 83/150 (55) 73/168 (43) 43/167 (26) 50/116 (43)
PA1 mutation Thrombocytosis Other Haematological disorder Iron deficiency anaemia Haemoglobinopathy Haematological malignancy Immunological disease‡ Cardiac disease Extracranial solid tumour Other§
3 1 0 32/170 (19) 15/170 (9) 1/170 (1) 16/170 (10) 6/170 (4) 5/167 (3) 2/170 (2) 8/170 (5)
Data are given as N (%) unless otherwise specified. *Risk factors and prothrombotic factors are not mutually exclusive. Multiple risk factors coexist in many patients. †Otitis, sinusitis and mastoiditis. ‡Inflammatory bowel disease, Kawasaki syndrome and macrophage activation syndrome. §Osteogenesis imperfecta, brain tumour, vein of Galen malformation and oral contraceptive use.
94/168 (56) 32/168 (19) 41/168 (24) 57/154 (37) 45/147 (31) 19/147 (13) 22/147 (15) 4/147 (3)
practices and short-term outcomes, and making the cumulative published data more widely generalisable. We found that children are more often treated with anticoagulation compared with neonates in the same centres. Jordan et al reported that 52% of neonates with CSVT enrolled in the IPSS database by the same centres as in our study received antithrombotic therapy compared with 83% of children with CSVT reported in our study. The basis for this finding cannot be determined from our study. Differences between neonates and children concerning risk factor profiles and prevalence of intracranial haemorrhage may have led to different judgements by practitioners about the potential risks vs. benefits of antithrombotic treatment. Moreover, the current treatment guidelines for the treatment of CSVT are more conservative for neonates compared with children.17 18 While the safety of anticoagulation for
131/156 (84) 11/156 (7) 10/156 (6) 4/75 (5) 81/167 (48) 72/167 (43) 6/167 (4) 8/167 (5)
Table 2 78/169 (46) 14/169 (8)
Antithrombotic treatment N (%)
Initial antithrombotic therapy Any antithrombotic therapy 141/170 UFH 55/170 LMWH 117/170 Warfarin 16/170 Aspirin 16/170 Secondary antithrombotic therapy LMWH 34/141 Warfarin 14/141 Aspirin 11/141 Antithrombotic therapy in all children according to geographical region
34/165 (17) 4/165 (2) 6/165 (4) 4/165 (2) 19/169(11) 2/169 (1) 34/170 (20) 2 0 5 4 9 2 2 12 1
(83) (32) (69) (9) (9) (24) (10) (8)
Medication chosen
Continued
4
Continued
Demographics
Demographics Age, median (IQR) Gender: number of men Region of origin Asia Australia Canada Europe South America USA Clinical signs and symptoms Headache Altered consciousness Focal deficits Seizures Treatment with anticonvulsants Radiological features Main thrombotic locations Superficial venous system Deep venous system Superficial and deep Infarct present Haemorrhage present Haemorrhagic infarct only Other haemorrhage Both Outcome Disposition at hospital discharge Home Rehabilitation hospital Other hospital Other Status at discharge Normal Neurological deficit Died Unknown Risk factors* Acute provoking illness Acute head/neck infection† Meningitis Acute systemic illness or injury Dehydration Acidosis Anoxia Shock Trauma Central venous line Underlying chronic disease condition Prothrombotic state* Protein C or S deficiency Antithrombin deficiency Factor V Leiden Prothrombin mutation Lipoprotein (a) elevation Antiphospholipid antibody Lupus anticoagulant MTHFR mutation Factor VIII elevation
Table 1
Geographic region
UFH
LMWH
p Value
Australia Canada Europe South America USA
10/11 (91) 25/65 (39) 8/26 (31) 5/7 (71) 21/75 (28)
1/11 40/65 18/26 2/7 46/75
ADC Online First, published on October 6, 2014 as 10.1136/archdischild-2014-306382 Original article
Paediatric cerebral sinovenous thrombosis: findings of the International Paediatric Stroke Study R N Ichord,1 S L Benedict,2 A K Chan,3 F J Kirkham,4,5 U Nowak-Göttl,6 for the International Paediatric Stroke Study Group 1
Department Neurology & Pediatrics, Children’s Hospital of Philadelphia, Perelman School of Medicine of the University of Pennsylvania, Philadelphia, Pennsylvania, USA 2 Department Neurology, Primary Children’s Hospital, University of Utah, Salt Lake City, Utah, USA 3 Division of Hematology and Oncology, Department of Pediatrics, McMaster University, Hamilton, Ontario, Canada 4 Neurosciences Unit, University College London Institute of Child Health, London, UK 5 University hospital Southampton NHS Foundation Trust, Southampton, UK 6 Department of Pediatric Hematology/Oncology, Universitätsklinikum SchleswigHolstein, Kiel, Germany Correspondence to Dr Rebecca Ichord, Department Neurology CTRB 10th Fl, Children’s Hospital of Philadelphia, 3501 Civic Ctr Blvd, Philadelphia, PA 19104, USA; [email protected] FJK and UNG shared senior authorship. Received 10 March 2014 Revised 7 September 2014 Accepted 16 September 2014
ABSTRACT Objectives We evaluated clinical features, treatment practices and early outcome in a multicentre cohort of children with cerebral sinovenous thrombosis (CSVT). Methods Children with CSVT from 10 countries were enrolled from January 2003 to July 2007 in the International Paediatric Stroke Study. We analysed clinical symptoms, underlying conditions, antithrombotic treatment and neurological outcome at hospital discharge in 170 children. Results Of 170 children enrolled, 60% were male; median age 7.2 years (IQR 2.9–12.4). Headache, altered consciousness, focal deficits and seizures were common presenting clinical features. Infarction affected 37% and intracranial haemorrhage 31%. Risk factors included chronic disease in 50%; acute systemic illness or head/ neck disorders 41%; prothrombotic state 20% and other haematological abnormality 19%. Discharge neurological status was normal in 48%, abnormal in 43% and unknown in 5%. Antithrombotic therapy was common, most often low molecular weight heparin was common, with significant regional variation in treatment practices. Mortality was low (4%) and was associated with no anticoagulation but not underlying chronic disease, anatomic extent of thrombosis or intracranial haemorrhage. Abnormal neurological status at discharge or death was associated with decreased level of consciousness at presentation and the presence of an identified prothrombotic state. Conclusions Our study extends the observations of previously published smaller studies in children with CSVT that this is a morbid disease with diverse underlying causes and risk factors. Divergent treatment practices among highly specialised centres as well as limited data on treatment efficacy and safety suggest that further study of this condition is warranted.
INTRODUCTION
To cite: Ichord RN, Benedict SL, Chan AK, et al. Arch Dis Child Published Online First: [ please include Day Month Year] doi:10.1136/archdischild2014-306382
Cerebral sinovenous thrombosis (CSVT), defined as blood clot in a vein carrying blood from the brain to the heart, affects 0.34–0.67/100 000 children annually, of which 27%–35% are neonates.1–3 Mortality is 3%–12%, and neurological sequelae affect up to 62% of survivors.4–6 Aetiology in the paediatric population includes a wide spectrum of underlying conditions and symptoms at clinical presentation.1–3 5 7 Recent studies suggest that haematological abnormalities, including prothrombotic disorders and iron deficiency, contribute to development of CSVT.1 5 8–12 Previously published cohort studies have described epidemiology, treatment practices and outcomes, but are limited due to small sample size or single-centre enrolment.
What is already known on this topic? ▸ Cerebral sinovenous thrombosis in childhood is fatal in 3%–12% and causes neurological sequelae in up to 62% of survivors. ▸ Previously published cohort studies have described epidemiology, treatment practices and outcomes, but are limited due to small sample size or single-centre enrolment. ▸ Uncertainty remains regarding risk factors for death and adverse neurological outcome and optimal treatment.
What this study adds? ▸ The major strengths of our study are its large sample size and representation of multiple international centres not included in smaller studies. ▸ This adds valuable new data regarding a rare condition, improving knowledge of risk factors, current treatment practices and outcomes. ▸ Our finding of increased mortality in patients not treated with systemic anticoagulation supports previous observations that have suggested there is probable benefit of anticoagulation.
The International Paediatric Stroke Study (IPSS) established a prospective multicentre registry to characterise epidemiology, treatment and outcome of infants and children with arterial ischaemic stroke (AIS) and CSVT. The objective of this paper is to analyse this large multicentre international data set for clinical presentation, underlying diseases and provoking conditions, treatment and outcome at hospital discharge among children with CSVT and to define risk factors for death and abnormal neurological status at hospital discharge.
METHODS AND DESIGN Ethics This study was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki. Research ethics approval was obtained at each site by the local institutional review boards to include data of consecutively enrolled patients in the IPSS database maintained in Toronto, Canada.
Ichord RN, et al. Arch 2014. Dis ChildProduced 2014;0:1–6. doi:10.1136/archdischild-2014-306382 1 Copyright Article author (or their employer) by BMJ Publishing Group Ltd (& RCPCH) under licence.
Downloaded from adc.bmj.com on October 11, 2014 - Published by group.bmj.com
Original article Setting The IPSS is an international paediatric stroke registry established in 2003 with the purpose of collecting data on demographic, clinical and radiological features, risk factors, treatment and outcomes at the time of hospital discharge among children from birth up to the age of 19 years with AIS and CSVT. We used the same methodology for enrolment, data abstraction and database management as previously described.13 This analysis included consecutively enrolled cases submitted to the database between 1 January 2003 and 1 July 2007. Sample size was determined by the number of cases enrolled in that time interval that met inclusion criteria. Local treating physicians determined diagnostic and treatment decisions.
Patients Site investigators identified cases and determined eligibility.14 Patients were included in the current childhood cohort study who met consensus-based, published clinical and radiological criteria for the diagnosis of childhood CSVT,6 as follows: (1) age >28 days up to 19 years at onset of symptoms, (2) acute neurological syndrome with signs and symptoms of CSVT, such as altered mental status, headache, papilloedema, seizures or focal neurological deficits and (3) radiological confirmation of CSVT by one or more modalities (MRI with MR venogram, CT venogram or cerebral catheter venography) showing thrombus or flow interruption within cerebral veins or dural venous sinuses, with or without venous distribution infarction.
Data collection
outcomes at intervals past hospital discharge were beyond the scope of the IPSS registry.
Statistical analysis Characteristics of the patient population were summarised with descriptive statistics, expressed as medians and IQRs. All statistical analyses were performed using the StatView 5 software package (SAS Institute, Cary, North Carolina, USA) and MedCalc (V.11.1.1., Mariakerke, Belgium). Continuous data are presented as median values and IQR. For categorical data, frequency distributions were compared between groups using χ2 test or Fisher’s exact test (F), as appropriate. The criterion for statistical significance was set at p=0.05 based on two-sided test. In order to evaluate (i) clinical symptoms (no clinical symptoms vs. any clinical symptoms), (ii) underlying medical conditions (no underlying medical condition vs. any underlying medical condition), (iii) thrombophilic risk factors (no thrombophilia vs. any thrombophilia) and (iv) initial anticoagulation (UFH or LMWH vs. no anticoagulation), analysis was performed using logistic regression. Adjustment variables for anticoagulation included recruitment areas (Australia, Canada, Europe, South America, the USA). Risk ratios and adjusted risks were expressed as ORs together with 95% CIs. Clustering by centre was not accounted for in the analysis because of small numbers contributed by each centre.
RESULTS Patient demographics Investigators at 30 sites in 10 countries enrolled 1187 evaluable cases of AIS and CSVT between January 2003 and July 2007 (https://app3.ccb.sickkids.ca/cstrokestudy/). Figure 1 shows enrolment characteristics by age group (neonate vs. nonneonate) and diagnostic subtype. A total of 170 patients aged 1 month to 19 years with a diagnosis of childhood CSVT were enrolled, representing 65% of patients in the IPSS cohort with CSVT and 14% of the combined IPSS cohort with either AIS or CSVT of all ages. Of these, 162 had CSVT alone and 8 had combined AIS and CSVT (figure 1). There was a male predominance, with 102 boys (60%) (table 1). The median age was 7.2 (IQR 2.9–12.4) years, with 25% of children under the age of 3 years.
Site investigators abstracted data from medical records onto IPSS case report forms (CRF). Data were entered via a webbased data entry system to the IPSS database, created and maintained in the IPSS data coordinating centre at the Toronto Hospital for Sick Children. Clinical data abstracted included demographics (age, gender, country of origin); clinical symptoms at presentation (altered mental status, headache, seizures, focal neurological deficits); underlying disease conditions (heart disease, genetic disorder, connective tissue disorder, prothrombotic state (including acquired and genetic thrombophilias), haematological disorder, malignancy, pre-existing cerebrovascular disorder); acute illness or provoking condition (systemic disorders, head or neck infections, trauma, treatment with a prothrombotic drug such as L-asparaginase, cranial surgery); neuroimaging findings and treatment modalities (antithrombotic treatment with unfractionated heparin (UFH) or low molecular weight heparin (LMWH), anticonvulsants, antibiotics, immunomodulatory therapy, neurosurgical interventions, transfusion or exchange transfusion). Initial and secondary antithrombotic treatments were reported, where secondary antithrombotic treatment refers to a treatment change from the initial agent to a different agent during admission. Data on prothrombotic states were reported to the registry database only for abnormal results. Practices varied widely across centres regarding testing for prothrombotic states. Data concerning the testing performed on each patient were not available as this was beyond the scope of the IPSS protocol. Data on neuroimaging included site of thrombosis, presence of infarction and presence and type of intracranial haemorrhage. Site of CSVT was classified as involving superficial or deep venous systems or both. Outcome data were collected on discharge disposition (home, inpatient rehabilitation unit, nursing facility), vital status at hospital discharge (alive or dead) and neurological status at discharge classified as normal or abnormal. Data on standardised functional
Table 1 summarises clinical and radiological characteristics, risk factors and discharge status. Common presenting signs and symptoms in children with CSVT included headache, altered consciousness, focal deficits and seizures. The superficial dural sinuses alone were most commonly involved (56%). Cerebral venous infarction affected 37% and intracranial haemorrhage 31% of cases. Clinical neurological status at discharge was normal in 48%, abnormal in 43% and unknown or not reported in 5%. Six children (4%) died before hospital discharge. In total, 84% of children were discharged to home, while 7% went to inpatient rehabilitation and 6% to another hospital or nursing facility. Chronic health disorders were reported in 50% of cases; acute systemic illness or acute head/neck disorders in 41% and neither or unknown in 9%. A prothrombotic state was identified in 34/170 (20%) of cases, 28 with isolated prothrombotic states without other chronic underlying conditions and the remainder associated with other chronic underlying disorders. As diagnostic practices varied widely across centres, all subjects were not tested for all prothrombotic factors. Data collection regarding
2
Ichord RN, et al. Arch Dis Child 2014;0:1–6. doi:10.1136/archdischild-2014-306382
Clinical and radiological features, risk factors, treatment and outcome
Downloaded from adc.bmj.com on October 11, 2014 - Published by group.bmj.com
Original article Figure 1 Enrolment characteristics of International Paediatric Stroke Study (IPSS) registry. DNQ, did not qualify or excluded for insufficient data.
The aim of this study was to describe clinical symptoms, underlying medical conditions, anticoagulant treatment and acute neurological outcome in 170 children with CSVT from a multicentre international registry. This is one of the largest cohort studies with the most diverse geographical representation published to date. Our study extends the observations of previously published smaller studies in children with CSVT that this is a morbid disease with diverse underlying causes and risk factors. The age distribution with median age of 7 years and male predominance of 60% are comparable with previous reports.2–5 As seen in other studies, acute head and neck disorders in otherwise healthy children are dominant causes of childhood CSVT. Certain chronic diseases known to predispose to CSVT are clearly represented in this cohort, including congenital heart disease, inflammatory bowel disease, haematological disorders and malignancy. Presenting signs and symptoms were dominated by non-specific symptoms of headache and altered
consciousness, with focal neurological deficits (43%) more common than reported in prior studies. Cerebral venous infarction or haemorrhage or both were more common than in previously published studies, possibly reflecting widespread use of MRI. The mortality rate of 4% is comparable to that reported in recent prospective cohort studies from Europe and Canada.3 4 Abnormal neurological status at discharge was documented in 43% of children in this study, further illustrating the serious nature of this disease. While long-term outcome data were not available in our study, these observations are consistent with reported rates of long-term neurological impairments and neurocognitive deficits in 40%–60% of survivors.3 5 8 12 15 16 Evaluation of risk factors for poor outcome or death identified coma as a risk factor for death and neurological sequelae, although this is only statistically significant for the latter group, probably because of the small number of deaths. Our finding of increased mortality in patients not treated with systemic anticoagulation is in keeping with other cohort studies showing probable benefit of anticoagulation.7 Further study in a larger cohort or in a clinical trial where confounding factors such as underlying disease, presence of infarction or haemorrhage can be measured and controlled for in the analysis would be helpful in clarifying the effect of anticoagulation treatment on mortality and long-term functional outcome. We found significant regional differences in antithrombotic treatment practices. A significant difference in the mode of initial anticoagulant administration was found with respect to different patient recruitment areas. In Canada, Europe and the USA, children with CSVT onset were treated with LMWH in approximately 60%–69% of cases, while in Australia and South America the use of UFH was more frequent. Our study has a number of strengths and limitations. Important limitations include wide variations in practices concerning evaluation of prothrombotic risk factors, differences across centres in modality and timing of imaging, choice of treatment and the lack of data on long-term outcome. Central review and adjudication of imaging would be ideal for any cohort study of paediatric CSVT, but it was beyond the scope of our study. The major strengths of our study are its large sample size and representation of multiple international centres. Notably, our study includes a large number of patients from the USA and other regions of the world (South America), which have previously been under-represented in cohort studies of paediatric CSVT. This adds valuable new data regarding a rare condition, thereby improving knowledge of risk factors, current treatment
Ichord RN, et al. Arch Dis Child 2014;0:1–6. doi:10.1136/archdischild-2014-306382
3
which patients were tested for each prothrombotic factor was not available as this was beyond the scope of the IPSS protocol. Specific prothrombotic abnormalities were reported by annotations on the CRFs, detailed in table 1. Haematological malignancy and iron deficiency were each reported in around 10% of cases, but only one patient had a haemoglobinopathy (sickle cell anaemia). The anatomical extent of the thrombosis was not associated with presence of prothrombotic factors. Table 2 summarises initial and secondary antithrombotic treatment. Antithrombotic treatment was given in 141 of the 170 (83%) children, of whom 33 (28%) had intracranial haemorrhage present at the time of treatment. Antithrombotic treatment was withheld in 29/170 children, of whom 10 had intracranial haemorrhage. Among the 6 children who died, 3 received antithrombotic treatment.
Clinical–radiological associations and outcome predictors Table 3 shows the associations of clinical and radiological features with outcome. There was an association of death and nontreatment with anticoagulation, with an OR of 5.2. Abnormal neurological status at discharge was associated with decreased level of consciousness at presentation, but was not associated with anticoagulation treatment. There was no association of death or abnormal neurological status with presence of an underlying chronic disease, extent of thrombosis or presence of intracranial haemorrhage at initial presentation.
DISCUSSION
Downloaded from adc.bmj.com on October 11, 2014 - Published by group.bmj.com
Original article Table 1 status
Demographic, clinical and radiological characteristics and discharge
7.2 years (2.9–12.4 years) 102 (60) 2 (1) 9 (5) 46 (27) 19 (11) 9 (5) 85 (50) 105/137 (76) 83/150 (55) 73/168 (43) 43/167 (26) 50/116 (43)
PA1 mutation Thrombocytosis Other Haematological disorder Iron deficiency anaemia Haemoglobinopathy Haematological malignancy Immunological disease‡ Cardiac disease Extracranial solid tumour Other§
3 1 0 32/170 (19) 15/170 (9) 1/170 (1) 16/170 (10) 6/170 (4) 5/167 (3) 2/170 (2) 8/170 (5)
Data are given as N (%) unless otherwise specified. *Risk factors and prothrombotic factors are not mutually exclusive. Multiple risk factors coexist in many patients. †Otitis, sinusitis and mastoiditis. ‡Inflammatory bowel disease, Kawasaki syndrome and macrophage activation syndrome. §Osteogenesis imperfecta, brain tumour, vein of Galen malformation and oral contraceptive use.
94/168 (56) 32/168 (19) 41/168 (24) 57/154 (37) 45/147 (31) 19/147 (13) 22/147 (15) 4/147 (3)
practices and short-term outcomes, and making the cumulative published data more widely generalisable. We found that children are more often treated with anticoagulation compared with neonates in the same centres. Jordan et al reported that 52% of neonates with CSVT enrolled in the IPSS database by the same centres as in our study received antithrombotic therapy compared with 83% of children with CSVT reported in our study. The basis for this finding cannot be determined from our study. Differences between neonates and children concerning risk factor profiles and prevalence of intracranial haemorrhage may have led to different judgements by practitioners about the potential risks vs. benefits of antithrombotic treatment. Moreover, the current treatment guidelines for the treatment of CSVT are more conservative for neonates compared with children.17 18 While the safety of anticoagulation for
131/156 (84) 11/156 (7) 10/156 (6) 4/75 (5) 81/167 (48) 72/167 (43) 6/167 (4) 8/167 (5)
Table 2 78/169 (46) 14/169 (8)
Antithrombotic treatment N (%)
Initial antithrombotic therapy Any antithrombotic therapy 141/170 UFH 55/170 LMWH 117/170 Warfarin 16/170 Aspirin 16/170 Secondary antithrombotic therapy LMWH 34/141 Warfarin 14/141 Aspirin 11/141 Antithrombotic therapy in all children according to geographical region
34/165 (17) 4/165 (2) 6/165 (4) 4/165 (2) 19/169(11) 2/169 (1) 34/170 (20) 2 0 5 4 9 2 2 12 1
(83) (32) (69) (9) (9) (24) (10) (8)
Medication chosen
Continued
4
Continued
Demographics
Demographics Age, median (IQR) Gender: number of men Region of origin Asia Australia Canada Europe South America USA Clinical signs and symptoms Headache Altered consciousness Focal deficits Seizures Treatment with anticonvulsants Radiological features Main thrombotic locations Superficial venous system Deep venous system Superficial and deep Infarct present Haemorrhage present Haemorrhagic infarct only Other haemorrhage Both Outcome Disposition at hospital discharge Home Rehabilitation hospital Other hospital Other Status at discharge Normal Neurological deficit Died Unknown Risk factors* Acute provoking illness Acute head/neck infection† Meningitis Acute systemic illness or injury Dehydration Acidosis Anoxia Shock Trauma Central venous line Underlying chronic disease condition Prothrombotic state* Protein C or S deficiency Antithrombin deficiency Factor V Leiden Prothrombin mutation Lipoprotein (a) elevation Antiphospholipid antibody Lupus anticoagulant MTHFR mutation Factor VIII elevation
Table 1
Geographic region
UFH
LMWH
p Value
Australia Canada Europe South America USA
10/11 (91) 25/65 (39) 8/26 (31) 5/7 (71) 21/75 (28)
1/11 40/65 18/26 2/7 46/75
