Epilepsy in Low-Grade Gliomas: The Impacton Cognitive Function and Quality of Life

Martin Klein, PhD,1 Nadine H. J. Engelberts, PhD,1 Henk M. van der Ploeg, PhD,1

Dorothee G. A. Kasteleijn-Nolst Trenite, MD, PhD,2 Neil K. Aaronson, PhD,3

Martin J. B. Taphoorn, MD, PhD,4 Hans Baaijen, MD,5 W. Peter Vandertop, MD, PhD,5

Martin Muller, MSc,3 Tjeerd J. Postma, MD, PhD,6 and Jan J. Heimans, MD, PhD6

Low-grade gliomas frequently are associated with epilepsy. The purpose of this study is to determine the impact ofepilepsy and antiepileptic drug (AED) treatment on cognitive functioning and health-related quality of life (HRQOL) inthese patients. One hundred fifty-six patients without clinical or radiological signs of tumor recurrence for at least 1 yearafter histological diagnosis and with an epilepsy burden (based on seizure frequency and AED use) ranging from noneto severe were compared with healthy controls. The association between epilepsy burden and cognition/HRQOL was alsoinvestigated. Eighty-six percent of the patients had epilepsy and 50% of those using AEDs actually were seizure-free.Compared with healthy controls, glioma patients had significant reductions in information processing speed, psychomo-tor function, attentional functioning, verbal and working memory, executive functioning, and HRQOL. The increase inepilepsy burden that was associated with significant reductions in all cognitive domains except for attentional andmemory functioning could primarily be attributed to the use of AEDs, whereas the decline in HRQOL could be ascribedto the lack of complete seizure control. In conclusion, low-grade glioma patients suffer from a number of neuropsycho-logical and psychological problems that are aggravated by the severity of epilepsy and by the intensity of the treatment.

Ann Neurol 2003;54:514–520

Epileptic seizures are the first symptom of intracranialtumors in 30 to 90% of patients.1 Patients with supra-tentorial low-grade gliomas have a higher incidence ofepileptic seizures than high-grade gliomas.2 The patho-physiological mechanism causing epileptic seizures inthese patients is unclear, however. Several studies sug-gest that seizures hardly ever originate from the masslesion itself but from adjacent brain tissue.3,4 Althoughboth ionizing radiation5 and neurosurgery3,6,7 may sig-nificantly reduce or eliminate medically refractory sei-zures associated with low-grade gliomas, the role ofthese treatment modalities in prolonging survival (asopposed to a policy of observation) is still controver-sial.8

With 5- and 10-year progression-free rates of 50 and12%, respectively, for supratentorial low-grade astrocy-tomas, low-grade oligodendrogliomas and mixed glio-mas,9 and a median overall survival of 16.7 years forthe latter two groups,10 low-grade glioma patients maysurvive in a stable state for many years after diagnosis.In a recently performed study,11 we found that apart

from the tumor itself and the relatively minor impactof radiotherapy, glioma patients taking AEDs hadlower levels of cognitive functioning than those with-out AEDs. The older AEDs are known to produce de-creased cognitive functioning, depression, and irritabil-ity,12 effects that are amplified with the administrationof multiple AEDs.13 Toxicity in AED polytherapy maybe related to serum AED concentrations, rather than tothe number of drugs administered.14 Inadequate, de-layed treatment of seizures may result in an increasingseizure frequency with a concomitant deterioration ofcognitive abilities15 which are furthermore also affectedby seizure type, duration, and severity, the underlyingneuropathology, and the anatomical site of the lesion.1

A higher epilepsy burden expressed by seizure type andfrequency has been shown to affect health-related qual-ity of life (HRQOL) in epilepsy patients.16

This study aimed at determining the impact of epi-leptic seizures and the associated use of AED mono-therapy or AED polytherapy on the mid to long-termcognitive functioning and HRQOL of low-grade gli-

From the 1Department of Medical Psychology, VU UniversityMedical Center; 2Medical Center Alkmaar; 3Division of Psychoso-cial Research and Epidemiology, Netherlands Cancer Institute, Am-sterdam; 4Department of Neurology, University Medical CenterUtrecht; Departments of 5Neurosurgery and 6Neurology, VU Uni-versity Medical Center, Amsterdam, The Netherlands.

Received Aug 27, 2002, and in revised form Mar 28. Accepted forpublication Jun 15, 2003.

Address correspondence to Dr Klein, VU University Medical Cen-ter, Department of Medical Psychology, Van der Boechorststraat 7,1081 BT Amsterdam, The Netherlands,E-mail: m.klein.psychol@med.vu.nl

514 © 2003 American Neurological AssociationPublished by Wiley-Liss, Inc., through Wiley Subscription Services

oma patients. We expected a higher epilepsy burden tohave an additional negative impact on the already com-promised cognitive functioning and HRQOL of low-grade glioma patients.

Patients and MethodsPatientsThis study was performed as part of a large nationwide studyinto the neuropsychological status and HRQOL of low-gradeglioma patients. The methodology of the study has been de-scribed in detail elsewhere.11 In short, patients (1) with low-grade astrocytomas, oligodendrogliomas or oligoastrocyto-mas, (2) without clinical signs of tumor recurrence for atleast 1 year after the histological diagnosis and primary treat-ment, and (3) without radiological signs of recurrence within3 months before testing were recruited from neurosurgicalcenters throughout the Netherlands (Vrije Universiteit Med-ical Center; Netherlands Cancer Institute/Slotervaart Hospi-tal; Academic Medical Center; Leiden University MedicalCenter; Westeinde Hospital; Leyenburg Hospital; ErasmusUniversity Medical Center Rotterdam; University MedicalCenter Utrecht; St. Elisabeth Hospital; Maastricht UniversityHospital; Maastricht University; Hospital De Wever; Uni-versity Hospital Nijmegen; Martini Hospital; and UniversityHospital Groningen). Eligibility was checked with the gen-eral practitioner and by case note review. Patients were ex-cluded if they used corticosteroids, did not have a basic pro-ficiency in the Dutch language, or were unable tocommunicate adequately. Irradiated patients were includedonly if they received radiotherapy as primary treatment.

In addition to glioma patients, normative data derivedfrom a cohort of healthy controls17 were included in thisstudy to provide an additional anchor for interpreting theresults. Healthy controls were matched individually with pa-tients for age, sex, and educational level. Informed consentprocedures preceded patients’ participation. Assessment in-cluded the collection of clinical data on functional status,cognitive evaluation, and self-reported HRQOL. Clinicaldata derived from the case note reviews included tumor char-acteristics (histology, site), epilepsy history, and treatmenthistory (ie, neurosurgery [biopsy vs resection], radiotherapytarget [focal vs whole-brain], radiotherapy total dose andfraction dose, boost, and medication use [corticosteroids, an-tiepileptic drugs]).

Because of the variety in seizure history and antiepilepticdrug (AED) use, patients were assigned to one of six levelson a single ordinal scale, with higher levels representing ahigher epilepsy burden: (level 1) epilepsy-free; (level 2) epi-lepsy, seizure-free in the year before testing without AEDs;(level 3) epilepsy, seizure-free in the previous year with AEDmonotherapy; (level 4) epilepsy, seizure-free in the previousyear with AED polytherapy; (level 5) epilepsy, less than sixseizures in the previous year and on AED monotherapy orpolytherapy; and (level 6) epilepsy, more than six seizures inthe previous year and on AED monotherapy or polytherapy.

Study MeasuresPERFORMANCE STATUS Performance status was assessedby means of the Karnofsky performance status (KPS) scale,18

which is an overall indicator of the patient’s level of physicalfunctioning.

The capacity to perform activities of daily living (ADL)was assessed by means of the Barthel Activities of Daily Liv-ing Index (BADLI).19 The BADLI consists of 10 items as-sessing continence of bowel and bladder, grooming, toiletuse, feeding, transfer, mobility, dressing, climbing stairs, andbathing.

COGNITIVE PERFORMANCE STATUS. Trained psychome-tricians, supervised by a board-certified neuropsychologist(M.K.) administered a neuropsychological test battery con-sisting of the following tests: Line Bisection Test,20 FacialRecognition Test,20 Judgment of Line Orientation Test,20

Letter-Digit Substitution Test,20 Visual Verbal LearningTest,20 Working Memory Task,20 Stroop Color-WordTest,20 Categoric Word Fluency Task,20 and Concept Shift-ing Test.21 To accomplish data reduction, we calculatedsummary measures to detect possible deficits in the cognitivedomains of (1) information processing speed, (2) psychomo-tor function, (3) attentional functioning, (4) verbal memory,(5) working memory, and (6) executive functioning. Con-struction of these cognitive domains was based on a PrincipalComponent Analysis (PCA) using Varimax rotation withKaiser normalization performed on the Z-scores (standardequivalents) of the healthy controls.17 The outcome of thePCA was a confirmation of the cognitive domains conven-tionally used in neuropsychological practice. Individual cog-nitive test scores of patients were converted to Z-scores, us-ing the means and standard deviations of the matchedhealthy controls as a reference. Cognitive domain summarymeasures (means) based on these Z-scores were calculated forall six patient groups.

HEALTH-RELATED QUALITY OF LIFE HRQOL was as-sessed by means of the Medical Outcomes Study (MOS)Short-Form Health Survey (SF-36).22 The SF-36 is a self-report questionnaire and is composed of 36 items, organizedinto 8 multiitem scales assessing physical functioning, rolelimitations due to physical health problems, bodily pain,general health perceptions, vitality, social functioning, rolelimitations due to emotional problems, and mental health.The SF-36 also yields two higher order component scores,one for Physical Health (PCS) and one for Mental Health(MCS).

Statistical Analysis�2 tests, Student’s t tests for independent samples, and anal-ysis of variance were conducted to test for significant differ-ences in sociodemographic and tumor/treatment characteris-tics between patient groups. Analysis of covariance,correcting for differences in age, sex, and education was con-ducted to test for significant differences between the com-bined patient groups and healthy controls in the six cognitivesummary measures and HRQOL. To identify whether ahigher epilepsy burden was associated with limitations incognitive functioning, we performed a series of stepwise lin-ear regression analyses with possible confounders (ie, age, sex,and education) forced into the model at the first step. Sub-sequently, epilepsy burden scored on the previously described

Klein et al: Epilepsy in Low-Grade Glioma 515

single ordinal scale was entered as independent variable. Thesix cognitive summary measures, the HRQOL higher ordercomponent scores (PCS and MCS), performance status(KPS), and the capacity to perform ADL activities (BADLI)were entered as the dependent variables. The level of statis-tical significance for all tests was set at p value less than 0.05.Probability for entry in the regression model was set at 0.05and probability for removal was set at 0.10.

ResultsSociodemographic and Clinical CharacteristicsOf the 239 adult patients with supratentorial low-gradegliomas who met the inclusion criteria, a total of 195patients agreed to participate in this nationwide study.Forty-four patients (18%) declined participation be-cause they found it too burdensome, or they were re-luctant to be confronted with what they considered tobe a “cured” illness. Approximately half of these pa-tients (53%) received conventional radiotherapy 1 to22 years previously. Focal radiotherapy was normallyapplied, although 10% had whole-brain radiotherapy.The main reason for whole-brain radiotherapy was alarge tumor volume, which made focal radiation verydifficult. Because of missing information, the data of

39 (20%) of the 195 patients recruited in this studydid not permit us to reliably calculate the epilepsy bur-den based on seizure frequency and AED use, and thusthese patients were excluded from the analysis.

Table 1 shows the sociodemographic, tumor, andtreatment characteristics of patients stratified by epi-lepsy burden, and of whom complete data sets wereretrievable from the case note reviews. Patients withmissing data concerning epilepsy did not differ signif-icantly from those patients for whom such data wereavailable. Patient groups did not differ significantly insex, educational level, histological diagnosis, tumor lat-eralization, neurosurgical intervention, or in the use ofradiotherapy. However, patients with a higher epilepsyburden tended to be older (p � 0.044). Althoughmost of the tumors were located in the frontal lobes,patient groups differed in the distribution of nonfron-tal tumors (�2 test, p � 0.005). A consistent findingwas that with increasing epilepsy burden there was ahigher proportion of tumors in the temporal lobes.

In line with the international literature on low-gradegliomas, we found that approximately 14% of all pa-tients did not have a seizure history. Seizure control

Table 1. Sociodemographic, Tumor, and Treatment Characteristics of 156 Glioma Patients Stratified by Epilepsy Burden

Characteristics

Glioma Patient Groupsa

1(n � 21)

2(n � 21)

3(n � 33)

4(n � 24)

5(n � 24)

6(n � 33)

Mean age at testing (SD)b (yr) 35.8 (10.1) 37.7 (11.5) 43.4 (11.7) 43.5 (12.2) 45.7 (13.2) 41.0 (8.5)Sex: M (%) 43 62 68 70 58 67Level of education (SD) 4.2 (2.2) 4.5 (2.2) 4.1 (1.8) 4.0 (2.3) 3.7 (2.0) 4.7 (2.1)Histological diagnosis

Astrocytoma (%) 81 71 56 70 67 79Oligodendroglioma (%) 14 24 35 26 33 12Oligoastrocytoma (%) 5 5 9 4 — 9

Tumor lateralizationLeft-sided (%) 43 62 53 48 42 58Right-sided (%) 52 29 47 52 50 39Bilateral (%) 5 10 — — 8 3

Tumor localizationc

Frontal (%) 33 43 68 33 58 34Parietal, occipital (%) 24 24 9 29 — 30Temporal (%) 19 19 23 38 38 30Deep structures (%) 14 9 — — — —Other (%) 10 5 — — 4 6

Neurosurgical interventionBiopsy (%) 38 33 52 57 33 55Resection (%) 62 67 48 43 67 45

RadiotherapyNo (%) 43 43 44 39 58 42Yes (%) 57 57 56 61 42 58

aPatient groups with higher levels representing a higher epilepsy burden; (level 1) epilepsy-free; (level 2) epilepsy, seizure-free in the year beforetesting without AED; (level 3) epilepsy; seizure-free in the previous year with AED monotherapy; (level 4) epilepsy, seizure-free in the previousyear with AED polytherapy; (level 5) epilepsy, � 6 seizures in the previous year and on AED monotherapy or polytherapy; (level 6) epilepsy,�6 seizures in the previous year and on AED monotherapy or polytherapy.bSignificant differences in mean age at testing between patient groups (analysis of variance p � 0.044).cSignificant differences in tumor localization between patient groups (�2 test, p � 0.005).

SD � standard deviation; AED � antiepileptic drug.

516 Annals of Neurology Vol 54 No 4 October 2003

was attained in 50% of patients using AEDs (ie,Groups 3, 4, 5, and 6). Table 2 shows the AED use ofpatients stratified by epilepsy burden. In this sample ofpatients, carbamazepine was prescribed relatively oftenas a first-line therapy. Only 4% of the patients receivedphenobarbital monotherapy, whereas newer AEDs,which included lamotrigine, oxcarbazepine, topiramate,and vigabatrin were only rarely prescribed and appar-ently are not common practice as first-line or second-line treatment in these patients.

Neuropsychological PerformanceSummary measures of cognitive functioning in the do-mains of information processing speed, psychomotorfunction, and attentional functioning are depicted inFigure 1. Analysis of covariance, controlling for differ-ences in age, sex, and education, demonstrated that thecombined patient groups had statistically significantlower scores in all cognitive domains (all: p � 0.000)than healthy controls. Linear regression analysis dem-onstrated that a higher epilepsy burden furthermorewas associated with a lower information processingspeed and impaired psychomotor function (p � 0.046and p � 0.001, respectively) but not with a deteriora-tion in attentional functioning.

Verbal memory capacity, working memory capacity,and executive functioning are illustrated in Figure 2.Although the deviations of patients from healthy con-trols are less outspoken than in the cognitive domainsdiscussed previously, patients’ verbal memory capacity,working memory capacity, and executive functioningare well below the levels attained by healthy controls

(analysis of covariance, all: p � 0.000). In the regres-sion analysis, epilepsy burden was not significantly as-sociated with the verbal memory capacity already com-promised by the glioma. However, a higher epilepsyburden was strongly associated with a decrease in bothworking memory capacity and in executive functioning(p � 0.000 and p � 0.003, respectively). Patients us-ing AED monotherapy or polytherapy (combinedGroups 3 through 6) had significantly more limitationsin information processing capacity (Student’s t test:p � 0.004), psychomotor function (Student’s t test:p � 0.001), attentional functioning (Student’s t test:p � 0.006), working memory capacity (Student’s ttest: p � 0.003), and executive functioning (Student’st test: p � 0.000) than those not using AEDs (com-bined Groups 1 and 2).

Health-Related Quality of Life, Performance Status,and Activities of Daily LivingCompared with healthy controls, the combined patientgroups attained lower levels of physical functioning(p � 0.003), role limitations due to physical healthproblems (p � 0.000), bodily pain (p � 0.001), gen-eral health perceptions (p � 0.000), vitality (p �0.000), social functioning (p � 0.013), and role limi-tations due to emotional problems (p � 0.002). Linearregression analyses controlling for differences in age,sex, and education indicated that a higher epilepsy bur-den was strongly associated with a decrease in bothPhysical Health (PCS: p � 0.000) and Mental Healthscores (MCS: p � 0.001).

Figure 3 depicts standardized PCS and MCS scores

Table 2. Antiepileptic Drug Use of 156 Glioma Patients Stratified by Epilepsy Burden

AED

Glioma Patient Groupsa

1(n � 21)

2(n � 21)

3(n � 33)

4(n � 24)

5(n � 24)

6(n � 33)

Type of AED (%)b

Carbamazepine (1) — — 50 — 29 18Valproic acid (2) — — 24 — 13 6Phenytoin (3) — — 20 — 21 24Phenobarbital (4) — — — — 4 —

1 � 2 — — — 26 8 121 � 3 — — — 13 — 31 � 4 — — — 9 — —2 � 3 — — — — 4 32 � 4 — — — — — 33 � 4 — — — 9 — —�2 AEDs — — — 9 4 18Other combinations or AEDs — — 6 34 17 13

aPatient groups with higher levels representing a higher epilepsy burden; (level 1) epilepsy-free; (level 2) epilepsy, seizure-free in the year beforetesting without AED; (level 3) epilepsy, seizure-free in the previous year with AED monotherapy; (level 4) epilepsy, seizure-free in the previousyear with AED polytherapy; (level 5) epilepsy, �6 seizures in the previous year and on AED monotherapy or polytherapy; (level 6) epilepsy,�6 seizures in the previous year and on AED monotherapy or polytherapy.bSignificant differences between patient groups in AED use (�2 test, p � 0.000).

AED � antiepileptic drug.

Klein et al: Epilepsy in Low-Grade Glioma 517

of patients stratified by epilepsy burden relative to thatof healthy controls represented by the “50”-line, withhigher scores indicating better HRQOL. Patients withongoing seizures (combined Groups 5 and 6) had sig-nificantly lower self-reported physical (PCS, Student’s ttest: p � 0.000) and mental health (MCS, Student’s ttest: p � 0.001) than seizure-free patients (combinedGroups 1 through 4).

A higher epilepsy burden was not associated withlower levels of physical functioning as measured by theKPS scale or by limitations in patients’ capacity to con-duct ADL activities as assessed by means of theBADLI.

DiscussionThe aim of this study was to determine the impact ofepilepsy and its treatment on the cognitive functioning

and HRQOL of low-grade glioma patients. The overallresults indicate that glioma patients, regardless of epi-lepsy burden, have lower levels of cognitive functioningand HRQOL than healthy controls. These lower levelsof cognitive functioning and health-related quality oflife in glioma patients compared with healthy controlsare not unexpected, because our patient sample com-prises the majority of patients from our former study.11

In this study, we moreover observed that a higher ep-ilepsy burden was associated with a more severe dete-rioration in several cognitive domains, indicating thatthe lower cognitive deficits of glioma patients are ag-gravated by the burden of epilepsy and/or its treatmentwith AEDs. An additional analysis indicated that pa-tients using AEDs performed worse in all cognitive do-mains, except for verbal memory performance thanthose not using AEDs. These findings suggest that pri-marily AED therapy rather than seizure frequency neg-atively affect cognitive function in these patients

Fig 1. Mean Z-scores on the cognitive domains of informationprocessing speed, psychomotor function, and attentional func-tioning of the 156 glioma patients stratified by epilepsy burdenrelative to that of age-, sex-, and education-matched healthycontrols represented by the “0”-line. A higher score means bet-ter performance. A higher epilepsy burden was associated witha lower information processing speed (p � 0.046) and im-paired psychomotor function (p � 0.001). Patients usingAEDs (Groups 3–6) had more limitations in informationprocessing capacity (p � 0.004), psychomotor function (Stu-dent’s t test: p � 0.001), attentional functioning (Student’s ttest: p � 0.006) than those not using AEDs (Groups 1 and2).

Fig 2. Mean Z-scores on summary measures of verbal memory,working memory, and executive functioning of the 156 gliomapatients stratified by epilepsy burden relative to that of age-,sex-, and education-matched healthy controls represented by the“0”-line (higher score means better performance). A higherepilepsy burden was associated with a decrease in workingmemory capacity (p � 0.000) and executive functioning (p �0.003). Patients using antiepileptic drugs (AEDs) (Groups3–6) had more limitations in working memory capacity (p �0.003) and executive functioning (Student’s t test: p �0.000) than those not using AEDs (Groups 1 and 2).

518 Annals of Neurology Vol 54 No 4 October 2003

mostly using established AEDs. The newer AEDs, in-cluding gabapentin, lamotrigine, and oxcarbazepine,might control seizures as effectively as the establishedAEDs, and possibly have fewer and less severe side ef-fects.23

This study confirms the findings of a large-scale Eu-ropean study in epilepsy patients that indicates thatHRQOL as measured by SF-36 scores varies accordingto epilepsy burden expressed by seizure type and fre-quency and by the stigma associated with it.16 Inter-estingly, the data of the European study suggest thatthe HRQOL of the glioma patients in this study isequivalent to that of epilepsy patients without gliomas.Regression analysis indicated that epilepsy burden waspredictive for self-perceived physical and mental healthas expressed by PCS and MCS scores, respectively. Inaddition, we found that ongoing seizures rather thanAED use had the most deleterious impact onHRQOL. This finding is supported by a prospective

study in which those patients who became completelyseizure-free reported significantly more positive changein HRQOL than those who did not.24

Although a higher epilepsy burden in this study wasassociated with both lower cognitive functioning andreduced HRQOL, this was not reflected by lower levelsof physical functioning (KPS) or by limitations in pa-tients’ capacity to perform activities of daily living(BADLI). This suggests that these scales may have lim-itations to accurately document differences between pa-tients differing in epilepsy characteristics or that theydo not capture relevant and meaningful problems orsymptoms within the low-grade glioma populationconfronted with epilepsy.

This study demonstrated that AED monotherapy orpolytherapy appears to be inadequate for seizure con-trol in 50% of glioma patients. This number is com-parable to that of a large study into the prognosticvalue of the epilepsy type and syndromes for seizurerecurrence and in which seizure-freedom was attainedin 46% of symptomatic patients with tumors.25 Themost common causes of AED treatment failure includeuse of a suboptimal medication for a particular seizuretype, inability to properly determine the seizure typeby standard examination and testing methods, occur-rence of certain neurological or psychological symp-toms that can mimic seizures, or failure to take medi-cation as prescribed. Currently, there is no “perfect”therapy for any particular seizure type that will be suit-able for all patients, and the decision to treat or not totreat at all after a single first seizure remains controver-sial for many patients.26,27 In low-grade glioma pa-tients, however, AED treatment with valproic acid orcarbamazepine usually is initiated after the first seizure.Prognosis is primarily a function of the underlying ep-ilepsy syndrome, and although treatment with AEDsdoes reduce the risk of subsequent seizures and addi-tional cognitive damage, it does not alter the long-termprognosis for seizure control and remission.27 If twomonotherapy trials are unsuccessful, the chance ofachieving seizure control with additional agents is lessthan 20%.28

Several studies suggest that better seizure controland/or reduction in AED use might be accomplishedin low-grade glioma patients by surgical interven-tion.3,6,7 This option should be considered specificallyin patients with AED resistance early in the course ofepilepsy29 and whose originating focus can be confi-dently localized to a well-defined resectable area (notinvolving eloquent cortex). If possible, surgery in thesepatients should not be limited to tumor resection, butalso include resection of epileptogenic foci. Whetherlow-grade glioma patients with refractory epilepsy in-deed may benefit from such a surgical intervention re-mains to be confirmed by randomized clinical trials.

Fig 3. Health-related quality-of-life SF-36 higher order com-ponent scores: Physical Health (PCS) and Mental Health(MCS) of 156 glioma patients stratified by epilepsy burdenrelative to that of age-, sex-, and education-matched healthycontrols represented by the “50”-line. Higher score means bet-ter HRQOL. A higher epilepsy burden was associated with adecrease in PCS (p � 0.000) and MCS scores (p � 0.001).Patients with ongoing seizures (Groups 5 and 6) had signifi-cantly lower self-reported physical (PCS, Student’s t test: p �0.000) and mental health (MCS, Student’s t test: p �0.001) than seizure-free patients (Groups 1–4).

Klein et al: Epilepsy in Low-Grade Glioma 519

This study was supported by a grant from the Dutch Cancer Society(VU96-1155, M.K.).

We thank G. J. Ossenkoppele, B. J. Slotman, W. Boogerd, S. Leen-stra, J. H. C. Voormolen, W. M. Mackay, R. Vriesendorp, N. Lam-booy, P. W. Wijermans, C. J. J. Avezaat, A. W. Dekker, H. Stru-ikmans, G. N. Beute, C. van der Heul, A. Twijnstra, H. C.Schouten, J. Jolles, P. J. J. Koehler, R. H. Boerman, G. E. M. Kien-stra, J. J. Mooij, and G. W. van Imhoff, for permission to includetheir patients, and T. A. Albersen, W. Cleijne, J. Grit, N. Gulde-mond, and W. Oomen for their invaluable help in tracing and test-ing patients.

References1. Rasmussen T. Surgery of epilepsy associated with brain tumors.

Adv Neurol 1975;8:227–239.2. Shaw EG, Scheithauer BW, Fallon O. Management of supra-

tentorial low-grade gliomas. Oncology 1993;7:97–104.3. Berger MS, Ghatan S, Haglund MM, et al. Low-grade gliomas

associated with intractable epilepsy: seizure outcome utilizingelectrocorticography during tumor resection. J Neurosurg 1993;79:62–69.

4. Haglund MM, Berger MS, Kunkel DD, et al. Changes ingamma-aminobutyric acid and somatostatin in epileptic cortexassociated with low-grade gliomas. J Neurosurg 1992;77:209–216.

5. Chalifoux R, Elisevich K. Effect of ionizing radiation on partialseizures attributable to malignant cerebral tumors. StereotactFunct Neurosurg 1996;67:169–182.

6. Hennessy MJ, Elwes RD, Honavar M, et al. Predictors of out-come and pathological considerations in the surgical treatmentof intractable epilepsy associated with temporal lobe lesions.J Neurol Neurosurg Psychiatry 2001;70:450–458.

7. Zentner J, Hufnagel A, Wolf HK, et al. Surgical treatment ofneoplasms associated with medically intractable epilepsy. Neu-rosurgery 1997;41:378–386; discussion, 386–387.

8. DeAngelis LM. Brain tumors. N Engl J Med 2001;344:114–123.

9. Leighton C, Fisher B, Bauman G, et al. Supratentorial low-grade glioma in adults: an analysis of prognostic factors andtiming of radiation. J Clin Oncol 1997;15:1294–1301.

10. Olson JD, Riedel E, DeAngelis LM. Long-term outcome oflow-grade oligodendroglioma and mixed glioma. Neurology2000;54:1442–1448.

11. Klein M, Heimans JJ, Aaronson NK, et al. Effect of radiother-apy and other treatment-related factors on mid-term to long-term cognitive sequelae in low-grade gliomas: a comparativestudy. Lancet 2002;360:1361–1368.

12. Meador KJ. Cognitive side effects of medications. Neurol Clin1998;16:141–155.

13. Vining EP. Cognitive dysfunction associated with antiepilepticdrug therapy. Epilepsia 1987;28:S18–S22.

14. Deckers CL, Hekster YA, Keyser A, et al. Reappraisal of poly-therapy in epilepsy: a critical review of drug load and adverseeffects. Epilepsia 1997;38:570–575.

15. Jokeit H, Ebner A. Long term effects of refractory temporallobe epilepsy on cognitive abilities: a cross sectional study.J Neurol Neurosurg Psychiatry 1999;67:44–50.

16. Baker GA, Jacoby A, Buck D, et al. Quality of life of peoplewith epilepsy: a European study. Epilepsia 1997;38:353–362.

17. Bosma H, van Boxtel MP, Ponds RW, et al. Pesticide exposureand risk of mild cognitive dysfunction. Lancet 2000;356:912–913.

18. Karnofsky DA, Abelmann WH, Craver LF. The use of nitrogenmustards in the palliative treatment of carcinoma. Cancer 1948;1:634–656.

19. Wade DT, Collin C. The Barthel ADL Index: a standard mea-sure of physical disability? Int Disabil Stud 1988;10:64–67.

20. Lezak MD. Neuropsychological assessment. 3rd ed. New York:Oxford University Press, 1995.

21. Houx PJ, Jolles J. Vulnerability factors for age-related cognitivedecline. In: Isaacson RL, Jensen KF, eds. The vulnerable brainand environmental risks. Vol 3. New York: Plenum Press,1994:25–41.

22. Stewart AL, Ware JE, eds. Measuring functioning and well-being: the medical outcomes study approach. Durham, NC:Duke University Press, 1992.

23. Brodie MJ. Monostars: an aid to choosing an antiepileptic drugas monotherapy. Epilepsia 1999;40:S17–S22; discussion,S73–74.

24. Birbeck GL, Hays RD, Cui X, Vickrey BG. Seizure reductionand quality of life improvements in people with epilepsy. Epi-lepsia 2002;43:535–538.

25. Semah F, Picot MC, Adam C, et al. Is the underlying cause ofepilepsy a major prognostic factor for recurrence? Neurology1998;51:1256–1262.

26. Hauser WA, Rich SS, Lee JR, et al. Risk of recurrent seizuresafter two unprovoked seizures. N Engl J Med 1998;338:429–434.

27. Musicco M, Beghi E, Solari A, Viani F. Treatment of firsttonic-clonic seizure does not improve the prognosis of epilepsy.First Seizure Trial Group (FIRST Group). Neurology 1997;49:991–998.

28. Mattson RH, Cramer JA, Collins JF, et al. Comparison of car-bamazepine, phenobarbital, phenytoin, and primidone in par-tial and secondarily generalized tonic-clonic seizures. N EnglJ Med 1985;313:145–151.

29. Kwan P, Brodie MJ. Early identification of refractory epilepsy.N Engl J Med 2000;342:314–319.

520 Annals of Neurology Vol 54 No 4 October 2003