G

C

E

BEa

b

ARRAA

KENP

1

ielpwludfdi

B

(((

0h

ARTICLE IN PRESS Model

LINEU-3370; No. of Pages 3

Clinical Neurology and Neurosurgery xxx (2013) xxx– xxx

Contents lists available at SciVerse ScienceDirect

Clinical Neurology and Neurosurgery

jou rn al h om epage: www.elsev ier .com/ locate /c l ineuro

pilepsy in children with periventricular leukomalacia

arıs Ekici a,∗, Nur Aydınlı a, Kubilay Aydınb, Mine C alıs kana,mine Eraslana, Meral Özmena

Istanbul University, Istanbul Medical Faculty, Department of Pediatric Neurology, Istanbul, TurkeyDepartment of Radiodiagnostics, Istanbul Medical Faculty, Istanbul, Turkey

a r t i c l e i n f o

rticle history:eceived 2 April 2012eceived in revised form 16 June 2013ccepted 24 June 2013vailable online xxx

eywords:pilepsyeonatal seizureeriventricular leukomalacia

a b s t r a c t

Objective: We aimed to analyze the development of epilepsy in a patient group with periventricularleukomalacia followed at a tertiary pediatric neurology center.Patients and methods: The study included 108 children aged between 2 and 8 years with radiologicallyproven periventricular leukomalacia who had been regularly observed at the Istanbul University, IstanbulFaculty of Medicine, Department of Pediatrics, Pediatric Neurology outpatient clinic between January2006 and December 2011.Results: Neonatal seizures were reported in 22 patients (20.3%), 14 of whom developed epilepsy. A signif-icant correlation was found between neonatal seizures and prematurity and newborn asphyxia (p = 0.013and p = 0.010, respectively). Epilepsy developed in 35 patients (32.4%), history of neonatal seizures andmore severe loss of white matter, periventricular hyperintensity and corpus callosum involvement werefound to be correlated with epilepsy (p = 0.001, p = 0.004, p = 0.016, and p = 0.004, respectively). The mostcommon seizure pattern observed was generalized tonic clonic seizures (n = 13) and complex partial

seizures (n = 11). Those with focal EEG findings had a significantly better neurodevelopmental and cogni-tive level than those with multifocal/generalized EEG findings (p = 0.024). Seizures continued with varyingfrequency in 14 epileptic patients (40%) despite antiepileptic treatment.Conclusion: Almost a third of patients with periventricular leukomalacia develop epilepsy that can beintractable in substantial part. Neonatal seizures and severe MRI findings are important clues that canindicate the development of epilepsy in these patients.

. Introduction

Periventricular leukomalacia (PVL) is caused by ischemic ornflammatory trauma of cerebral white matter adjacent to the lat-ral ventricles. It is more common in premature infants with veryow birthweight but can also occur in full term infants. Along withrematurity, PVL etiology includes maternal chorioamnionitis asell as respiratory distress and sepsis of the newborn [1]. PVL often

eads to cerebral movement disorder and spastic diplegia. Frequentse of magnetic resonance imaging (MRI) has shown that PVL canevelop without concomitant cerebral palsy [2]. It can lead to motor

Please cite this article in press as: Ekici B, et al. Epilepsy in children

http://dx.doi.org/10.1016/j.clineuro.2013.06.014

unction disorders, vision problems, cognitive and behavioral disor-ers and epilepsy [3]. As there have been few studies researching

ncidence of epilepsy, risk factors, seizure types and prognosis in

∗ Corresponding author at: Ortaköy dereboyu cad. Arkeon sitesi, A5 blok D3,es iktas , Istanbul, Turkey. Tel.: +90 505 7396812.

E-mail addresses: ekicibaris@yahoo.com (B. Ekici), nuraydinli@superonline.comN. Aydınlı), dr.aydink@superonline.com (K. Aydın), minecal@istanbul.edu.trM. C alıs kan), emineeraslan@gmail.com (E. Eraslan), mozmen@istanbul.edu.trM. Özmen).

303-8467/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.clineuro.2013.06.014

© 2013 Elsevier B.V. All rights reserved.

PVL patients, we set out to analyze the development of epilepsy ina patient group observed at our tertiary pediatric neurology center.

2. Methods

The study included children aged between 2 and 8 years withradiologically proven periventricular leukomalacia who had beenregularly observed at the Istanbul University, Istanbul Faculty ofMedicine, Department of Pediatrics, Pediatric Neurology outpatientclinic between January 2006 and December 2011. Demographic andpast medical information of patients was obtained from clinicalfiles and detailed interviews with parents. All patients underwentneurological examination and intelligence scoring using Stanford-Binet Intelligence Scales or Bayley Mental, Motor and Social ScalesIII. Neurodevelopment and cognition were considered normal inpatients with Bayley composite cognitive scores or IQ scores ofover 85. Patients with scores of 70–84 were considered slightly

with periventricular leukomalacia. Clin Neurol Neurosurg (2013),

impaired, and those with scores under 70–significantly impaired.Epilepsy was defined as two or more unprovoked seizures,

excluding neonatal convulsions. Diagnosis was made based onpatient records, clinical description and electroencephalogram

ARTICLE IN PRESSG Model

CLINEU-3370; No. of Pages 3

2 B. Ekici et al. / Clinical Neurology and Neurosurgery xxx (2013) xxx– xxx

Table 1Factors associated with epilepsy development.

Epileptic patients with PVL (n) Non-epileptic patients with PVL (n) p

Had neonatal convulsions (n) 14/35 8/73 0.001

Preterm birth (n) 23/35 58/73 0.097Intraventricular hemorrhage (n) 8/35 17/73 0.608Had neonatal sepsis (n) 7/35 25/73 0.129Had mechanical ventilation (n) 21/35 46/73 0.763Normal/slightly impaired/impaired neurodevelopment or cognition/n) 7/13/15 35/23/15 0.010Diplegic/hemiplegic/quadriplegic cerebral palsy (n) 7/11/13 26/20/13 0.069Mild/moderate/severe white matter loss (n) 8/24/3 41/27/5 0.004

Mild/moderate/severe corpus callosum thinning (n) 5/19/11 37/21/15 0.004Trigonal/centrum semiovale/frontal hyperintensities (n) 8/14/13 38/19/16 0.016Had Wallerian degeneration (n) 6/35 4/73 0.580

09 (±2

p

fiopstsiwf

tltiilconapw

yawc

3

2wtdtnnnodNefr2wsc

Mean total MRI score 6.

< 0.05 is considered as significant and indicated in bold.

ndings. Seizure types were classified according to the criteriaf the International League Against Epilepsy defined in 1981. Allatients with epilepsy had interictal EEG records available fortudy. A 10-channel EEG device was used for the recordings, andhe electrodes were used according to the international 10–20ystem. The recordings of younger children were obtained whilen spontaneous or sedated sleep. Electroencephalogram findings

ere categorized as normal, focal epileptiform activity and multi-ocal/generalized epileptiform activity.

All MRI findings were reevaluated by a neuroradiologist blindedo the neurological condition of the participants. Periventricu-ar hyperintensities were graded as follows: grade I, localized torigon; grade II, extending to centrum semiovale; grade III, extend-ng frontally. White matter loss was graded as follows: grade I,rregularity of lateral ventricule borders; grade II, dilatation ofateral ventricule; grade III, dilatation extending cortex. Corpusallosum thinning was graded as follows: grade I, thinning of genur splenium; grade II, thinning of both parts; grade III, global thin-ing of corpus callosum. Presence of Wallerian degeneration waslso noted. Total MRI score was calculated by addition of gradeslus 1 extra point if Wallerian degeneration was present. Patientsith MRI findings suggesting near-term hypoxia were excluded.

The SPSS 11.5 software package was used for biostatistical anal-sis. The level of significance was set at p = 0.05. The study waspproved by the Ethics Committee of the Istanbul University andas carried out in accordance with the Helsinki Declaration prin-

iples.

. Results

Enrolled in the study were 108 PVL patients aged between4 and 96 months. Average patient age was 62.2 ± 21.5 months,hile average follow-up period was 44.2 ± 25.4 months. Neona-

al seizures were reported in 22 patients (20.3%), 14 of whomeveloped epilepsy, while the remaining 8 were seizure free afterhe neonatal period. A significant correlation was found betweeneonatal seizures and prematurity (<37 gestational weeks) andewborn asphyxia (p = 0.013 and p = 0.010, respectively), whileone was found between newborn sepsis, intraventricular hem-rrhage and mechanical ventilation (minimum p = 0.506). Epilepsyeveloped in 35 patients (32.4%) and febrile seizures in 6 (5.5%).one of the patients with epilepsy had any family history ofpilepsy, and only one patient had a close relative with a history ofebrile seizures. Out of 6 patients with febrile seizures, 2 had closeelatives with history of febrile seizures. Out of 35 epilepsy patients,

Please cite this article in press as: Ekici B, et al. Epilepsy in children

http://dx.doi.org/10.1016/j.clineuro.2013.06.014

1 were female and 14 were male. With the exception of thoseho had the first seizure during the neonatal period, average first

eizure age was 28.9 months (range: 5–96 months). Status epilepti-us developed in 3 patients and was connected to febrile seizures in

.14) 4.95 (±2.05) 0.009

1 of them. Twenty-nine (82.8%) epileptic patients had 1 seizure typewhile 6 had multiple types (17.2%). Seizures were generalized tonicclonic in 13 patients, generalized tonic in 6, complex partial in 11,simple partial in 2, partial with secondary generalization in 2, infan-tile spasms in 3, atypical absence in 2, myoclonic in 1 and atonicin 1. Out of patients with febrile convulsions, the one with historyof status epilepticus had bilateral frontocentral sharp waves on theinterictal EEG, while the other 5 had normal EEG findings. Amongepileptic patients, 3 had normal interictal EEG findings, 16 had focalfindings and another 16 had multifocal/generalized findings. Thosewith focal EEG findings had a significantly better neurodevelop-mental and cognitive level than those with multifocal/generalizedEEG findings (p = 0.024).

While 4 epilepsy patients could be taken off antiepileptic drugs,12 stayed seizure free for over a year on monotherapy, and 5required polytherapy. In 4 patients on monotherapy and 2 on poly-therapy, seizures continued at a rate less than once a month, whilein 8 (22.8%) patients they continued at a rate of more than oncea month despite polytherapy. Three patients who developed Westsyndrome (aged 3 years 8 months, 4 years 6 months and 6 years 5months upon evaluation) remained seizure free after monotherapy.

When epilepsy-related factors were analyzed, history of neona-tal seizures was found to be strongly correlated with developmentof epilepsy (p = 0.001). Prematurity (<37 gestational weeks), new-born asphyxia, intraventricular hemorrhage and newborn sepsisdid not show correlation with epilepsy (respectively, p = 0.097,p = 0.105, p = 0.608 and p = 0.129), and neither did mechanicalventilation (p = 0.763). Severe neurodevelopmental or cognitiveimpairment was correlated with epilepsy (p = 0.010). Cerebral palsywas found in 90 patients (83.3%). Of the patients with cerebralpalsy, 33 had diplegia, 31 had hemiplegia and 26 had quadriple-gia. There was no correlation between the type of cerebral palsyand epilepsy (p = 0.069) (Table 1). Gross motor development wascompatible with their age in 4 of the epileptic patients (11.4%).

When radiological findings were evaluated, more severe loss ofwhite matter, periventricular hyperintensity and corpus callosuminvolvement were found to be correlated with epilepsy (respec-tively, p = 0.004, p = 0.016, and p = 0.004) Total MRI score was higherin epileptic patients (p = 0.009) (Table 1).

4. Discussion

Approximately one fifth of the PVL patients observed at ourpediatric neurology outpatient clinic had a history of neona-tal seizures. Similarly, a population-based observational study of

with periventricular leukomalacia. Clin Neurol Neurosurg (2013),

neonatal seizures in 545 very low birthweight infants with PVLreports 18.7% seizure incidence [4]. The same study reports the fol-lowing significant independent predictors of seizures among theseinfants: decreasing gestational age, intraventricular hemorrhage,

ING Model

C

and N

piaicitmab

sro

bP6cWmtcr

wtsrlasat

ettgeP

cwwibrTn

httMP

[

[

[

[

[

[

ARTICLELINEU-3370; No. of Pages 3

B. Ekici et al. / Clinical Neurology

osthemorrhagic hydrocephalus, sepsis, and necrotizing enterocol-tis. Prematurity was correlated with neonatal seizures in our studys well, but the fact that our patient group also included full termnfants points to newborn asphyxia being a major factor. Molecularascades induced by hypoxia had been proposed as a crucial factorn the pathophysiology of epileptogenesis [5]. Hypoxia-inducibleranscription factors are potential future diagnostic and prognostic

arkers as they modulate cerebral hypoxic stress responses andctivate endogenous neuroprotective systems in the developingrain [6].

We attribute our inability to find a correlation between sep-is and epilepsy in our study group to the fact that patients wereeferred to us from different neonatal centers, where the diagnosisf sepsis was often made based on clinical evaluation.

After scanning the data of PVL patients at a cerebral palsy reha-ilitation center, Humphreys et al. reported epilepsy in 26% of 154VL patients, while Fukuda et al. report a 30% rate in their study of3 PVL patients observed for at least 5 years [7,8]. In our tertiaryenter-based study, this rate was also approximately one in three.

e consider newborn seizures to be a major red flag for develop-ent of epilepsy. Similarly to our study, Humphreys et al. reported

hat only the presence of neonatal seizures was significantly asso-iated with epilepsy once other risk factors were controlled in theegression model [7].

The most common description of seizure patterns in our patientsas consistent with generalized tonic clonic and complex partial

ypes of seizures. These types were also the most common in thetudy by Humphreys et al., who relied on parent interviews andetrospective evaluation of data [7]. Considering that symptomaticocalization-related epilepsy is expected in PVL patients, we think

portion of seizures described as generalized tonic clonic is in factecondary generalized partial. And indeed, studies by Gurses et al.nd Fukuda et al. report partial seizures as the most frequent seizureype [8,9].

In our study, seizures continued with varying frequency in 14pileptic patients (40%) despite antiepileptic treatment. The facthat only 4 patients could be taken off antiepileptics indicateshat PVL-related epilepsy is refractory as the whole cerebral palsyroup [10,11]. Taking into account that 4 of our patients developedpilepsy without cerebral palsy, we would also like to stress thatVL can lead to different neurological problems.

According to studies that have evaluated epilepsy incidence inhildren with cerebral palsy, epilepsy is more common in patientsith pronounced cognitive impairment [12,13]. In our study asell, those with extensive EEG findings had more severe cognitive

mpairment, which leads us to think that this correlation worksoth ways. Our results were not compatible with former studieseporting the high prevalence of epilepsy in hemiplegic patients.his disparity might be explained by exclusion of patients withear-term hypoxic MRI findings.

The correlation between MRI findings and neurologic prognosisas been studied before. Fujimoto et al. [14] have linked white mat-

Please cite this article in press as: Ekici B, et al. Epilepsy in children

http://dx.doi.org/10.1016/j.clineuro.2013.06.014

er loss to severity of cerebral palsy and cognitive impairment. Inwo studies based on ventricular dimensions and signal changes on

RI findings, Serdaroglu et al. [15] found that patients with stage 4VL had more extensive motor and mental involvement than those

[

[

PRESSeurosurgery xxx (2013) xxx– xxx 3

with stage 1 PVL, while Hashimoto et al. [16] found that radiologicstages were correlated with neurologic prognosis. In our study,more severe loss of white matter, periventricular hyperintensityand corpus callosum involvement were found to be correlated withepilepsy. The link between epilepsy and severity of radiologicalfindings seen in our study was also detected by Humphreys et al.,who speculated that this link pointed to the epileptogenic natureof isolated white matter lesions [7]. However, we think that whatleads to seizures is the abnormalities that develop following whitematter trauma and involve cerebral neuronal regions includingboth the cortex and deep nuclear structures [17].

To summarize, almost a third of PVL patients in our study devel-oped epilepsy, which was intractable in approximately half of thecases. Epilepsy may develop in PVL patients even when cerebralpalsy does not. In these patients, neonatal seizures and severe MRIfindings are important clues that can indicate the development ofepilepsy.

References

[1] Bass WT. Periventricular leukomalacia. Neoreviews 2011;12:76–83.[2] Woodward LJ, Anderson PJ, Austin NC, Howard K, Inder TE. Neonatal MRI to pre-

dict neurodevelopmental outcomes in preterm infants. New England Journalof Medicine 2006;355:685.

[3] Fazzi E, Bova S, Giovenzana A, Signorini S, Uggetti C, Bianchi P. Cognitive visualdysfunctions in preterm children with periventricular leukomalacia. Develop-mental Medicine and Child Neurology 2009;51:974–81.

[4] Kohelet D, Shochat R, Lusky A, Reichman B. Israel Neonatal Network Risk factorsfor seizures in very low birthweight infants with periventricular leukomalacia.Journal of Child Neurology 2006;21:965–70.

[5] Schaller B. Influences of brain tumor-associated pH changes and hypoxia onepileptogenesis. Acta Neurologica Scandinavica 2005;111:75–83.

[6] Trollmann R, Gassmann M. The role of hypoxia-inducible transcription factorsin the hypoxic neonatal brain. Brain and Development 2009;31:503–9.

[7] Humphreys P, Deonandan R, Whiting S, Barrowman N, Matzinger MA, Briggs V,et al. Factors associated with epilepsy in children with periventricular leuko-malacia. Journal of Child Neurology 2007;22:598–605.

[8] Fukuda K, Kirino T, Fujiwara Y, Nagai S, Endo S. Clinical aspects of epilepsyin children with periventricular leukomalacia. No To Hattatsu 2010;42:291–5.

[9] Gurses C, Gross DW, Andermann F, Bastos A, Dubeau F, Calay M, et al. Periven-tricular leukomalacia and epilepsy: incidence and seizure pattern. Neurology1999;52:341–5.

10] Kułak W, Sobaniec W. Risk factors and prognosis of epilepsy in chil-dren with cerebral palsy in north-eastern Poland. Brain and Development2003;25:499–506.

11] Singhi P, Jagirdar S, Khandelwal N, Malhi P. Epilepsy in children with cerebralpalsy. Journal of Child Neurology 2003;18:174–9.

12] Carlsson M, Hagberg G, Olsson I. Clinical and aetiological aspects of epilepsyin children with cerebral palsy. Developmental Medicine and Child Neurology2003;45:371–6.

13] Mert GG, Incecik F, Altunbasak S, Herguner O, Mert MK, Kiris N, et al. Fac-tors affecting epilepsy development and epilepsy prognosis in cerebral palsy.Pediatric Neurology 2011;45:89–94.

14] Fujimoto S, Togari H, Banno T, Takashima S, Funato M, Yoshioka H, et al.Correlation between magnetic resonance imaging and clinical profiles ofperiventricular leukomalacia. Tohoku Journal of Experimental Medicine1999;188:143–51.

15] Serdaroglu G, Tekgul H, Kitis O, Serdaroglu E, Gökben S. Correlativevalue of magnetic resonance imaging for neurodevelopmental outcome inperiventricular leukomalacia. Developmental Medicine and Child Neurology2004;46:733–9.

with periventricular leukomalacia. Clin Neurol Neurosurg (2013),

16] Hashimoto K, Hasegawa H, Kida Y, Takeuchi Y. Correlation between neuroimag-ing and neurological outcome in periventricular leukomalacia: diagnosticcriteria. Pediatrics International 2001;43:240–5.

17] Inder TE, Warfield SK, Wang H, Hüppi PS, Volpe JJ. Abnormal cerebral structureis present at term in premature infants. Pediatrics 2005;115:286–94.