Conventional cerebral angiography in children with ischemic stroke

Conventional Cerebral Angiography in Children With Ischemic Stroke Vijeya Ganesan, MB, ChB*, Lloyd Savvy, MBBS†, W. Kling Chong, MBBS, MD†, and Fenella J. Kirkham, MB, BChir* A retrospective review of conventional cerebral angiograms of 46 children with ischemic stroke was undertaken. Comparison was made with findings on magnetic resonance imaging and magnetic resonance angiography where available. Thirty-six children (78%) underwent magnetic resonance angiography in addition to conventional cerebral angiography. Seven patients had normal cerebral angiograms. Magnetic resonance angiography was diagnostic in 25 of 28 patients with large vessel occlusion, stenosis, or moyamoya syndrome. Conventional angiography was abnormal in four of nine patients with a normal magnetic resonance angiography. All patients with normal conventional angiograms also had normal magnetic resonance angiograms. Conventional angiography, either diagnostic or yielding further information, altered management in five patients with arterial dissection, one patient with large vessel occlusion, one patient with large vessel stenosis, and four patients with arteritis. On the basis of this experience, a clinical algorithm for the use of conventional cerebral angiography in the investigation of ischemic stroke in children is proposed. © 1999 by Elsevier Science Inc. All rights reserved. Ganesan V, Savvy L, Chong WK, Kirkham FJ. Conventional cerebral angiography in children with ischemic stroke. Pediatr Neurol 1999;20:38-42.

Introduction The role of conventional cerebral angiography in the investigation of the child with ischemic stroke is uncertain at present. With the increasing availability of magnetic resonance angiography (MRA), it has not been clear whether this alone is sufficient or whether there is a risk of missing significant pathologic findings requiring alterna-

From the *Neurosciences Unit; Institute of Child Health; and † Department of Radiology; Great Ormond Street Hospital for Children NHS Trust; London, United Kingdom.

38

PEDIATRIC NEUROLOGY

Vol. 20 No. 1

tive management strategies. The aim of this study was to investigate whether there is any diagnostic or therapeutic advantage to undertaking conventional cerebral angiography if MRA is available. On the basis of this experience, the authors propose a clinical algorithm for the use of conventional cerebral angiography in the investigation of children presenting with ischemic stroke. Material and Methods Children presenting with ischemic stroke to the Great Ormond Street Hospital for Children between 1985 and 1996 were eligible for inclusion if a conventional cerebral angiogram had been performed during the acute admission. Stroke was defined as an acute focal neurologic deficit of greater than 24 hours duration, with evidence of cerebral infarction on cranial computed tomography or magnetic resonance imaging (MRI). Patients with stroke in the neonatal period (less than 28 days of age) and hemorrhagic stroke were excluded. Because of the authors’ recognized interest in pediatric cerebrovascular disease, the study population included patients with ischemic stroke and moyamoya syndrome who were referred from other regions. All angiograms were reviewed by two neuroradiologists who were unaware of the clinical and other radiologic data. The site of any abnormality was classified as affecting large (proximal to the first division of a large intracranial vessel), medium (beyond the first division but proximal to the third division of a major artery), or small (the smallest arterioles seen before the capillary phase) vessels. Where possible, these findings were compared with the results of MRA studies carried out during the same admission. MRAs were acquired using a time-of-flight technique centered on the circle of Willis. A short-echo time technique (less than 8 ms) and high spatial resolutions (256 or 512 matrix in a field of view of between 18 and 22 cm) were used. In most cases, MRA was performed at presentation, before the conventional angiogram, as part of the MRI protocol for children presenting with acute focal neurologic deficits in the authors’ unit.

Results Of a total of 128 children with ischemic stroke seen between 1985 and 1996, 69 (54%) underwent conven-

Communications should be addressed to: Dr. Ganesan; Newcomen Centre; Guys Hospital; Saint Thomas Street; London SE1 9RT, UK. Received January 5, 1998; accepted July 8, 1998.

© 1999 by Elsevier Science Inc. All rights reserved. PII S0887-8994(98)00112-X ● 0887-8994/99/$20.00

tional cerebral angiography. Patient selection for cerebral angiography and subsequent management decisions were entirely at the discretion of the clinician in charge, without any predefined guidelines. MRA became readily available in the authors’ unit in 1994; however, the proportions of children investigated with conventional angiography before and after 1994 were similar (51% and 56%). Of the 69 studies performed during this period, 21 studies (19 of which were known to be abnormal) could not be traced; two children who presented with acute hemiparesis without evidence of cerebral infarction were excluded from this study. Conventional cerebral angiograms from 46 patients fulfilling the definition of acute stroke were available for review. These patients’ ages ranged from 9 months to 17 years (median age was 6 years, 7 months); 26 were male. The clinical details and results of radiologic investigations are summarized in Table 1. Risk factors for ischemic stroke were identified in 23 children (50%). A possible complication of angiography was noted in one case (Patient 12), a child with middle cerebral artery stenosis who developed occlusion of this vessel and whose infarct extended within hours of the procedure. Seven patients had normal cerebral angiograms. Abnormalities in the remaining 39 patients predominantly affected the anterior cerebral circulation. Large vessel abnormalities were identified in 35 patients; five patients had abnormalities of the small vessels, which were confined to this site in four patients (Patients 34, 35, 37, and 38). The radiologic diagnoses in the patients with abnormal studies were large vessel occlusion in nine, large vessel stenosis in nine, moyamoya syndrome in 10, cerebral vasculitis in three, diffuse small vessel disease in two, focal arteritis in one, and arterial dissection in five. Thirty-six patients (78%) underwent MRA in addition to conventional angiography. Of the 12 children with an angiographic diagnosis of large vessel stenosis or occlusion, 11 had a similar diagnosis made on MRA (Fig 1). Patient 9 who had evidence of occlusion of the right superior cerebellar artery on angiography had a normal MRA. Eight of the 10 children with moyamoya syndrome underwent MRA, which was diagnostic of moyamoya syndrome in all five children with bilateral involvement and in one of the three children with unilateral involvement. Although large vessel abnormalities were identified on MRA in the remaining two patients, the collateral vessels were not apparent. MRA was considered abnormal but not diagnostic in four of five patients with arterial dissection. The remaining child, Patient 33, who had vertebral artery dissection, had a normal MRA. Of the five patients with abnormalities of the small vessels, one child (Patient 36) also had large vessel abnormalities that were identified on MRA. However, the identification of small vessel disease implied more diffuse disease and altered clinical management. Patient 38 had evidence of turbulent flow in both proximal middle cerebral arteries on MRA but had evidence of diffuse small

vessel disease, with normal middle cerebral arteries, on conventional angiography. Of nine children with normal MRA findings, four (Patients 9, 33, 35, and 37) had abnormalities detected on conventional angiography. None of the patients with normal conventional arteriograms had an abnormal MRA. Conventional angiography identified abnormalities not apparent on the MRA in 13 patients; in 11 of these, the additional findings on the conventional angiogram altered subsequent clinical management. In five patients in whom the final diagnosis was arterial dissection, anticoagulant therapy was initiated as a result of the angiographic findings. Two children (Patients 37 and 38) had evidence of diffuse small vessel disease and were treated with aspirin. Patient 36, with a final diagnosis of isolated cerebral angiitis, was treated with steroids and cyclophosphamide. Patient 34, who was known to have systemic vasculitis, was immunosuppressed with steroids and azathioprine after a stroke on the basis of angiographic appearances of cerebral vasculitis. Patient 18 had irregularity in the basilar tip, suggesting previous occlusion and recanalization, whereas the MRA demonstrated only a small basilar artery. He was treated with anticoagulants for 2 years on the basis of the angiographic findings. Patient 9 was treated with dipyridamole and aspirin on the basis of the angiographic findings of large vessel abnormalities.

Discussion This study suggests that conventional cerebral angiography has a continuing role in the identification of potentially treatable cerebrovascular abnormalities in children with ischemic stroke. These include moyamoya syndrome [1,2], arterial dissection [3], and cerebral vasculitis [4]. MRA is a sensitive imaging modality for moyamoya syndrome, especially for the detection of large vessel stenosis and occlusion. However, collateral circulation is poorly visualized and the degree of large vessel stenosis tends to be overestimated [5]. Conventional cerebral angiography remains mandatory for preoperative evaluation. MRA may be of particular value in screening patients at risk, such as family members [5-7], for serial monitoring in patients who are known to be affected, or after surgery [8]. Although the diagnosis may be considered in patients with bilateral cerebrovascular disease on MRA or in those with suggestive clinical features, such as recurrent ischemic events or unexplained cognitive decline, differentiation from unilateral isolated large vessel disease may be difficult without conventional angiography. This differentiation is important because surgical revascularization may be a therapeutic option for patients with collateral vessels. MRI and MRA are sensitive diagnostic investigations for arterial dissection, especially in combination with duplex ultrasound of the carotid vessels in the neck [3,9]. The diagnostic sensitivity of MRI can be further improved

Ganesan et al: Angiography in Childhood Stroke 39

Table 1.

Summary of risk factors, findings on cerebral angiography, MRA, and outcome

Patient No. 1 2 3 4 5 6 7 8 9* 10

Risk Factor

Angiographic Abnormality

11 12 13 14 15

Aortic coarctation, hypertension Nonspecific febrile illness Hemolytic-uremic syndrome Preceding chickenpox None None None None None High titer of anti-Mycoplasma pneumoniae IgM FVL heterozygote FVL heterozygote None None None

16*

None

17 18* 19 20 21 22 23 24 25 26 27 28 29* 30* 31* 32* 33* 34*

None (subsequently had two additional strokes) None Ventricular septal defect Facial nevus Preceding chickenpox Type 1 protein S deficiency None None None Cranial irradiation None None Craniocervical trauma Cranial trauma Cranial trauma Cranial trauma None Systemic vasculitis

35*

Polyarteritis nodosa, hypertension

36*

None (this study performed after second stroke) Chemotherapy for abdominal lymphoma Severe hypertension

37* 38* 39 40 41 42 43 44 45 46

None Homozygous sickle cell disease Previous meningitis Previous surgery for craniopharyngioma, APCr, familial hypercholesterolaemia None None None None

L ICA occlusion Partial occlusion of R MCA by embolus L MCA occlusion R MCA occlusion L MCA occlusion L MCA occlusion Basilar tip occlusion, filling defect in L VA Occlusion of L PICA Occlusion of R SCA; irregularity of L PCA Stenosis R terminal ICA/MCA

MRA not performed R MCA occlusion L MCA occlusion R MCA occlusion MRA not performed Reduced flow L ICA MRA not performed MRA not performed Normal MRA not performed

Stenosis of R terminal ICA Proximal L MCA stenosis Terminal L ICA and MCA stenosis R terminal ICA/MCA stenosis Terminal ICA and MCA stenosis

Reduced flow R ICA, MCA, ACA L MCA stenosis L MCA stenosis MRA not performed Absent flow proximal R MCA and ACA Abnormal flow R MCA and PCA

R MCA stenosis; posterior circulation normal R MCA stenosis

40

PEDIATRIC NEUROLOGY

MCA MRA PCA PICA SCA VA

Vol. 20 No. 1

5 5 5 5 5 5

Proximal R MCA occlusion

Irregular basilar tip Bilateral moyamoya syndrome Bilateral moyamoya syndrome Bilateral moyamoya syndrome Bilateral moyamoya syndrome Bilateral moyamoya syndrome Bilateral moyamoya syndrome Bilateral moyamoya syndrome L sided moyamoya syndrome L sided moyamoya syndrome L sided moyamoya syndrome L ICA dissection R ICA dissection R VA dissection R VA dissection L VA dissection Irregularity/focal stenosis of multiple vessels compatible with cerebral vasculitis Attenuated and tortuous small vessels compatible with arteritis Stenosis and tortuosity of multiple vessels suggestive of cerebral vasculitis Irregularity of small vessels Attenuation of all vessels compatible with chronic hypertension Focal arteritis of R MCA Normal four-vessel study Normal four-vessel study Normal four-vessel study

Small basilar artery Bilateral moyamoya syndrome MRA not performed Bilateral moyamoya syndrome Bilateral moyamoya syndrome Bilateral moyamoya syndrome Bilateral moyamoya syndrome Bilateral moyamoya syndrome Reduced flow L MCA L MCA stenosis L sided moyamoya syndrome Reduced flow L ICA, MCA, ACA Reduced flow R ICA, MCA Occlusion of distal R VA Reduced flow R VA Normal MRA not performed

Normal Normal Normal Normal

Normal MRA not performed Normal MRA not performed

four-vessel study four-vessel study R carotid study L carotid and vertebral study

* Patients in whom the angiographic findings differed from those on MRA. Abbreviations: ACA 5 Anterior cerebral artery APCr 5 Activated protein C resistance FVL 5 Factor V Leiden ICA 5 Internal carotid artery

Findings on MRA

Middle cerebral artery Magnetic resonance angiography Posterior cerebral artery Posterior inferior cerebellar artery Superior cerebellar artery Vertebral artery

Normal L MCA occlusion Normal Abnormal flow proximal MCA bilaterally Small R ICA, MCA Normal Normal Normal

Figure 2. Contrast cerebral angiogram (left internal carotid injection) of Patient 17 demonstrating irregular stenosis of the left middle cerebral artery (arrow) suggestive of cerebral vasculitis.

Figure 1. Magnetic resonance angiogram revealing occlusion of the proximal left middle cerebral artery (arrow).

by taking fine cuts through the skull base and using fat saturation techniques [9]. However, such techniques are not routinely included in MR examinations, which may partly account for the relative lack of sensitivity of MRI and MRA for arterial dissection in the authors’ patients. A history of trauma, as well as other suggestive clinical features, such as facial pain, may be helpful in targeting these techniques. However, definite exclusion of the diagnosis may not be possible with conventional angiography. Although large vessel involvement in diffuse cerebral

vasculitis may be apparent on MRA, the distinction from focal large vessel stenosis or occlusion may be impossible without imaging smaller vessels (Fig 2). Although the angiographic appearances of cerebral vasculitis may be mimicked by other processes [10], the appearance of multiple intracranial vascular stenoses or dilatations is suggestive of the diagnosis. If patients are imaged early in the course of their illness, it also may not be possible to make the distinction from localized vessel disease. This is well illustrated by the cases of Patients 17 and 36 who had recurrent strokes and severe residual disability. Although initial angiography was suggestive of focal large vessel disease, subsequent studies in both patients have revealed

Figure 3. Clinical algorithm for use of conventional angiography in the investigation of acute ischemic stroke in children. MRA 5 magnetic resonance angiography; MRV 5 magnetic resonance venography; TCD 5 transcranial Doppler ultrasound.

Ganesan et al: Angiography in Childhood Stroke 41

multiple vessel stenosis and irregularity, compatible with cerebral vasculitis. The diagnosis should, therefore, be considered in patients with recurrent or progressive symptoms. Noninvasive detection of cerebrovascular disease in children has been previously investigated in children with sickle cell disease. In this population, MRA and transcranial Doppler ultrasound individually have demonstrated a high sensitivity and specificity compared with conventional angiography in the detection of large vessel stenosis and occlusion [11,12]. However, the resolution of MRA only permits examination of vessels with a diameter of 1 mm [13], and detection of abnormalities in the posterior circulation is relatively poor [11]. Transcranial Doppler ultrasound is not widely available to pediatricians and cannot be used to exclude pathology in small vessels. These issues are less important in children with sickle cell disease because the detection of large vessel disease is of primary importance in determining management. However, in other children presenting with ischemic stroke, the spectrum of cerebrovascular pathologic findings and potential treatments is more diverse, and a different approach is required. In the light of the experience reviewed in this study, a clinical algorithm for guiding the use of cerebral angiography in the investigation of children with ischemic stroke is proposed in Figure 3. Conventional cerebral angiography remains the gold standard for cerebrovascular imaging. As more therapies become available for ischemic stroke, conventional angiography may assume an even greater role in guiding their use. Drs. Ganesan and Kirkham were supported by the Wellcome Trust.

42

PEDIATRIC NEUROLOGY

Vol. 20 No. 1

References [1] George BD, Neville BGR, Lumley JSP. Transcranial revascularisation in childhood and adolescence. Dev Med Child Neurol 1993; 35:675-82. [2] Olds MV, Griebel RW, Hoffman HJ. The surgical treatment of childhood moyamoya disease. J Neurosurg 1987;60:675-80. [3] Sturzenegger M. Spontaneous internal carotid artery dissection: Early diagnosis and management in 44 patients. J Neurol 1994;242: 231-8. [4] Barron TF, Ostrov BE, Zimmerman RA, Packer RJ. Isolated angiitis of CNS: Treatment with pulse cyclophosphamide. Pediatr Neurol 1993;9:73-5. [5] Battistella PA, Carollo C, Pellegrino PA, Soriani S, Scarpa P. Magnetic resonance angiography in moyamoya disease. Childs Nerv Syst 1995;11:329-34. [6] Ganesan V, Isaacs E, Kirkham FJ. Variable presentation of cerebrovascular disease in monovular twins. Dev Med Child Neurol 1997;39:628-31. [7] Houkin K, Tanaka N, Takahashi A, Kamiyama H, Abe H, Kajii N. Familial ocurrence of moyamoya disease: Magnetic resonance angiography as a screening test for high-risk subjects. Childs Nerv Syst 1994;10:421-5. [8] Kikuchi M, Hayakawa H, Takahashi I, et al. Moyamoya disease in three siblings—follow-up study with magnetic resonance angiography (MRA). Neuropediatrics 1995;26:33-6. [9] Ozdoba C, Sturzenegger M, Schroth G. Internal carotid artery dissection: MR imaging features and clinical radiological correlation. Radiology 1996;199:191-8. [10] Moore PM. Diagnosis and management of isolated angiitis of the central nervous system. Neurology 1989;39:167-73. [11] Kandeel AY, Zimmerman RA, Ohene-Frempong K. Comparison of magnetic resonance angiography and conventional angiography in sickle cell disease: Clinical significance and reliability. Neuroradiology 1996;38:409-16. [12] Adams RJ, Nichols FT, Figueroa R, McKie V, Lott T. Transcranial Doppler correlation with cerebral angiography in sickle cell disease. Stroke 1992;23:1073-7. [13] Sellar RJ. Imaging the blood vessels of the head and neck. J Neurol Neurosurg Psychiatry 1995;59:225-37.