Epilepsy & Behavior 9 (2006) 181–185 www.elsevier.com/locate/yebeh

Brief Communication

Functional MRI reveals declined prefrontal cortex activation in patients with epilepsy on topiramate therapy Jacobus F.A. Jansen a,b,*, Albert P. Aldenkamp c,d, H.J. Marian Majoie c,d, Rianne P. Reijs c,d, Marc C.T.F.M. de Krom c, Paul A.M. Hofman b, M. Eline Kooi Klaas Nicolay a,b, Walter H. Backes a,b a

a,b

,

Biomedical NMR, Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands b Department of Radiology, Maastricht University Hospital, Maastricht, The Netherlands c Department of Neurology, Maastricht University Hospital, Maastricht, The Netherlands d Epilepsy Centre Kempenhaeghe, Heeze, The Netherlands Received 9 March 2006; revised 27 April 2006; accepted 3 May 2006

Abstract Functional magnetic resonance imaging of covert word generation was used to examine brain activation abnormalities associated with topiramate-induced cognitive language impairment in patients with epilepsy. Compared with a control epilepsy group, in the topiramatetreated group, there was significantly less activation in the language-mediating regions of the prefrontal cortex; the topiramate group also had significantly lower neuropsychological language scores. These findings suggest that topiramate has a critical effect on the cerebral neural systems that mediate expressive language.  2006 Elsevier Inc. All rights reserved. Keywords: Epilepsy; Topiramate; Cognition; Language; Neuropsychology; Functional magnetic resonance imaging

1. Introduction Cognitive impairment induced by antiepileptic drugs is a major issue in the treatment of epilepsy [1]. Many therapeutic dilemmas arise when adequate seizure control can be achieved only with medication that is associated with cognitive side effects. An example of a drug with tolerability-related issues is topiramate (TPM). It is a broad-spectrum antiepileptic drug used both in adjunctive therapy and in monotherapy for patients with epilepsy, and has also shown efficacy in the treatment of several other neurological and psychiatric diseases, including migraine [2]. Although TPM is beneficial for patients with epilepsy in terms of efficacy [3], studies based on subjective complaints have revealed a high frequency of TPM-induced cognitive adverse events [4,5]. These cognitive side effects are an impor*

Corresponding author. Fax: +31 43 387 6909. E-mail address: [email protected] (J.F.A. Jansen).

1525-5050/$ - see front matter  2006 Elsevier Inc. All rights reserved. doi:10.1016/j.yebeh.2006.05.004

tant reason for TPM discontinuation, even when the drug has a favorable effect on seizure frequency [6]. Additionally, neuropsychological studies provide clear clinical evidence of TPM-induced cognitive impairment [7,8]. It has been reported that primarily frontal lobe-associated executive functions, like verbal fluency and working memory, decline in patients taking TPM [8]. Secondarily, nonverbal functions may be affected as well, if these functions have a working memory or short-term memory component [9,10]. Hence, certain function-specific brain areas (i.e., prefrontal language- and memory-related areas) may be more greatly affected by TPM treatment than other areas. Possible alterations in the functional neuroanatomy of language caused by TPM treatment can be visualized with functional magnetic resonance imaging (fMRI) of covert word generation. The combination of fMRI and neuropsychological testing has greater value than neuropsychological testing alone, as fMRI enables the determination of differences in sensitivity of brain regions to TPM

182

Brief Communication / Epilepsy & Behavior 9 (2006) 181–185

treatment. Here we examine the effect of TPM treatment on patients with epilepsy using both methods, to gain insight into the underlying cortical activation patterns and modification thereof.

logical language score to each patient, on a nominal scale ranging from 1 (severe cognitive language problems) to 5 (no problems). Differences in scores between the two groups were analyzed using the Wilcoxon/ Mann–Whitney test.

2.3. Imaging procedures 2. Methods 2.1. Study subjects Five patients with epilepsy using TPM as add-on treatment (TPM group) and 10 control patients with epilepsy not taking TPM, but all on add-on polytherapy (control group), were selected during their visits to the outpatient neurology department of the Maastricht University Hospital or the Epilepsy Centre Kempenhaeghe (a tertiary epilepsy referral and care center). All patients were identified as having refractory epilepsies. Seizure classification and syndrome classification were based on history taking, a hetero-anamnesis, additional interictal electroencephalography, and MRI. The two groups did not differ with respect to most patient demographics and characteristics (e.g., age at onset of epilepsy, age at examination, and seizure frequency) (Table 1). All patients on TPM had a starting dose of 25 mg, and a titration schedule of 25 mg per week. These patients experienced self-reported and psychometric speech/language dysfunction, as well as attentional difficulties, during treatment. All patients were right-handed, as assessed by the Edinburgh Inventory handedness test [11], and had been seizure-free at least 3 weeks before inclusion. The study was approved by the Commission on Medical Ethics of Maastricht University Hospital, and all subjects gave written informed consent.

MRI was performed with a 1.5-T unit (Philips Intera, Philips Medical Systems, Best, The Netherlands). Functional MRI data were acquired using a whole-brain single-shot three-dimensional (3D) blood oxygen level-dependent echo-planar imaging sequence, with repetition time 2 s, echo time 50 ms, flip angle 90, voxel size 3.5 · 3.5 · 3.5 mm, matrix 64 · 64, 34 contiguous slices per volume, and 96 volumes per acquisition. For anatomical reference, we acquired a 3D T1-weighted fast field-echo image, with repetition time 11 ms, echo time 3.5 ms, flip angle 90, matrix 256 · 256, 150 contiguous slices, and voxel size 3.5 · 3.5 · 3.5 mm.

2.4. Cognitive language task During fMRI, subjects completed a standard expressive language task aimed at frontal activation (which includes naming, comprehension, and repetition), as opposed to receptive language, which encompasses more temporal/parietal activation. The task comprised the covert generation of words beginning with a different input letter (U-N-K-A-E-P) visually presented [13]. The activation paradigm consisted of six word-generation condition blocks (32 s each) alternating with six baseline rest condition blocks (32 s each).

2.5. Data analysis 2.2. Neuropsychological tests Language was assessed with an aphasia test battery that includes the word generation task and the naming test [12]. Using the results of these tasks, an experienced clinical neuropsychologist assigned a neuropsycho-

Spatial data preprocessing was performed with SPM2 software (Wellcome Department of Cognitive Neurology). The number of activated voxels was determined in predefined brain regions essential for language processing [14]. These bilateral regions comprise the inferior and middle

Table 1 Patient demographics and characteristics

Number of patients Male/female Age at onset of epilepsya Age at examinationa Seizure frequency (month 1)a Pathology Type epilepsy Seizure focusc Duration of epilepsy before introduction of TPM (years)a Duration of TPM treatment (days)a TPM dosage (mg)a TPM drug loada,d Full drug loada,d Other medication Neuropsychological language scorea Number of voxels with an activation level >0.9% in the language areasa

Patients with epilepsy on TPM

Patients with epilepsy not on TPM

5 0/5 11.0 ± 4.3 32.6 ± 11.6 4.8 ± 8.5 1 DNET, 4 UKb 5 CLRE 2 LT, 1 MF, 2 BL 19.6 ± 12.0

10 3/7 9.4 ± 3.5 35.3 ± 9.1 3.1 ± 1.5 8 HS, 1 Cyste, 1 UK 10 CLRE 5 LT, 5 RT NA

128.8 ± 120.6 170 ± 115 0.7 ± 0.5 2.0 ± 0.7 2 LEV, 2 CBZ, 1 GBP, 1 LTG 1.8 ± 0.8

NA NA NA 1.4 ± 0.7 1 LEV, 7 CBZ, 3 LTG, 1 VPA, 3 OXC 4.3 ± 0.7

291 ± 155

1861 ± 1150

Statistical significance

P = 0.18 P = 0.17 P = 0.31

P = 0.13 P = 0.002 P = 0.01

The bold P values denote P-values lower than 0.05, and therefore denote statistical significance. a Mean ± SD. b TPM, topiramate; HS, hippocampal sclerosis; DNET, dysembryoplastic neuroepithelial tumor; UK, unknown; CLRE, cryptogenic localizationrelated epilepsy; LT, left temporal; RT, right temporal; MF, multifocal; BL, bilateral; NA, not applicable; LEV, levetiracetam; OXC, oxcarbazepine; CBZ, carbamazepine; GBP, gabapentin; LTG, lamotrigine; VPA, valproate. c Based on the electroencephalogram. d Drug load is a quantitative assessment of total antiepileptic drug use calculated by standardizing the doses using the ratio of prescribed daily dose to defined daily dose [21].

Brief Communication / Epilepsy & Behavior 9 (2006) 181–185 prefrontal cortex (Brodmann areas, BA 44–47), the superior prefrontal cortex (BA 6, 9, and 10), the anterior cingulate cortex (ACC, BA 24 and 32), the anterior temporal lobe (BA 38), the posterior and inferior temporal lobe (BA 20 and 21), and the temporoparietal junction (BA 39 and 40). Quantitative analysis consisted of counting all voxels with an activation level, expressed as percentage signal change, higher than a threshold of 0.9% (mean + 2SD) with respect to baseline signal, within the predefined language areas of all individual activation maps. Activation differences between the TPM group and the control group were analyzed using a two-tailed Student t test. To visualize which brain regions display the largest or, conversely, the smallest differences in activation level in the TPM and control groups, the average differences with corresponding SD were calculated on a voxel-byvoxel basis for the whole brain. Then, from all voxels for which the absolute value of the activation difference level was larger than its SD, the strength of the deviation of those voxels from the global averaged activation difference level was calculated.

3. Results Neuropsychological assessment revealed that the language score was significantly lower in the TPM group than in the control group (Table 1). Furthermore, fMRI demonstrated that the main region activated in both groups was an extensive area of the known expressive language network, including the left inferior prefrontal cortex (IPC, i.e., Broca’s area (BA 44–45)), A

Control group

TPM group

3%

0.9 %

L

+10 mm

+10 mm

B

10 %

-10 % -22 mm

L

+20 mm

+44 mm

Fig. 1. (A) Coronal images of group mean functional MRI activation maps, obtained for the covert word generation paradigm overlaid on a spatially normalized T1-weighted MR image, with the control group on the left and the TPM group on the right. The activation maps consist of all voxels within the language areas that have an activation level >0.9%. For the control group, activation is clearly visible in the left inferior prefrontal cortex (IPC) and the medial prefrontal cortex (MPC), whereas the TPM group exhibits significant underactivation in these regions. (B) Transverse images of average underactivation level for the whole brain of the TPM group as compared with that of the control group. The prefrontal cortex (IPC and MPC) uniformly exhibits higher-than-average underactivation, whereas the occipital cortex contains subregions with lower-than-average as well as higher-than-average decreases in activation. Slice positions are given in stereotaxic Talairach coordinates.

183

the medial prefrontal cortex (MPC, i.e., ACC (BA 24 and 32) and medial superior prefrontal cortex (BA 6 and 9)), and predominantly the left posterior parietal lobule (part of BA 40). Quantitative analysis revealed that relative to the control group, in the TPM group, the language areas were significantly underactivated (i.e., fewer voxels with an activation level higher than the threshold of 0.9%) (Table 1, Figs. 1A and 2A). Results from the neuropsychological assessment and fMRI analysis are illustrated in Fig. 2B. A separate analysis of the female-only subsets of the two groups revealed similar results. In the TPM group, there was significant underactivation throughout the whole brain; however, to elucidate which brain regions are more sensitive to TPM, we calculated which regions had a lower-than-average or a higher-thanaverage underactivation level. This analysis revealed that the language areas (IPC and MPC) of the TPM group uniformly displayed a larger-than-average decrease in activation, whereas the occipital cortex had subregions with lower-than-average as well as higher-than-average decreases in activation (Fig. 1B). 4. Discussion The study described here was performed to expand our understanding of TPM-induced language impairment through fMRI investigation of the underlying cortical activation pattern changes. The covert word generation paradigm used reveals significant underactivation in the language areas of patients with epilepsy on TPM therapy, which is in accordance with the type of neuropsychologically assessed cognitive language impairment (i.e., dysphasia). It has often been reported that most cognitive adverse events emerge particularly at high doses, and increase in a dose-related fashion [15]. However, this TPM group displayed cognitive language impairment on a relatively slow titration schedule and at a relatively low final dose level, which is in accordance with previous results [16]. These observations indicate that TPM induces a local dysfunction of the left prefrontal cortex. The current findings emphasize that not only its efficacy profile, but also its tolerability profile should be thoroughly considered before prescription of TPM. Especially because TPM has demonstrated its potential utility in neurological disorders other than epilepsy [2], the importance of its tolerability profile should be acknowledged in the treatment of these other disorders as well. An explanation for the generally lower activation level for the TPM group throughout the traditional expressive language areas is that TPM likely has an effect on the entire cerebrum. However, the fact that the prefrontal languagerelated areas displayed relatively stronger decreases in activation indicates that these regions have a higher sensitivity to TPM-induced activation-modifying effects. TPM has multiple mechanisms of action that have been hypothesized to contribute to its seizure control effects, some of which

184

Brief Communication / Epilepsy & Behavior 9 (2006) 181–185 4

A

10

SD 0

TPM group

2SD 0

Control group 3

Number of voxels

10

2

10

1

10

0

10

0

0.5

1

1.5

2

Activation level (%) B 3500 TPM group

Number of voxels

3000

Control group

2500 2000 1500 1000 500 0 0

1

2 3 4 5 Neuropsychological language score

6

Fig. 2. (A) Distribution of voxels within the language areas as a function of activation level. The black curves represent the TPM group, and the gray curves, the control group. Solid lines represent mean values, and dotted lines, 95% confidence intervals. The dashed vertical lines indicate the global mean activation levels (averaged over all patients) plus one (SD0) or two (2SD0) standard deviations for brain regions other than language areas, respectively. The global mean plus 2SD0 equals the threshold value of 0.9%, which is used for quantitative analysis. (B) Number of suprathresholded voxels within the language areas as a function of neuropsychological language score. A significant difference between the two groups is observed both for the neuropsychological language score and for the number of voxels. Error bars represent the standard deviation.

may also be involved in the induction of cognitive impairment, including sodium and calcium channel blockade, c-aminobutyric acid (GABA) potentiation, and glutamate receptor antagonism [17]. Preclinical studies link disruptions in the neurotransmission system of GABA with disturbances in the frontal lobe [18]. Because it has been shown that daily TPM therapy increases brain GABA concentrations [19], a TPM-induced increase in the inhibiting effect of the total GABA pool may be responsible for the cognitive language problems, as these problems originate in the frontal lobe. However, not all GABAergic antiepi-

leptic drugs are associated with cognitive impairment. For example, gabapentin, which also is known to increase brain GABA concentrations [19], is associated with minimal cognitive effects [7]. This indicates that TPM-induced cognitive language impairment is possibly caused by the complex interaction of all its mechanisms of action. Because the two groups in this study both have epilepsy, we can only describe the effect of TPM therapy on patients with epilepsy. To study the effect of TPM alone, or the interaction effect between epilepsy and TPM, a control group without epilepsy, for example, patients with migraine, would

Brief Communication / Epilepsy & Behavior 9 (2006) 181–185

be required. However, at present, it is difficult to recruit controls without epilepsy on TPM because of their limited availability. Moreover, the doses of TPM used in migraine are much lower than those used in epilepsy. The current study has some limitations that restrict generalization of TPM-induced brain function abnormalities within the cerebral language system. Because of the cross-sectional design, lack of randomization, lack of interpretation of the tests by blinded investigators, and limited number of patients in this study, the observed effect cannot be unambiguously attributed to TPM alone. As the separate female-only analysis yielded similar results, we argue for the exclusion of a gender effect. Furthermore, there is an obvious correspondence between type and severity of cognitive (i.e., language) impairment and the areas observed to have decreased activation (prefrontal cortex). Hence TPM seems to play an important role in the origin of the observed underactivation. Another shortcoming is the lack of neuropsychological assessment of the patients before TPM administration. However, all patients were living independently in society and were at normal function levels (without aphasic disorders), as evaluated by a trained neurologist. Therefore, additional larger clinical studies may help provide further support for cognitive language impairment induced by TPM therapy. To elucidate left– right hemispheric differential effects, it would be valuable to include further fMRI examinations aimed at revealing nondominant hemispheric activation (e.g., [20]). To provide solid scientific support for our findings, ideally a randomized, double-blind, controlled study design is needed, preferably in patients randomized to TPM monotherapy. In conclusion, by revealing brain function abnormalities within the language system and determining which regions are involved, the present findings help to visualize cognitive language impairment induced by TPM. References [1] Aldenkamp AP, De Krom M, Reijs R. Newer antiepileptic drugs and cognitive issues. Epilepsia 2003;44(Suppl. 4):21–9. [2] Brandes JL, Saper JR, Diamond M, et al. Topiramate for migraine prevention: a randomized controlled trial. JAMA 2004;291:965–73. [3] Faught E, Wilder BJ, Ramsay RE, et al. For the Topiramate YD Study Group. Topiramate placebo-controlled dose-ranging trial in refractory partial epilepsy using 200-, 400-, and 600-mg daily dosages. Neurology 1996;46:1684–90.

185

[4] Ketter TA, Post RM, Theodore WH. Positive and negative psychiatric effects of antiepileptic drugs in patients with seizure disorders. Neurology 1999;53:S53–67. [5] Tatum IV WO, French JA, Faught E, et al. Postmarketing experience with topiramate and cognition. Epilepsia 2001;42:1134–40. [6] Bootsma HP, Coolen F, Aldenkamp AP, et al. Topiramate in clinical practice: long-term experience in patients with refractory epilepsy referred to a tertiary epilepsy center. Epilepsy Behav 2004;5:380–7. [7] Martin R, Kuzniecky R, Ho S, et al. Cognitive effects of topiramate, gabapentin, and lamotrigine in healthy young adults. Neurology 1999;52:321–7. [8] Thompson PJ, Baxendale SA, Duncan JS, Sander JW. Effects of topiramate on cognitive function. J Neurol Neurosurg Psychiatry 2000;69:636–41. [9] Kockelmann E, Elger CE, Helmstaedter C. Cognitive profile of topiramate as compared with lamotrigine in epilepsy patients on antiepileptic drug polytherapy: relationships to blood serum levels and comedication. Epilepsy Behav 2004;5:716–21. [10] Fritz N, Glogau S, Hoffmann J, Rademacher M, Elger CE, Helmstaedter C. Efficacy and cognitive side effects of tiagabine and topiramate in patients with epilepsy. Epilepsy Behav 2005;6:373–81. [11] Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9:97–113. [12] Deelman BG, Liebrand WB, Koning-Haanstra M, van den Burg W. Measurements of aphasic disorders: a brief description of the SANbattery. Gerontologie 1980;11:17–21. [13] McCarthy G, Blamire AM, Rothman DL, Gruetter R, Shulman RG. Echo-planar magnetic resonance imaging studies of frontal cortex activation during word generation in humans. Proc Natl Acad Sci USA 1993;90:4952–6. [14] Ojemann G, Ojemann J, Lettich E, Berger M. Cortical language localization in left, dominant hemisphere: an electrical stimulation mapping investigation in 117 patients. J Neurosurg 1989;71:316–26. [15] Biton V, Edwards KR, Montouris GD, Sackellares JC, Harden CL, Kamin M. Topiramate titration and tolerability. Ann Pharmacother 2001;35:173–9. [16] Aldenkamp AP, Baker G, Mulder OG, et al. A multicenter, randomized clinical study to evaluate the effect on cognitive function of topiramate compared with valproate as add-on therapy to carbamazepine in patients with partial-onset seizures. Epilepsia 2000;41:1167–78. [17] Shank RP, Gardocki JF, Streeter AJ, Maryanoff BE. An overview of the preclinical aspects of topiramate: pharmacology, pharmacokinetics, and mechanism of action. Epilepsia 2000;41(Suppl. 1):S3–9. [18] Petty F. GABA and mood disorders: a brief review and hypothesis. J Affect Disord 1995;34:275–81. [19] Kuzniecky R, Ho S, Pan J, et al. Modulation of cerebral GABA by topiramate, lamotrigine, and gabapentin in healthy adults. Neurology 2002;58:368–72. [20] Staudt M, Lidzba K, Grodd W, Wildgruber D, Erb M, KragelohMann I. Right-hemispheric organization of language following early left-sided brain lesions: functional MRI topography. NeuroImage 2002;16:954–67. [21] Lammers MW, Hekster YA, Keyser A, Meinardi H, Renier WO, van Lier H. Monotherapy or polytherapy for epilepsy revisited: a quantitative assessment. Epilepsia 1995;36:440–6.