Professional Documents
Culture Documents
©AANS, 2014
Christian von der Brelie, M.D.,1,2 Stefan Kuczaty, M.D., 3 and Marec von Lehe, M.D.1,4
Departments of 1Neurosurgery and 3Epileptology, University of Bonn Medical Center, Bonn; 2Department of
Neurosurgery, Unfallkrankenhaus Berlin; and 4Department of Neurosurgery, University of Bochum, Germany
Object. Sufficient data on surgical treatment and seizure outcome of pediatric patients with different types of
epilepsy, especially drug-resistant epilepsy and associated cerebral cavernous malformations, are scarce. The aim of
this study was to carefully evaluate seizure outcome using the International League Against Epilepsy (ILAE) clas-
sification with regard to the presurgical symptom duration.
Methods. Fifty-one pediatric patients younger than 19 years with cerebral cavernous malformations of all CNS
localizations have been surgically treated at the authors’ institution. Twenty-two patients with seizures or epilepsy
who harbored cortically located supratentorial cerebral cavernous malformations underwent surgical treatment and
were retrospectively analyzed.
Results. More extensive resections were used in 82% of all patients with epilepsy symptoms for longer than 2
years. Eighty-two percent of patients with symptom duration shorter than 2 years underwent circumscribed lesionec-
tomy including the surrounding hemosiderotic rim. The overall rate of mild permanent, unanticipated postoperative
deficits was 4.5%; the rate of anticipated neurological deficits was 9%. The mean follow-up was longer than 117
months in all groups. Seizure outcome was excellent in the group with symptom duration shorter than 2 years (100%
ILAE Class 1). Seizure outcome was significantly worse in the group with longer symptom duration (p = 0.02). Seven
patients were seizure free after surgery. Seizure outcome was stable over the years.
Conclusions. Since seizure outcome is worse with longer seizure duration, early surgery and, if needed, inter-
disciplinary intervention, is recommended. Even in cases of multiple cerebral cavernous malformations and epilepsy,
surgery should be considered.
(http://thejns.org/doi/abs/10.3171/2014.2.PEDS13361)
V
ascular malformations are often accompanied by ing forms of epilepsy or even drug-resistant epilepsy.
epileptic seizures and represent a common neuro- Four percent of drug-resistant epilepsy patients of all ages
logical and neurosurgical problem. Since the ad- are diagnosed with a CCM.2,6,16
vent of MRI, cerebral cavernous malformations (CCMs) The rate of pediatric patients with CCM-associated
have been increasingly recognized as being responsible drug-resistant epilepsy is estimated to be around 30%.1,13
for structural epileptic seizures and epilepsy. Their over- Furthermore, children with longer symptom duration were
all incidence is 0.1%–0.5%.7,23 Around 25% of CCMs are considered to have a worse outcome regarding seizure
diagnosed in pediatric patients.22,24 These lesions make up freedom.13 Without epileptic seizures, the indication for
about 2%–18% of all vascular malformations in the pedi- the surgical therapy of CCM is the prevention of bleeding
atric brain.18 Cerebral cavernous malformations typically and its consequences. With ongoing seizure activity, the
present either with hemorrhage or epileptic seizures.17,20,24 indication changes and seizure control is the focus. There-
Regarding the CCM-associated type of epilepsy, symp- fore, it is advisable to adapt the surgical strategy and the
toms may present as sporadic seizures and longer-stand- extent of the resection.
Data on CCM-associated epilepsy in the pediatric
population are scarce. Table 1 gives an overview of the
Abbreviations used in this paper: AED = antiepileptic drug; CCM available literature. It becomes clear that most of the stud-
= cerebral cavernous malformation; EEG = electroencephalography; ies display results of rather small patient populations. In
ILAE = International League Against Epilepsy. particular, patients with drug-resistant epilepsy are under-
represented. Comparisons between the different studies ILAE classification. Only patients with follow-up longer
are difficult since useful seizure outcome scales, such as than 12 months were included in the study.
Engel’s classification or the International League Against
Epilepsy (ILAE) classification, are usually not reported. Categorization of Localization
Overall, there is a shortage of detailed and useful data Localization of the CCMs was analyzed on the basis
regarding pediatric patients with different types of epi- of radiological images, radiological reports, and surgical
lepsy and long-term epileptological follow-up after sur- records. Localization of symptomatic CCMs was divided
gery.1 The purpose of this paper is to generate data for a into different categories. Temporal localization was sub-
thorough counseling of patients and caregivers. divided into temporolateral neocortical and temporome-
sial archicortical localization. Furthermore, symptomatic
Methods CCMs were considered to be in an eloquent location if the
symptomatic CCM was located in the precentral or post-
Inclusion and Exclusion Criteria central gyrus or in Wernicke’s (angularis region) or Bro-
ca’s area of the presumed dominant hemisphere, as well as
The neurosurgical index database and the epilepsy in the primary visual cortex. All other CCMs were con-
surgery database were screened for procedures per- sidered as being localized in noneloquent brain regions.
formed between 1988 and 2010. A total of 168 patients
with CCM-associated epilepsy were identified. Among Categorization of Duration of Epilepsy
these, 22 patients were younger than 19 years (13%).
Fifty-one pediatric patients with CCMs of all CNS lo- Patients were assigned to the shorter symptom dura-
calizations underwent surgery at our institution in this tion group (Group A) if they had symptoms for less than 2
period. Twenty-nine patients harboring CCMs in non- years. If seizure duration was longer than 2 years, patients
epileptogenic areas, for example, basal ganglia, thalamus, were categorized in Group B.
cerebellum, brainstem, or spinal CCMs, were excluded, Presurgical Epileptological Workup
except for patients with multiple CCMs. Six patients with
multiple CCMs and epilepsy in whom all epileptogenic Patients with drug-resistant epilepsy underwent de-
CCMs were located cortically or subcortically were in- tailed epileptogical evaluation assessing medical his-
cluded. We included only cases with histopathological tory with special focus on seizure semiology. Thorough
confirmation of CCM. All biopsy specimens underwent epileptological workup also included high-quality MRI.25
neuropathological analysis, and only cases with concor- Long-term video-electroencephalography (EEG) moni-
dant neuropathological, neuroradiological, and intraoper- toring and in some cases invasive monitoring with strip or
ative findings, particularly unequivocal exclusion of arte- grid electrodes was performed. Standards and techniques
riovenous malformations, were considered for this study. of presurgical evaluation have been published.5
Analyzed data included patient age, sex, age at seizure
onset and type of seizures, duration of symptoms, surgi- Surgical Technique and Extent of Resection
cal approach, extent of resection, and surgical morbidity Extent of resection was evaluated by analysis of surgi-
and mortality, as well as seizure outcome according to cal records and, if available, postoperative MRI findings.
Resection types were classified into lesionectomy (the TABLE 2: Demographic data of 22 patients with CCM-associated
hemosiderotic rim was completely or partially removed) epilepsy*
and extended lesionectomy. Extended lesionectomy was
usually performed as tailored resection (lesionectomy Parameter Total Group B Group A p Value
including rim and adjacent neocortex). In selected cases, no. of pts 22 11 11
lesionectomy including temporomesial resection (cases
M/F 14/8 7/4 7/4
in which the lesion and the hemosiderotic rim as well
as amygdala and hippocampus/parahippocampal gyrus mean age at seizure 11.3 10.6 12 NS
were removed) was performed. onset (yrs)
mean age at op (yrs) 13.9 15.3 12.5 NS
Follow-Up and Seizure Outcome mean duration of 30.4 57.8 3.6 0.0005
Follow-up data were obtained by regular outpatient symptoms (mos)
visits, routinely performed at 3, 6, and 12 months post- no. w/ multiple CCMs 6/22 (27.3) 3/6 (50) 3/6 (50)
operatively. Also, all patients and/or caregivers and/or
their general practitioners/neurologists were contacted * Values are number of patients (%) unless stated otherwise. NS = not
by phone with standardized interviews and were asked to significant.
provide information on further seizure history and anti-
epileptic medication, subsequent neurosurgical or neuro- less than 2 years (Group A). The remaining 11 patients
logical hospital admissions, or any other subsequent neu- had symptoms for more than 2 years (Group B). Among
rosurgical procedure. Seizure outcome was determined the 11 patients in Group B, 8 patients were diagnosed as
using ILAE classification.26 having drug-resistant epilepsy.
The mean age at seizure onset in Group A was 12
Statistical Analysis years, and the mean age at surgery was 12.5 years. The
To analyze descriptive population statistics and po- mean duration of seizures in Group B was 57.8 months.
tential associations of factors with postoperative seizure The mean age at seizure onset was 10.6 years, and the
outcome we used SPSS (version 20, IBM). Statistical sig- mean age at surgery was 15.3 years in Group B (see also
nificance was reached at p < 0.05. Table 2 for other details of the groups).
Preoperative Workup
Results
Detailed epileptological evaluation was done in 9
Patient and Demographic Data (41%) of 22 patients. Detailed workup was done in 8 pa-
We identified 22 patients younger than 19 years; 14 tients in Group B (72.7%) and in 1 patient in Group A
(9%). In 7 (87.5%) of the 8 patients in Group B, electro-
(64%) of these patients were male. The mean age at onset
physiological diagnostics showed concordance between
of seizure activity was 11.3 ± 4.6 years ([± SD] range 2–18
localization of the CCM and ictal focus; the remaining
years), and the mean age at surgery was 13.9 ± 4.3 years
patient underwent invasive diagnostic testing (see below).
(range 3–18 years). The mean duration of seizures was Detailed epileptological workup was done in 4 (67%)
30.4 ± 39.2 months (range 0.03–144 months). In 4 patients of 6 patients with multiple CCMs to identify the symp-
seizures occurred as a symptom of recent significant in- tomatic CCM. Electrophysiological data were conclusive
tracerebral hemorrhage from CCM (see Tables 2 and 3 in 3 of those patients (75%); the remaining patient under-
for details). went invasive monitoring.
Localization of Symptomatic CCMs Invasive Monitoring
Sixteen patients had single supratentorial CCMs. Overall, invasive EEG recording was performed in
Nine (56.2%) of these 16 CCMs were located in the left 2 patients. Findings from noninvasive epileptological
hemisphere. In 19% the symptomatic CCM was located workup were inconclusive in these patients, and therefore
in eloquent brain regions; 6 CCMs (38%) were temporal they underwent subdural grid electrode and depth elec-
(Fig. 1). Temporal localization was statistically not pre- trode implantation. One patient in Group B (Case 4, Table
disposing for longstanding symptom duration (p > 0.05). 3) had drug-resistant epilepsy potentially due to a tem-
Six patients with multiple CCMs harbored 27 CCMs poromesial CCM located in the parahippocampal gyrus.
overall (mean 4.5, range 3–6) (Fig. 1). In these 6 patients, Noninvasive surface EEG was not sufficient to determine
a total of 8 symptomatic CCMs were removed. Regarding if the epileptogenic area included the temporomesial
the localization of the removed lesions, 37.5% of these structures. This patient underwent stereotactic implanta-
were located in the temporal lobe and 37.5% in eloquent tion of bilateral hippocampal depth electrodes and 2 left-
brain regions. sided subtemporal 4-contact strip electrodes.
The other patient (Case 13, Table 3), who was in
Clinical Data and Duration of Symptoms
Group A, harbored 3 superficial multiple CCMs in the
The patients were divided into 2 groups according to left parasagittal frontal cortex, in the left occipitobasal
their preoperative duration of seizures, which they expe- cortex, and in the left frontolateral region. Surface EEG
rienced presurgically. Eleven patients had symptoms for recorded 2 seizures for which no focus could be delineat-
* AHE = amygdalohippocampectomy; ECoG = electrocorticography; ext = extended; incl = including; MST = multiple subpial transection.
ed. Ictal SPECT evaluation was also not conclusive. This toms of epilepsy, the more extensive the resection (p =
patient was implanted with a total of 8 strip electrodes 0.008; Fig. 2).
covering all CCMs.
Neither of these patients developed a postoperative Surgical Morbidity and Mortality
complication due to the implantation procedure, and both In 8 of 11 patients in Group B, the postoperative
were seizure free at last available outcome. course was uneventful. One patient developed quadran-
tanopia, as anticipated (a calculated deficit that was dis-
Surgical Management
cussed and accepted prior to surgery). One patient devel-
In Group B, 9 (81.8%) of 11 patients underwent ex- oped transient mild anomia and mild amnesic aphasia.
tended lesionectomy. In these cases the lesion itself and The patient recovered well with speech therapy and was
the hemosiderotic rim as well as the adjacent cortex were not impaired at the last available follow-up. Another pa-
removed. Two patients in Group B (18.2%) underwent le- tient developed a permanent sensory deficit (calculated
sionectomy. In 1 patient (12.5%), multiple subpial tran- deficit) due to removal of a symptomatic postcentral
sections were added to the extended resection. In this CCM. In Group A, 1 patient developed a permanent un-
patient (Case 2), the epileptological workup revealed an anticipated mild motor aphasia due to pure lesionectomy
overlapping of the epileptogenic zone and the Wernicke of a symptomatic CCM in the proximity of Broca’s area.
area (Table 3). In Group A the surgical strategy was dif- The overall rate of mild permanent, unanticipated postop-
ferent. Nine patients (81.8%) underwent lesionectomy in- erative deficits was 4.5%. The rate of anticipated deficits
cluding complete or partial removal of the hemosiderotic was 9%. There were no systemic or local complications in
rim (Fig. 2). In summary, the longer the clinical symp- the entire study group.
Fig. 1. Localization of CCMs in patients with a single CCM (16 patients) and multiple CCMs (6 patients with 27 CCMs). Frontal
localization was found most often in patients with single supratentorial CCMs and multiple CCMs.
Discussion
The present study analyzed one of the largest single-
center cohorts of pediatric patients with CCM-associated Fig. 2. Lesionectomy (Lx) was performed more often in patients
seizures and epilepsy.1,13,17 One of the most common symp- with a shorter seizure history. Extended lesionectomy (extended Lx)
toms of CCMs are seizures, and there is still uncertainty was done more frequently in patients with more longstanding seizure
regarding if and when to treat patients surgically. Should symptoms.
epilepsy at one centre from 1988-2009. J Neurol Neurosurg temporal lobe epilepsy: pathological correlations. Ann Neu-
Psychiatry 84:54–61, 2013 rol 22:341–347, 1987
3. Buckingham MJ, Crone KR, Ball WS, Berger TS: Manage- 17. Lee JW, Kim DS, Shim KW, Chang JH, Huh SK, Park YG,
ment of cerebral cavernous angiomas in children presenting et al: Management of intracranial cavernous malformation in
with seizures. Childs Nerv Syst 5:347–349, 1989 pediatric patients. Childs Nerv Syst 24:321–327, 2008
4. Clusmann H: Predictors, procedures, and perspective for 18. Lena G, Ternier J, Paz-Paredes A, Scavarda D: [Central ner-
temporal lobe epilepsy surgery. Semin Ultrasound CT MR vous system cavernomas in children.] Neurochirurgie 53:
29:60–70, 2008 223–237, 2007 (Fr)
5. Clusmann H, Kral T, Schramm J: Present practice and per- 19. Loring DW, Meador KJ, Lee GP: Determinants of quality of
spective of evaluation and surgery for temporal lobe epilepsy. life in epilepsy. Epilepsy Behav 5:976–980, 2004
Zentralbl Neurochir 67:165–182, 2006 20. Maraire JN, Awad IA: Intracranial cavernous malformations:
6. Convers P, Bierme T, Ryvlin P, Revol M, Fischer C, Froment lesion behavior and management strategies. Neurosurgery
JC, et al: [Contribution of magnetic resonance imaging in 100 37:591–605, 1995
cases of refractory partial epilepsy with normal CT scans.] 21. Mazza C, Scienza R, Beltramello A, Da Pian R: Cerebral
Rev Neurol (Paris) 146:330–337, 1990 (Fr) cavernous malformations (cavernomas) in the pediatric age-
7. Del Curling O Jr, Kelly DL Jr, Elster AD, Craven TE: An group. Childs Nerv Syst 7:139–146, 1991
analysis of the natural history of cavernous angiomas. J Neu- 22. Mottolese C, Hermier M, Stan H, Jouvet A, Saint-Pierre G,
rosurg 75:702–708, 1991 Froment JC, et al: Central nervous system cavernomas in the
8. Di Rocco C, Iannelli A, Tamburrini G: Surgical management pediatric age group. Neurosurg Rev 24:55–73, 2001
of paediatric cerebral cavernomas. J Neurosurg Sci 41:343– 23. Robinson JR, Awad IA, Little JR: Natural history of the cav-
347, 1997 ernous angioma. J Neurosurg 75:709–714, 1991
9. Dodrill CB: Problems in the assessment of cognitive effects 24. Scott RM, Barnes P, Kupsky W, Adelman LS: Cavernous an-
of antiepileptic drugs. Epilepsia 33 (Suppl 6):S29–S32, 1992 giomas of the central nervous system in children. J Neuro-
10. Fortuna A, Ferrante L, Mastronardi L, Acqui M, d’Addetta R: surg 76:38–46, 1992
Cerebral cavernous angioma in children. Childs Nerv Syst 25. Urbach H: High-field magnetic resonance imaging for epilep-
5:201–207, 1989 sy. Neuroimaging Clin N Am 22:173–189, ix–x, 2012
11. Gallassi R, Morreale A, Lorusso S, Procaccianti G, Lugaresi 26. Wieser HG, Blume WT, Fish D, Goldensohn E, Hufnagel A,
E, Baruzzi A: Cognitive effects of valproate. Epilepsy Res King D, et al: ILAE Commission Report. Proposal for a new
5:160–164, 1990 classification of outcome with respect to epileptic seizures fol-
12. Giulioni M, Acciarri N, Padovani R, Frank F, Galassi E, Gaist lowing epilepsy surgery. Epilepsia 42:282–286, 2001
G: Surgical management of cavernous angiomas in children. 27. Xia C, Zhang R, Mao Y, Zhou L: Pediatric cavernous mal-
Surg Neurol 42:194–199, 1994 formation in the central nervous system: report of 66 cases.
13. Hugelshofer M, Acciarri N, Sure U, Georgiadis D, Baumgart- Pediatr Neurosurg 45:105–113, 2009
ner RW, Bertalanffy H, et al: Effective surgical treatment 28. Yeh HS, Privitera MD: Secondary epileptogenesis in cerebral
of cerebral cavernous malformations: a multicenter study of arteriovenous malformations. Arch Neurol 48:1122–1124,
79 pediatric patients. Clinical article. J Neurosurg Pediatr 1991
8:522–525, 2011
14. Josephson CB, Leach JP, Duncan R, Roberts RC, Counsell
CE, Al-Shahi Salman R: Seizure risk from cavernous or ar-
teriovenous malformations: prospective population-based Manuscript submitted July 11, 2013.
study. Neurology 76:1548–1554, 2011 Accepted February 27, 2014.
15. Kraemer DL, Awad IA: Vascular malformations and epilepsy: Please include this information when citing this paper: pub-
clinical considerations and basic mechanisms. Epilepsia 35 lished online April 4, 2014; DOI: 10.3171/2014.2.PEDS13361.
(Suppl 6):S30–S43, 1994 Address correspondence to: Christian von der Brelie, M.D.,
16. Kuzniecky R, de la Sayette V, Ethier R, Melanson D, Ander- Department of Neurosurgery, Unfallkrankenhaus Berlin, Warener
mann F, Berkovic S, et al: Magnetic resonance imaging in Strasse 7, 12683 Berlin, Germany. email: CvdB@gmx.net.