Stroke Patterns in Neonatal Group B Streptococcal

Stroke Patterns in Neonatal Group B Streptococcal Meningitis Marta I. Hern andez, MD*, Carmen C. Sandoval, MD†, Jose L. Tapia, MD‡, Tomas Mesa, MD*, Raul Escobar, MD*, Isidro Huete, MD§, Xing-Chang Wei, MD¶, and Adam Kirton, MD, MSc# Neonatal group B streptococcus meningitis causes neurologic morbidity and mortality. Cerebrovascular involvement is a common, poorly studied, and potentially modifiable pathologic process. We hypothesized that imaging patterns of focal brain infarction are recognizable in neonatal group B streptococcal meningitis. A consecutive case series included term neonates with the following: (1) bacterial meningitis, (2) acute group B streptococcal infection (positive cerebrospinal fluid/ blood culture), (3) brain magnetic resonance imaging within 14 days, and (4) acute intraparenchymal focal infarctions (restricted diffusion). Lesions within known arterial territories were classified as arterial ischemic stroke. Clinical presentations, investigations, and neurologic outcomes were recorded. Eight newborns (50% female) with focal infarction were identified. Five presented early (3 months) by interpersonal or nosocomial horizontal transmission. Despite screening and antibiotic prophylaxis, recent rates of early GBS meningitis seem to be increasing in the United States [3]. Neurologic consequences of bacterial meningeal inflammation include extension to brain parenchyma (cerebritis), blood-brain barrier disruption (cerebral edema), and impaired cerebrospinal fluid (CSF) circulation resulting in hydrocephalus. The effects of bacterial meningitis on the cerebral vasculature and resulting ischemic stroke are less well understood. Subarachnoid inflammation directly approximates the major vessels of the circle of Willis and in particular the smaller perforating lenticulostriate and thalamostriate arteries. Over the surface of the brain, both pial arteries and the venous system (cortical veins and dural venous sinuses) are potentially affected. Such vasculopathy induced by regional infection and inflammation may combine with prothrombotic processes to result in arterial and/or venous infarction. Such processes likely begin early in neonatal meningitis but may persist even after sterilization of the CSF [4]. Large studies of neonatal stroke (both arterial and venous) confirm an association with bacterial meningitis [5,6]. However, to our knowledge, no study has explored

From the *Department of Pediatrics †Neonatal Unit, ‡Department of Pediatric Neonatology, and §Department of Radiology, Pontificia Universidad Cat olica de Chile, Santiago, Chile; and ¶Department of Radiology and #Calgary Pediatric Stroke Program, Division of Neurology, Alberta Children’s Hospital, University of Calgary, Calgary, Alberta, Canada.

Communication should be addressed to: Dr. Kirton; Division of Neurology; Alberta Children’s Hospital; 2888 Shaganappi Trail NW; Calgary, AB, Canada T3B 6A8. E-mail: [email protected] Received July 6, 2010; accepted November 8, 2010.

282 PEDIATRIC NEUROLOGY Vol. 44 No. 4

Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.pediatrneurol.2010.11.002 0887-8994/$ - see front matter

Downloaded from ClinicalKey.com at University of Chile Catholic ALERTA August 30, 2016. For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.

stroke patterns specific to GBS meningitis. A large series of 166 children of all ages with meningitis described infarction in 10% [7]. Additional small series have described stroke in neonatal meningitis but have not looked at organism-specific disease [8]. Studies of neonatal stroke that result from meningitis have been further limited by heterogeneity of both the types of strokes and the bacterial agent responsible. Pathogen-specific patterns of morbidity in childhood meningitis are recognized and may dictate specific differences in management (e.g., corticosteroid therapy for Haemophilus influenzae meningitis) [9]. Ischemic stroke may represent an important, potentially treatable mechanism of brain injury in neonatal GBS meningitis. Dedicated analysis of specific outcomes such as stroke in specific types of meningitis is therefore required. We report a case series of acute focal brain infarction associated with neonatal GBS meningitis and describe 2 recognizable patterns of injury. Material and Methods Patient Identification Patients were identified from the registry of infant stroke database of Pontificia Universidad Catolica de Chile Hospital (2002-2009) and the Calgary Pediatric Stroke Program (2007-2009). Inclusion criteria included the following: (1) term-age neonate 50/mm3), (3) confirmation of GBS as responsible organism by positive CSF or blood culture, (4) magnetic resonance imaging (MRI) of the brain completed within 14 days of clinical onset of illness including diffusion weighted imaging, and (5) one or more intraparenchymal brain lesions demonstrating restricted diffusion consistent with focal infarction.

Data Collection All data elements were collected according to standardized institutional protocols. The lead authors (MH, AK) are coinvestigators in the International Pediatric Stroke Study (IPSS) [10], with both centers actively enrolling patients during the period of study. Patients with stroke are therefore evaluated according to protocol for stroke-related data elements. Data were extracted from the medical records directly in addition to other factors relevant to meningitis and the patient’s condition. Data elements included demographic, clinical presentation, infectious disease, stroke risk factors, diagnostic tests performed, blood and CSF testing results, and treatments provided. Neurologic outcome was based on most recent follow-up with the treating neurologist. Outcomes were dichotomized as good or poor on the basis of the presence or absence of clinically significant neurologic deficits or epilepsy.

Imaging Santiago magnetic resonance images were obtained by a 1.5 T superconducting system (Philips Achieva). Examinations included axial and coronal fast spin-echo T2-weighted (TR/TE, 2,270/100 ms; echo-train length, 30; 2 acquisitions; section thickness 5 mm; FOV 200 mm; matrix size 384 240), axial fast spin-echo T2 fluid-attenuated inversion recovery images (TR/TE, 11.000/127 ms, TI 2,800 ms; echo-train length 52; section thickness 5 mm; field of view 200 mm; matrix size 288 182); axial T1-weighted spin-echo images with and without gadolinium (gadoversetamide 0.5 mmol/ml; 0.2 mL/kg) (TR/TE, 450/15 ms, section thickness 5 mm, 1 acquisition), and axial diffusion-weighted images (b-factor

1,000 s/mm2, section thickness 5 mm, TR/TE 3,300/87 ms, 1 acquisition, and parallel imaging acceleration factor of 2). Calgary magnetic resonance images were obtained with a 1.5 T scanner (Siemens Avanto) with a 12-channel receiver-only phased-array surface coil and high-performance gradients (45 mT/m; 200-second rise time). Diffusion-weighted images (b-factor 0, 500, and 1,000 s/mm2, thickness/gap 5/1 mm, TR/TE 3,000/79 ms, number of excitations 4, and parallel imaging acceleration factor of 2) were acquired in axial and coronal planes. T1-weighted images were acquired as spin-echo, TR/TE 555/15 ms, thickness/gap 5/1 mm, number of excitation 2. T2-weighted images were acquired a fast spin-echo, TR/TE 4,590/99 ms, thickness/gap 5/1 mm, echo-train length 15, number of excitation 2. Fluid-attenuated inversion recovery images used TR/TE/TI 8,500/98/2,400 ms, ETL 13, NEX 2). Time-of-flight magnetic resonance angiography used 3-dimensional multiple overlapping thin slab acquisition, TR/TE 23/7 ms, flip angle 25 degrees. Slice thickness was 0.8 mm. All imaging was independently reviewed by the investigator neuroradiologists and 2 pediatric neurologists, at least 2 of whom were blinded to all other patient information. Lesions clearly respecting known arterial territories were classified as acute arterial ischemic stroke (AIS), with or without angiographic evidence of arterial occlusion. Lesions observed in the same venous territory where evidence of venous sinus thrombosis was confirmed on magnetic resonance venogram or computed tomographic venogram were classified as cerebral sinovenous thrombosis (CSVT). Focal areas of restricted diffusion limited to the cerebral cortex were classified simply as cortical infarctions because definitive distinction of arterial or venous ischemia from each other or direct cerebritis could not be made.

Results A total of 8 children were studied (50% female, 1 preterm). The population is summarized in Table 1. Case details are provided below. Prenatal GBS screening was completed in only 3 of 7 eligible cases (one child born at 31 weeks), and 2 of these were negative. Median age at presentation was 8 days (mean 9.7 days, range 2-28 days). Clinical presentations were typical and included fever, poor feeding, lethargy, and seizures. All of patients had evidence of elevated acute-phase reactants, hemodynamic instability, and/or septic shock at presentation. Clear imaging patterns of acute infarction were evident in all patients. Original MRI was completed at a median of 5 days (mean 5.8 days, range 3-14 days). A deep pattern of AIS in the territories of the penetrating lenticulostriate and thalamostriate arteries affecting the basal ganglia, thalamus, and deep white matter was seen in 7/8 children (88%; Fig 1). Superficial cortical infarction was observed in 6 children (75%; Fig 2), with all of the 4 lobes affected in at least one patient. Therefore, most children had combined deep and superficial patterns (5/8, 63%). No evidence of large artery ischemic stroke was observed. All children had time-of-flight magnetic resonance angiography completed in the acute setting, none of which demonstrated large artery disease. Definitive CSVT co-occurring in 1 patient was discovered on follow-up imaging and was not associated with parenchymal infarction. All children had follow-up MRI that demonstrated chronic sequelae of the previous injury, including atrophy of affected structures and hydrocephalus (Figs 1 and 2). Neurologic outcome was documented in all patients. Median time to follow-up was 24 months (mean 25.9

Hernandez et al: Neonatal Stroke in GBS Meningitis 283 Downloaded from ClinicalKey.com at University of Chile Catholic ALERTA August 30, 2016. For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.

Table 1. Summary of 8 cases of stroke associated with neonatal GBS meningitis Patient (time)

Stroke

Sex/Age

Clinical Findings

1

AIS

F/2 d

Fever, seizure, LOC, feeding intolerance Fever, seizure, feeding intolerance Fever, feeding intolerance Fever, LOC

2

AIS

F/3 d

3

AIS

M/2 d

4

AIS/CSVT F/28 d

5

AIS

M/14 d

Fever, seizures feeding intolerance

6

AIS

F/6 d

Fever, seizures

7

AIS

M/14 d* Fever, seizure, LOC

8

AIS

M/10 d

Fever, seizure

MRI

Acute DWI Lesions

3 d, 4 mo

Bilateral basal ganglia and occipital cortex 5 d, 20 d Bilateral basal ganglia and L thalamus 5 d, 19 d L parietal, occipital cortex. R atrial white matter 3 d, 15 d L + R frontal, R occipital cortex; sagittal sinus CSVT 14 d, 2 mo Bilateral frontal, parietal, occipital cortex and thalamus 8 d, 1 mo Bilateral basal ganglia 5 d, 12 mo Bilateral basal ganglia, thalamus, deep WM, temp + parietal cortex 4 d, 10 d, R thalamus, bilateral 6 wk occipital and frontal lobes

Pattern

Duration of Follow-up

D, S

60 mo

D

30 mo

D, S

10 mo

S

18 mo

D, S

48 mo

D

6 mo

D, S

32 mo

D, S

3 mo

Outcome CP (quadriplegia), visual impairment Delayed speech Global delays, epilepsy, hydrocephalus Global delays, visual impairment, hydrocephalus Global delays, epilepsy

Global delays, hydrocephalus, death CP (quadriplegia), global delays, hydrocephalus Normal†

* Corrected gestational age. † Short-term follow-up only. Abbreviations: AIS = Arterial ischemic stroke CP = Cerebral palsy CSVT = Cerebral sinovenous thrombosis D = Deep pattern of infarction (perforating arteries affecting basal ganglia thalamus deep white matter; see Fig 1) DWI = Diffusion weighted imaging LOC = Decreased level of consciousness MRI = Magnetic resonance imaging S = Superficial pattern of infarction (lesions limited to cortical surface; see Fig 2)

months, range 3-60 months). Neurologic outcomes were poor in 6 of 7 (86%) children with long-term follow-up (another ‘‘normal’’ child had only 3 months’ follow-up). Neurologic deficits included motor problems (cerebral palsy, 86%), language deficits (86%), global developmental delays (71%), visual impairment (29%), hydrocephalus (57%), and epilepsy (29%). No occurrences of recurrent stroke were described over this time period. Only 1 child was treated with antithrombotic therapy (ASA, 4 mg/kg/d) without complication. Case Reports CASE 1 A term female newborn was born to a healthy mother after an uneventful pregnancy. GBS screening at 38 weeks was negative. At 36 hours, she manifested grunting and fever, followed by partial seizures that progressed to status epilepticus. Initial analysis of blood revealed leukopenia (4,700 103/mm3), elevated C-reactive protein (CRP, 23 mg/dL), lactic acidosis (11 mmol/L), and positive blood culture for GBS. CSF revealed hypoglycorrhachia (0 mg/dL), pleocytosis (1,300 mm3, 91% neutrophils),

increased protein (272 mg/dL), and positive latex and culture for GBS. Treatment with ampicillin and gentamicin was initiated and maintained for 15 days. Triple-drug anticonvulsant therapy was required for seizure control, and mechanical ventilation was maintained for 10 days. MRI on day 3 revealed recent acute AIS of bilateral putamina, internal capsule, and occipital cortex. Follow-up MRI at 4 months demonstrated atrophy of the previously affected regions (Fig 1A). Neurologic evaluation at 5 years demonstrated global delays, spastic quadriplegia, and marked visual impairment. CASE 2 The mother of this term newborn girl had an uneventful pregnancy. GBS screening was positive at 34 weeks’ gestation, so she was treated with ampicillin, and repeated screening at 36 weeks was negative. On the third day of life, the infant manifested fever and lethargy. Initial analysis of blood revealed leucocytosis (22,000 103/mm3), CRP of 3.1 mg/dL, gram-positive bacilli, and positive GBS blood culture at 10 hours. Lumbar puncture could not initially be performed. Therapy with amikacin and ampicillin was initiated. Twenty-four hours later, she developed leftsided tonic-clonic seizures that resolved with phenobarbital.

284 PEDIATRIC NEUROLOGY Vol. 44 No. 4 Downloaded from ClinicalKey.com at University of Chile Catholic ALERTA August 30, 2016. For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.

Figure 1. Deep pattern of arterial ischemic stroke in neonatal group B streptococcus meningitis. Infarction occurs within the territories of small perforating arteries, presumably as a result of infection or inflammation of the subarachnoid space and meninges. As a result, deep gray structures (thalamus, basal ganglia) and deep white matter tracts are preferentially affected. Acute lesions demonstrate restricted diffusion (left), while follow-up studies demonstrate volume loss in these areas and ventricular dilatation (right). Examples are cases 1, 2, 6, and 7, respectively.

Figure 2. Superficial cortical infarction pattern in neonatal group B streptococcus meningitis. Acute lesions are evident over the cortical surface, often bilaterally. Differentiation of focal arterial ischemic stroke from superficial venous infarction, from direct extension of bacterial infection (cerebritis), is challenging. Some may conform to known arterial territories, but many do not and are diffuse and bilateral. Acute lesions demonstrate restricted diffusion (left), while follow-up studies demonstrate volume loss in these areas (right). Examples are cases 4, 5, 3, and 8, respectively.

Opisthotonic posturing was observed. MRI on day 5 revealed acute AIS in the caudate nuclei bilaterally and left thalamus (Fig 1B). Repeated MRI just 15 days later confirmed areas of encephalomalacia and gliosis in the previously infarcted regions. Follow-up at 30 months demonstrated language impairments requiring regular therapy. CASE 3 This term male newborn was born to a teenage mother who received no prenatal care. After an uncomplicated delivery at 37 weeks, the neonate developed lethargy,

poor feeding, grunting respirations, poor perfusion, and fever at 36 hours of life. Analysis of blood revealed WBC of 1,300 103/mm3, CRP of 5.5 mg/dL, and gram-positive bacilli. CSF was turbid with hypoglycorrhachia (1.9), and d-dimers and line-related thrombosis (femoral and jugular) treated with low


molecular weight heparin. MRI at day 4 demonstrated focal areas of restricted diffusion in the right thalamus and areas of the cortical surface of the occipital and frontal lobes (Fig 2D). Repeated MRI on day 10 suggested possible cortical vein thrombosis, but the patient was already anticoagulated. Repeated MRI at 6 weeks demonstrated resolution of the previous abnormalities with patent venous sinuses and no hydrocephalus. At 3 months, the child seems normal. Discussion We describe recognizable acute cerebral infarction patterns in a case series of neonatal GBS meningitis. Adverse neurologic outcomes from neonatal meningitis are well described. Specific neurologic morbidities are easily defined, with common examples being cerebral palsy, language and cognitive impairments, deafness, vision disorders, epilepsy, and hydrocephalus [1,2]. What is less clear, however, are the mechanisms by which damage occurs within the nervous system to create such deficits. Only by better understanding the mechanisms of injury can preventative strategies be designed to improve outcomes. Although many infants and children with meningitis likely experience stroke, only a small fraction of AIS in the pediatric population is related to meningitis. A large international study of 248 cases of neonatal AIS identified bacterial meningitis in only 3% of cases (IPSS, unpublished data). In the Canadian Pediatric Ischemic Stroke Registry study of nearly 1,000 children with AIS, meningitis was associated with 50 cases (5.3%) [5]. Cerebral infarction is also recognized as a complication of bacterial meningitis later in childhood, although recognition of distinct infarction patterns or mechanisms also remains poorly defined. A case series of 49 children with bacterial meningitis imaged with computed tomography suggested ‘‘infarction’’ in 27% [11]. Defining vascular mechanisms and differentiating such lesions from focal cerebritis or other forms of injury will require further studies of pathogen-specific meningitis with careful examination of acute cerebral and vascular images. Our results support an AIS mechanism in the deep infraction pattern observed in most patients. The major cerebral arteries course through the subarachnoid space, giving rise to perforating lenticulostriate and thalamostriate small arteries known to supply specific structures, namely the deep gray matter of the thalamus and basal ganglia as well as deep white matter. These small vessels therefore originate and course through the actively infected and inflamed meninges to reach the brain parenchyma. Although pathologic studies have been lacking, a resulting localized inflammatory vasculopathy with resultant thrombosis leading to occlusion seems likely. The sensitivity of acute diffusion imaging not only confirms such infarction, but is also able to demonstrate a punctuate pattern of injury resulting from occlusion of individual perforating arteries (Fig 1D). The common involvement of these deeper perfo-

rator arteries may relate to a greater extent of bacterial disease and inflammation in the basal meninges, although our study could not confirm this. The pathophysiology of the more superficial pattern of cortical infarction is not as easily understood, potentially relating to a similar process in small pial arteries, superficial cortical veins with venous infarction, or even direct extension of bacterial infection (cerebritis). Each of these suggestions is only speculation, and pathologic studies linked to detailed imaging will be required to fully understand both patterns of injury. Our sample size is not adequate to try and correlate patterns of stroke with outcome, although a connection between bilateral deep infarction and motor delays seems probable, while cortical lesions might be expected to correlate more with epilepsy. Such a presumed AIS mechanism for deep infarction suggests the possibility of treatment. A regional inflammatory arteriopathy might promote platelet aggregation, regional thrombosis, and subsequent arterial occlusion, although this mechanism has not been proven. This suggests an agent like ASA might be beneficial, although additional thrombotic mechanisms more amenable to anticoagulation therapies cannot be excluded. Three consensus-based guidelines for the management of stroke in children have been published. These include UK guidelines from 2004 [12], the Chest guidelines updated in 2008 [13], and the American Heart Association recommendations published in 2008 [14]. These guidelines agree that ASA in pediatric stroke seems safe, but they differ in numerous management recommendations, with some not addressing neonatal stroke at all and none addressing neonatal meningitis-associated stroke specifically [15]. A recent case series of 22 children with bacterial meningitis and stroke treated with ASA and/or anticoagulation documented no hemorrhagic complications (personal communication, Rand Askalan, MD, PhD, Hospital for Sick Children, Toronto, Canada). Although a randomized clinical trial is likely not feasible, larger studies of antithrombotic use in neonatal meningitis are required. A selection bias may have existed for our population whereby newborns revealing neurologic dysfunction may have been selectively imaged. More routine use of MRI in neonatal meningitis may help complete the full spectrum of focal infarction in neonatal GBS meningitis. Noninvasive imaging of the cerebral arteries could potentially help identify children at risk of meningitis-related infarction. As discussed above, much of the AIS pattern observed seemed to be limited to small perforating arteries of the basal ganglia and thalamus, vessels not typically visible on magnetic resonance angiography. However, dramatic examples of GBS meningitis-related disease of larger arteries, including the internal carotid, basilar, and middle cerebral arteries, have been reported [16]. Cerebrovascular imaging in neonates can be complicated and was not always of the highest quality in our case series. Omission of cerebrovascular imaging may also be a common problem in perinatal stroke in general [17]. We would suggest that magnetic resonance angiography should be included

Hernandez et al: Neonatal Stroke in GBS Meningitis 287 Downloaded from ClinicalKey.com at University of Chile Catholic ALERTA August 30, 2016. For personal use only. No other uses without permission. Copyright ©2016. Elsevier Inc. All rights reserved.

in all children with neonatal GBS meningitis undergoing MRI, although absence of large artery involvement obviously does not remove the risk of vascular disease and stroke. Improved technologies combined with more routine use may better define the role of vasculopathy in neonatal meningitis. The long-term effects of GBS meningitis on cerebrovascular health and stroke risk are also not determined. Recent studies from childhood stroke where arteriopathy that may often be parainfectious and/or inflammatory in nature suggests that the presence of arteriopathy may be the single greatest predictor of stroke recurrence [18,19]. Furthermore, a recent case series of adults with good early recovery from streptococcal meningitis subsequently developed acute and severe infarctions of the same deep penetrating arterial territories described in our case series [20]. Collectively, these observations suggest a potential long-term risk of stroke recurrence after neonatal GBS meningitis. Potential preventative strategies include long-term antiplatelet therapy, which is supported by consensus guidelines in many other forms of childhood stroke, as well as attention to other long-term modifiable arterial health factors, such as diet, dyslipidemia, and hypertension. Our study also suggested limitations in prenatal GBS screening in 2 countries, although this was not a primary objective of the study. A proportion of mothers classified as receiving ‘‘routine prenatal care’’ were not tested, including one with a history of GBS colonization in a previous pregnancy. Another 2 of 3 who screened as negative still had children with neonatal GBS sepsis and meningitis. Screening of GBS during pregnancy at 35 to 37 weeks’ gestation and subsequent antibiotic therapy during labor prevents early-onset GBS disease, likely reducing the incidence by more than 50% [1,12]. However, ideal prenatal GBS screening practices in terms of both disease prevention and cost-effectiveness have been questioned [21]. Few other causes of neonatal stroke are preventable, highlighting the need for optimal GBS meningitis prevention. Conclusion GBS meningitis produces recognizable patterns of acute brain infarction that suggest specific underlying mechanisms of brain injury. Attention to such details is required to better define stroke pathogenesis so that prevention strategies can evolve to improve outcomes. References [1] Chang CJ, Chang WN, Huang LT, et al. Neonatal bacterial meningitis in southern Taiwan. Pediatr Neurol 2003;29:288-94.

[2] Heath PT, Nik Yusoff NK, Baker CJ. Neonatal meningitis. Arch Dis Child Fetal Neonatal Ed 2003;88:F173-8. [3] [Anonymous], Trends in perinatal group B streptococcal disease—United States, 2000-2006. MMWR Morb Mortal Wkly Rep 2009;58:109-12. [4] Takeoka M, Takahashi T. Infectious and inflammatory disorders of the circulatory system and stroke in childhood. Curr Opin Neurol 2002;15:159-64. [5] deVeber G, Kirton A for the Canadian Pediatric Ischemic Stroke Study Group. Arterial ischemic stroke in Canadian children. Ann Neurol 2006;60(S10):S115. [6] deVeber G, Andrew M; Canadian Pediatric Ischemic Stroke Study Group. Cerebral sinovenous thrombosis in children. N Engl J Med 2001;345:417-23. [7] Chang CJ, Chang WN, Huang LT, et al. Cerebral infarction in perinatal and childhood bacterial meningitis. QJM 2003;96:755-62. [8] Fitzgerald KC, Golomb MR. Neonatal arterial ischemic stroke and sinovenous thrombosis associated with meningitis. J Child Neurol 2007;22:818-22. [9] van de Beek D, de Gans J, McIntyre P, Prasad K. Corticosteroids for acute bacterial meningitis. Cochrane Database Syst Rev 2007;(1): CD004405. [10] International Pediatric Stroke Study. Available at: https:// app3.ccb.sickkids.ca/cstrokestudy. Accessed July 1, 2010. [11] Snyder RD, Stovring J, Cushing AH, Davis LE, Hardy TL. Cerebral infarction in childhood bacterial meningitis. J Neurol Neurosurg Psychiatry 1981;44:581-5. [12] Paediatric Stroke Working Group. Stroke in childhood: Clinical guidelines for diagnosis, management and rehabilitation. London: Royal College of Physicians, 2004. [13] Monagle P, Chalmers E, Chan A, et al. Antithrombotic therapy in neonates and children: American College of Chest Physicians evidence-based clinical practice guidelines (8th edition). Chest 2008; 133(6 Suppl):887S-968. [14] Roach ES, Golomb MR, Adams R, et al. Management of stroke in infants and children: A scientific statement from a Special Writing Group of the American Heart Association Stroke Council and the Council on Cardiovascular Disease in the Young. Stroke 2008;39: 2644-91. [15] deVeber G, Kirkham F. Guidelines for the treatment and prevention of stroke in children. Lancet Neurol 2008;7:983-5. [16] Iijima S, Shirai M, Ohzeki T. Severe, widespread vasculopathy in late-onset group B streptococcal meningitis. Pediatr Int 2007;49: 1000-3. [17] Kirton A, deVeber G. Advances in perinatal ischemic stroke. Pediatr Neurol 2009;40:205-14. [18] 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. [19] 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;119:495-501. [20] Schut ES, Brouwer MC, de Gans J, Florquin S, Troost D. van de BD. Delayed cerebral thrombosis after initial good recovery from pneumococcal meningitis. Neurology 2009;73:1988-95. [21] Colbourn TE, Asseburg C, Bojke L, et al. Preventive strategies for group B streptococcal and other bacterial infections in early infancy: Cost effectiveness and value of information analyses. BMJ 2007;335 (7621):655.
