Epilepsy Research (2014) 108, 861—871
journal homepage: www.elsevier.com/locate/epilepsyres
Default mode network hypometabolism in epileptic encephalopathies with CSWS Noémie Ligot a, Frédérique Archambaud b,c,d, Nicola Trotta a, Serge Goldman a, Patrick Van Bogaert a, Catherine Chiron b,d,e, Xavier De Tiège a,∗ a
Laboratoire de Cartographie fonctionnelle du Cerveau (LCFC), UNI — ULB Neuroscience Institute, Université libre de Bruxelles (ULB), Brussels, Belgium b Inserm U1129, Paris, France c Université Paris Sud, France d CEA, I2BM, Service Hospitalier Frédéric Joliot, Orsay, France e Université Paris Descartes, France Received 4 December 2013; received in revised form 12 February 2014; accepted 16 March 2014 Available online 27 March 2014
KEYWORDS Continuous spike-wave during sleep; Positron emission tomography; Default mode network; Thalamus; Age-related changes; Cerebral glucose metabolism
Summary Previous studies investigating cerebral metabolic changes associated with continuous spike-waves during sleep (CSWS) compared the metabolism of children with CSWS with that of healthy adults, precluding any assessment in brain areas showing physiologic age-related metabolic changes. Here, we investigated the metabolic and connectivity changes characterizing the acute phase of CSWS activity by comparing awake brain metabolism of children with CSWS with that of pediatric pseudo-controls. Positron emission tomography using [18F]-fluorodeoxyglucose (FDG-PET) was performed in 17 awake children with cryptogenic CSWS (5 girls, age: 5—11 years). Voxel-based analyses identified significant metabolic changes in CSWS patients compared with 18 pediatric pseudo-controls (12 girls, age: 6—11 years, non-CSWS focal cryptogenic epilepsy with normal FDG-PET). CSWSinduced changes in the contribution of brain areas displaying metabolic changes to the level of metabolic activity in other brain areas were investigated using pathophysiological interaction. Hypermetabolism in perisylvian regions bilaterally and hypometabolism in lateral and mesial prefrontal cortex, precuneus, posterior cingulate cortex and parahippocampal gyri characterized the acute phase of CSWS (p < 0.05 FWE). No change in thalamic metabolism was disclosed. Altered functional connectivity was found between hyper- and hypometabolic regions in CSWS patients compared with pediatric pseudo-controls.
∗ Corresponding author at: Laboratoire de Cartographie fonctionnelle du Cerveau (LCFC), UNI — ULB Neuroscience Institute, Université libre de Bruxelles (ULB), 808 Lennik Street, Brussels, Belgium. Tel.: +32 2 555 89 62; fax: +32 2 555 47 01. E-mail address: [email protected] (X. De Tiège).
http://dx.doi.org/10.1016/j.eplepsyres.2014.03.014 0920-1211/© 2014 Elsevier B.V. All rights reserved.
862
N. Ligot et al. This study demonstrates hypometabolism in key nodes of the default mode network (DMN) in awake patients with CSWS, in relation with a possible phenomenon of sustained remote inhibition from the epileptic foci. This hypometabolism might account for some of the acquired cognitive or behavioral features of CSWS epileptic encephalopathies. This study failed to find any evidence of thalamic metabolic changes, which supports the primary involvement of the cortex in CSWS genesis. © 2014 Elsevier B.V. All rights reserved.
Introduction Epileptic encephalopathies are conditions in which the epileptic activity itself contributes to severe cognitive or behavioral impairments above and beyond what might be expected from the underlying cause of epilepsy (Berg et al., 2010). The epileptic encephalopathies with continuous spike-waves during sleep (CSWS) are considered as a prototype of these epileptic conditions (Holmes and LenckSantini, 2006; Nabbout and Dulac, 2003). They are indeed age-related conditions in which epileptic EEG activity occurring almost exclusively during sleep and for prolonged periods of time leads to heterogeneous acquired neuropsychological and behavioral disturbances during the awake state (for reviews, see Van Bogaert, 2013). Previous positron emission tomography with [18F]fluorodeoxyglucose (FDG-PET) studies performed during the awake state in patients with CSWS have demonstrated long lasting changes in brain function characterized by the association of focal increase in metabolism at the onset(s) or propagation(s) sites of CSWS activity and decreases in neuronal activity in distant associative cortical areas not primarily involved in CSWS activity (De Tiège et al., 2004, 2006, 2008, 2013). One pathophysiological model proposed to explain these findings is based on the ‘‘surrounding and remote inhibition theory’’, which suggests the existence of inhibition of neurons that surround or are remotely connected to the hypermetabolic epileptic focus via corticocortical or polysynaptic pathways (for a review, see De Tiège et al., 2009). These data also incite to attribute the neurological regression not only to neuronal impairment at the epileptic focus site, but also to epilepsy-induced neurophysiological changes in distant and connected brain areas (De Tiège et al., 2004, 2006, 2008, 2013). Although previous FDG-PET studies contributed to the understanding of epileptic encephalopathies with CSWS pathophysiology, the interpretation of their results was limited by the use of a control group of young healthy adults for the voxel-based statistical assessment of metabolic and connectivity changes (De Tiège et al., 2004, 2006, 2008, 2013). This limitation was due to obvious ethical constraints, which preclude FDG-PET data acquisition in healthy children. The comparison of FDG-PET data obtained in children with CSWS with that of an adult control group indeed impeded any assessment in brain areas with physiologic agerelated metabolic changes such as the thalamus, cingulate or mesial cortical areas (De Tiège et al., 2004, 2008, 2013; Van Bogaert et al., 1998). However, several experimental evidences coming from functional magnetic resonance imaging combined with EEG (EEG-fMRI) suggest the existence of transient changes in neuronal activity within those
brain regions during CSWS activity (De Tiège et al., 2007; Siniatchkin et al., 2010). In addition, the thalamus is considered as playing a key role in the pathophysiology of CSWS considering the neurophysiological activity changes within the thalamocortical network associated with slow sleep (Maquet et al., 1995; Siniatchkin et al., 2010; Steriade, 2005, 2006) and the potential association between perinatal thalamic ischemic lesions and CSWS activity (Guzzetta et al., 2005; Sanchez Fernandez et al., 2012). Therefore, to fully characterize the pathophysiology of CSWS syndromes and their associated diurnal cognitive and behavioral regression, it appeared essential to determine if CSWS activity is associated with sustained changes in metabolic activity within those regions. Here, we reanalyzed awake FDG-PET data obtained in a large group of children with CSWS by comparing them with a pediatric pseudo-control group composed of children with non-CSWS cryptogenic refractory focal epilepsy (Archambaud et al., 2013) in order to investigate the metabolic and connectivity changes characterizing the acute phase of CSWS. The main purpose of this comparison was to identify the metabolic abnormalities associated with CSWS activity.
Methods Patients and control subjects Among the 39 patients with an epileptic encephalopathy with CSWS studied by FDG-PET at the ULB-Hôpital Erasme (Brussels, Belgium) between November 1999 and April 2010, a group of 17 children (12 boys and 5 girls, mean age: 7.4 years, age range: 5—11 years) was retrospectively selected based on the following inclusion criteria: (1) normal structural cerebral MRI, (2) FDG-PET performed at the acute phase of CSWS, and (3) significant local hypermetabolism on FDG-PET based on voxel-based analysis of individual FDG-PET data (see below). Only patients with local hypermetabolism were included in this study as this metabolic abnormality represents the hallmark of the acute phase of CSWS and its absence probably reflects an evolution toward the recovery phase of CSWS (De Tiège et al., 2004, 2008). Furthermore, only patients with normal MRI were selected for this study in order to avoid any confounding effects of brain lesions on regional brain metabolism. Sixteen patients (out of 39) were therefore rejected from this study because of symptomatic epilepsy. Two other patients younger than 5 years old were not included in order to avoid any FDG-PET normalization artifact as we used the MNI (adult) template for normalization (see below). Four patients were rejected because they only had regional
Epileptic encephalopathies with CSWS Table 1
863
Summary of clinical, treatment and EEG data.
Patient
Sex/agea / syndrome
AED (mg/kg/day)
Neuropsychology
EEG during PET 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
M/5/LKS M/5/LKS F/6/LKS M/8/LKS M/8/LKS M/9/LKS M/10/LKS F/8/GMD F/8/GMD M/8/GMD M/10/GMD M/5/FS M/5/NDHS M/6/NDHS M/11/ARE
None None CB (1) VPA (30), LEV (50) None HC (0.8), CZP (0.1) LTG (5), OXC (30) VPA (20) LTG (3), LEV (45) LTG (5) LEV (60) VPA (35), TPM (3.5) VPA (33), CBZ (22) TPM (2), LTG (10) LTG (11), DPH (4)
Frequent IEDs (C3, C4) Rare IEDs (C3, C4) Frequent IEDs (C3, T4) Rare IEDs (T3, T4) Frequent IEDs (T3, T4) Rare IEDs (T3) Frequent IEDs (C3, C4) Episodic IEDs (C3, C4) Episodic IEDs (C3, C4) Frequent IEDs (C3, C4, Fp1) Rare IEDs (C3) Episodic IEDs (T3, T4) Episodic IEDs (C4) Frequent IEDs (C4) Frequent IEDs (C4)
16
F/6/ARE
VPA (30)
17
F/8/OS
VPA (27), LTG (2)
pIQ (Leiter): 93 pIQ (WISC): 112 pIQ (WPPSI): 72 pIQ (Leiter): 75 pIQ (SON-R): 109 IQ: NA IQ: NA pIQ (Leiter):50 pIQ (WISC): 50 pIQ (WISC): 50 pIQ (Leiter): 53 IQ (WPPSI): 89 IQ (McCarthy): 78 pIQ (Leiter): 70 Disinhibition, IQ (WISC-III): 77 Agitation, impulsivity, IQ (WPPSI): 77 pIQ (Leiter): 63
Frequent IEDs (C4) Frequent IEDs (C3, C4)
AED: antiepileptic drugs, M: male, F: female, LKS: Landau—Kleffner syndrome, FS: frontal syndrome, NDHS: non-dominant hemispheric syndrome, OS: opercular syndrome, GMD: global mental deterioration, CB: clobazam, VPA: valproate, LEV: levetiracetam, HC: hydrocortisone, CZP: clonazepam, LTG: lamotrigine, OXC: oxycarbamazepine, TPM: topiramate, CBZ: carbamazepine, DPH: diphantoine, pIQ: performance IQ, NA: not available at the time of FDG-PET, IEDs: interictal epileptic discharges. a Years.
hypometabolism. All the FDG-PET data used in this study came from CSWS patients that have been included in previous studies from our group on FDG-PET or therapeutic trials in CSWS, patients’ clinical details can be found in the respective papers (De Tiège et al., 2004, patients 1, 3, 9, 11, 15—18; De Tiège et al., 2008, patients 3—9; Buzatu et al., 2009, patients 3, 5, 9, 10, 11, 14; De Tiège et al., 2013, patients 1—4). All patients underwent a clinical evaluation that included neurological and neuropsychological examinations, awake and sleep video-EEG monitoring performed in the last 24 h preceding FDG-PET, structural cerebral MRI, and FDG-PET. The tests used for the neuropsychological evaluations were adapted to patients’ age and collaboration and are detailed in previous papers from our groups (De Tiège et al., 2008, 2013). According to the neurological and neuropsychological evaluations, six syndromes were considered for patients’ classification purpose: Landau—Kleffner syndrome (7 patients) (Landau and Kleffner, 1957), atypical rolandic epilepsy (2 patients) (Tovia et al., 2011), frontal syndrome (1 patient) (Roulet Perez et al., 1993; Veggiotti et al., 2001), global mental retardation (4 patients) (Patry et al., 1971), non-dominant hemisphere syndrome (2 patients) (Maquet et al., 1995), and opercular syndrome (1 patient) (Shafrir and Prensky, 1995; Tachikawa et al., 2001). Table 1 summarizes the CSWS patients’ clinical details. The ULB-Hôpital Erasme Ethics Committee gave approval for conducting this retrospective study. The ULB-Hôpital Erasme Ethics Committee waived the requirement for obtaining parents’ written informed consent in the context of this retrospective study.
A group of 18 pseudo-control children (pediatric pseudocontrols, 12 girls and 6 boys, mean age: 9.6 years, age range: 6—11 years) was selected in a pediatric pseudocontrol group previously described by the Service Hospitalier Frédéric Joliot (Archambaud et al., 2013) (Orsay, France) to match as close as possible age range and mean of the group of CSWS patients. This pediatric pseudo-control group consisted of 18 children selected among 24 patients with non-CSWS refractory focal epilepsy with negative findings on MRI and FDG-PET. The use of these FDG-PET data was approved by the institutional ethical standards committee on human experimentation, and written informed consent was obtained from all guardians of patients participating in the study. A group of 26 healthy adult volunteers (adult controls, 16 females and 10 males, mean age 28 years, 18—42 years) studied with FDG-PET at the ULB-Hôpital Erasme was used as an adult control group for FDG-PET data analyses. This population of adult controls has been used in previous FDGPET studies from our group (De Tiège et al., 2004, 2006, 2008, 2013). The ULB-Hôpital Erasme Ethics Committee gave approval for the PET investigations in adult control subjects. Written informed consent was obtained from all adult control subjects.
PET data acquisitions in CSWS patients and adult controls PET scans were obtained using a CTI-Siemens ECAT 962 (HR+, Siemens Medical Solutions, Munich, Germany) tomograph
864 installed at the ULB-Hôpital Erasme (Brussels, Belgium). All participants were awake at the time of FDG injection, placed with the eye-closed in a quiet dark room, fasted for at least 4 h and received an intravenous bolus injection of 2—3 mCi (74—111 MBq) of FDG before PET acquisition in 3D mode. They were not sedated for PET data acquisition, which started 30 min after FDG injection. The current antiepileptic treatment was unchanged. All patients were free of clinical seizures for at least 72 h. EEG was monitored during the PET procedure.
PET data acquisitions in pediatric pseudo-controls Patients were examined using ECAT 962 (HR+, Siemens Medical Solutions, Munich, Germany) installed at the Institute of Biomedical Imaging (I2BM, CEA, Service Hospitalier F. Joliot, Orsay, France). The patients were investigated in a fasting and resting state, in a quiet, dimly lit environment. The last seizure had occurred more than 6 h before PET examination except for 3 patients (between 1 and 6 h in 3). All patients received an intravenous bolus injection of 3.7 MBq/kg (maximum 180 MBq) of FDG while lying in the scanner. Patients were not sedated for PET data acquisition, which started 30 min after FDG injection. EEG was not monitored during the procedure. The fact that patients had their last seizure more than 6 h before FDG-PET (except for 3 patients) and that EEG was not monitored during FDG incorporation is not an issue since all patients had negative FDG-PET findings.
PET data analyses All FDG-PET data were analyzed using SPM8 software (http://www.fil.ion.ucl.ac.uk/spm, Wellcome Department of Imaging Neuroscience, London, UK) implemented in Matlab 7.1 (Mathworks, Sherborn, MA). The PET images were first spatially normalized into the MNI space and subsequently smoothed using a 16-mm full width at half-maximum isotropic kernel. Global activity normalization was performed using proportional scaling. Then, subtractive analyses were conducted both at the individual and group levels using the same methodology that was previously described (De Tiège et al., 2004, 2008). Briefly, individual-level analyses first compared PET data of each CSWS patient with those of the pediatric pseudo-control group. For these analyses, a design matrix containing PET data of each CSWS patient and those of the pediatric pseudo-controls taken as a group was built. Then, t-contrasts identified brain regions where glucose metabolism was significantly lower or higher in each CSWS patient than in pediatric pseudo-controls. Group-level analyses were subsequently performed to compare (1) PET data of CSWS patients taken as a group with those of pediatric pseudo-controls, (2) PET data of CSWS patients taken as a group with those of adult controls, and (3) PET data of pediatric pseudo-controls taken as a group with those of adult controls. For these analyses, a design matrix containing PET data of CSWS patients, pediatric pseudocontrols and adult controls taken as separated groups was built. Using this design matrix, t-contrasts first identified brain regions where glucose metabolism was significantly lower or higher in CSWS patients than in the pediatric
N. Ligot et al. pseudo-control group. For this analysis, age and sex were used as covariates of no interest to avoid the potential confounding effects of these variables on the results. Then, t-contrasts identified brain regions where glucose metabolism was significantly lower or higher in CSWS patients or pediatric pseudo-controls than in the adult control group. Subsequently, conjunction analyses (Friston et al., 1999) between CSWS patients versus adult controls t-contrasts and pediatric pseudo-controls versus adult controls t-contrasts were performed to identify metabolic changes characterizing both pediatric populations. Finally, we searched for pathophysiological interactions (PathoPI) that we previously defined as disease-related changes in the contribution of a brain area to the level of metabolic activity in another brain area in the group of CSWS patients compared with the pediatric pseudocontrol group (De Tiège et al., 2004, 2008). PathoPI analyses were conducted based on the a priori hypothesis of altered effective connectivity between hyper- and hypometabolic brain regions (De Tiège et al., 2004). In practice, the peak voxel values of the most significant hypo and hypermetabolic areas found in the group-level analysis comparing CSWS patients with pediatric pseudo-controls were used as a covariate of interest centered around condition means and interacting with each condition. These covariates of interest were introduced in separate design matrices that included CSWS patients’ scans and those of the pediatric pseudo-controls taken as separated groups. PathoPI analyses identified throughout the brain the regions that showed CSWS-related differences in modulation with the considered hypo- and hypermetabolic areas in the group of CSWS patients compared with pediatric pseudo-controls. For these analyses, age and sex were used as covariates of no interest to avoid the potential confounding effect of these variables on the results. All results of SPM analyses were considered significant at p
journal homepage: www.elsevier.com/locate/epilepsyres
Default mode network hypometabolism in epileptic encephalopathies with CSWS Noémie Ligot a, Frédérique Archambaud b,c,d, Nicola Trotta a, Serge Goldman a, Patrick Van Bogaert a, Catherine Chiron b,d,e, Xavier De Tiège a,∗ a
Laboratoire de Cartographie fonctionnelle du Cerveau (LCFC), UNI — ULB Neuroscience Institute, Université libre de Bruxelles (ULB), Brussels, Belgium b Inserm U1129, Paris, France c Université Paris Sud, France d CEA, I2BM, Service Hospitalier Frédéric Joliot, Orsay, France e Université Paris Descartes, France Received 4 December 2013; received in revised form 12 February 2014; accepted 16 March 2014 Available online 27 March 2014
KEYWORDS Continuous spike-wave during sleep; Positron emission tomography; Default mode network; Thalamus; Age-related changes; Cerebral glucose metabolism
Summary Previous studies investigating cerebral metabolic changes associated with continuous spike-waves during sleep (CSWS) compared the metabolism of children with CSWS with that of healthy adults, precluding any assessment in brain areas showing physiologic age-related metabolic changes. Here, we investigated the metabolic and connectivity changes characterizing the acute phase of CSWS activity by comparing awake brain metabolism of children with CSWS with that of pediatric pseudo-controls. Positron emission tomography using [18F]-fluorodeoxyglucose (FDG-PET) was performed in 17 awake children with cryptogenic CSWS (5 girls, age: 5—11 years). Voxel-based analyses identified significant metabolic changes in CSWS patients compared with 18 pediatric pseudo-controls (12 girls, age: 6—11 years, non-CSWS focal cryptogenic epilepsy with normal FDG-PET). CSWSinduced changes in the contribution of brain areas displaying metabolic changes to the level of metabolic activity in other brain areas were investigated using pathophysiological interaction. Hypermetabolism in perisylvian regions bilaterally and hypometabolism in lateral and mesial prefrontal cortex, precuneus, posterior cingulate cortex and parahippocampal gyri characterized the acute phase of CSWS (p < 0.05 FWE). No change in thalamic metabolism was disclosed. Altered functional connectivity was found between hyper- and hypometabolic regions in CSWS patients compared with pediatric pseudo-controls.
∗ Corresponding author at: Laboratoire de Cartographie fonctionnelle du Cerveau (LCFC), UNI — ULB Neuroscience Institute, Université libre de Bruxelles (ULB), 808 Lennik Street, Brussels, Belgium. Tel.: +32 2 555 89 62; fax: +32 2 555 47 01. E-mail address: [email protected] (X. De Tiège).
http://dx.doi.org/10.1016/j.eplepsyres.2014.03.014 0920-1211/© 2014 Elsevier B.V. All rights reserved.
862
N. Ligot et al. This study demonstrates hypometabolism in key nodes of the default mode network (DMN) in awake patients with CSWS, in relation with a possible phenomenon of sustained remote inhibition from the epileptic foci. This hypometabolism might account for some of the acquired cognitive or behavioral features of CSWS epileptic encephalopathies. This study failed to find any evidence of thalamic metabolic changes, which supports the primary involvement of the cortex in CSWS genesis. © 2014 Elsevier B.V. All rights reserved.
Introduction Epileptic encephalopathies are conditions in which the epileptic activity itself contributes to severe cognitive or behavioral impairments above and beyond what might be expected from the underlying cause of epilepsy (Berg et al., 2010). The epileptic encephalopathies with continuous spike-waves during sleep (CSWS) are considered as a prototype of these epileptic conditions (Holmes and LenckSantini, 2006; Nabbout and Dulac, 2003). They are indeed age-related conditions in which epileptic EEG activity occurring almost exclusively during sleep and for prolonged periods of time leads to heterogeneous acquired neuropsychological and behavioral disturbances during the awake state (for reviews, see Van Bogaert, 2013). Previous positron emission tomography with [18F]fluorodeoxyglucose (FDG-PET) studies performed during the awake state in patients with CSWS have demonstrated long lasting changes in brain function characterized by the association of focal increase in metabolism at the onset(s) or propagation(s) sites of CSWS activity and decreases in neuronal activity in distant associative cortical areas not primarily involved in CSWS activity (De Tiège et al., 2004, 2006, 2008, 2013). One pathophysiological model proposed to explain these findings is based on the ‘‘surrounding and remote inhibition theory’’, which suggests the existence of inhibition of neurons that surround or are remotely connected to the hypermetabolic epileptic focus via corticocortical or polysynaptic pathways (for a review, see De Tiège et al., 2009). These data also incite to attribute the neurological regression not only to neuronal impairment at the epileptic focus site, but also to epilepsy-induced neurophysiological changes in distant and connected brain areas (De Tiège et al., 2004, 2006, 2008, 2013). Although previous FDG-PET studies contributed to the understanding of epileptic encephalopathies with CSWS pathophysiology, the interpretation of their results was limited by the use of a control group of young healthy adults for the voxel-based statistical assessment of metabolic and connectivity changes (De Tiège et al., 2004, 2006, 2008, 2013). This limitation was due to obvious ethical constraints, which preclude FDG-PET data acquisition in healthy children. The comparison of FDG-PET data obtained in children with CSWS with that of an adult control group indeed impeded any assessment in brain areas with physiologic agerelated metabolic changes such as the thalamus, cingulate or mesial cortical areas (De Tiège et al., 2004, 2008, 2013; Van Bogaert et al., 1998). However, several experimental evidences coming from functional magnetic resonance imaging combined with EEG (EEG-fMRI) suggest the existence of transient changes in neuronal activity within those
brain regions during CSWS activity (De Tiège et al., 2007; Siniatchkin et al., 2010). In addition, the thalamus is considered as playing a key role in the pathophysiology of CSWS considering the neurophysiological activity changes within the thalamocortical network associated with slow sleep (Maquet et al., 1995; Siniatchkin et al., 2010; Steriade, 2005, 2006) and the potential association between perinatal thalamic ischemic lesions and CSWS activity (Guzzetta et al., 2005; Sanchez Fernandez et al., 2012). Therefore, to fully characterize the pathophysiology of CSWS syndromes and their associated diurnal cognitive and behavioral regression, it appeared essential to determine if CSWS activity is associated with sustained changes in metabolic activity within those regions. Here, we reanalyzed awake FDG-PET data obtained in a large group of children with CSWS by comparing them with a pediatric pseudo-control group composed of children with non-CSWS cryptogenic refractory focal epilepsy (Archambaud et al., 2013) in order to investigate the metabolic and connectivity changes characterizing the acute phase of CSWS. The main purpose of this comparison was to identify the metabolic abnormalities associated with CSWS activity.
Methods Patients and control subjects Among the 39 patients with an epileptic encephalopathy with CSWS studied by FDG-PET at the ULB-Hôpital Erasme (Brussels, Belgium) between November 1999 and April 2010, a group of 17 children (12 boys and 5 girls, mean age: 7.4 years, age range: 5—11 years) was retrospectively selected based on the following inclusion criteria: (1) normal structural cerebral MRI, (2) FDG-PET performed at the acute phase of CSWS, and (3) significant local hypermetabolism on FDG-PET based on voxel-based analysis of individual FDG-PET data (see below). Only patients with local hypermetabolism were included in this study as this metabolic abnormality represents the hallmark of the acute phase of CSWS and its absence probably reflects an evolution toward the recovery phase of CSWS (De Tiège et al., 2004, 2008). Furthermore, only patients with normal MRI were selected for this study in order to avoid any confounding effects of brain lesions on regional brain metabolism. Sixteen patients (out of 39) were therefore rejected from this study because of symptomatic epilepsy. Two other patients younger than 5 years old were not included in order to avoid any FDG-PET normalization artifact as we used the MNI (adult) template for normalization (see below). Four patients were rejected because they only had regional
Epileptic encephalopathies with CSWS Table 1
863
Summary of clinical, treatment and EEG data.
Patient
Sex/agea / syndrome
AED (mg/kg/day)
Neuropsychology
EEG during PET 1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
M/5/LKS M/5/LKS F/6/LKS M/8/LKS M/8/LKS M/9/LKS M/10/LKS F/8/GMD F/8/GMD M/8/GMD M/10/GMD M/5/FS M/5/NDHS M/6/NDHS M/11/ARE
None None CB (1) VPA (30), LEV (50) None HC (0.8), CZP (0.1) LTG (5), OXC (30) VPA (20) LTG (3), LEV (45) LTG (5) LEV (60) VPA (35), TPM (3.5) VPA (33), CBZ (22) TPM (2), LTG (10) LTG (11), DPH (4)
Frequent IEDs (C3, C4) Rare IEDs (C3, C4) Frequent IEDs (C3, T4) Rare IEDs (T3, T4) Frequent IEDs (T3, T4) Rare IEDs (T3) Frequent IEDs (C3, C4) Episodic IEDs (C3, C4) Episodic IEDs (C3, C4) Frequent IEDs (C3, C4, Fp1) Rare IEDs (C3) Episodic IEDs (T3, T4) Episodic IEDs (C4) Frequent IEDs (C4) Frequent IEDs (C4)
16
F/6/ARE
VPA (30)
17
F/8/OS
VPA (27), LTG (2)
pIQ (Leiter): 93 pIQ (WISC): 112 pIQ (WPPSI): 72 pIQ (Leiter): 75 pIQ (SON-R): 109 IQ: NA IQ: NA pIQ (Leiter):50 pIQ (WISC): 50 pIQ (WISC): 50 pIQ (Leiter): 53 IQ (WPPSI): 89 IQ (McCarthy): 78 pIQ (Leiter): 70 Disinhibition, IQ (WISC-III): 77 Agitation, impulsivity, IQ (WPPSI): 77 pIQ (Leiter): 63
Frequent IEDs (C4) Frequent IEDs (C3, C4)
AED: antiepileptic drugs, M: male, F: female, LKS: Landau—Kleffner syndrome, FS: frontal syndrome, NDHS: non-dominant hemispheric syndrome, OS: opercular syndrome, GMD: global mental deterioration, CB: clobazam, VPA: valproate, LEV: levetiracetam, HC: hydrocortisone, CZP: clonazepam, LTG: lamotrigine, OXC: oxycarbamazepine, TPM: topiramate, CBZ: carbamazepine, DPH: diphantoine, pIQ: performance IQ, NA: not available at the time of FDG-PET, IEDs: interictal epileptic discharges. a Years.
hypometabolism. All the FDG-PET data used in this study came from CSWS patients that have been included in previous studies from our group on FDG-PET or therapeutic trials in CSWS, patients’ clinical details can be found in the respective papers (De Tiège et al., 2004, patients 1, 3, 9, 11, 15—18; De Tiège et al., 2008, patients 3—9; Buzatu et al., 2009, patients 3, 5, 9, 10, 11, 14; De Tiège et al., 2013, patients 1—4). All patients underwent a clinical evaluation that included neurological and neuropsychological examinations, awake and sleep video-EEG monitoring performed in the last 24 h preceding FDG-PET, structural cerebral MRI, and FDG-PET. The tests used for the neuropsychological evaluations were adapted to patients’ age and collaboration and are detailed in previous papers from our groups (De Tiège et al., 2008, 2013). According to the neurological and neuropsychological evaluations, six syndromes were considered for patients’ classification purpose: Landau—Kleffner syndrome (7 patients) (Landau and Kleffner, 1957), atypical rolandic epilepsy (2 patients) (Tovia et al., 2011), frontal syndrome (1 patient) (Roulet Perez et al., 1993; Veggiotti et al., 2001), global mental retardation (4 patients) (Patry et al., 1971), non-dominant hemisphere syndrome (2 patients) (Maquet et al., 1995), and opercular syndrome (1 patient) (Shafrir and Prensky, 1995; Tachikawa et al., 2001). Table 1 summarizes the CSWS patients’ clinical details. The ULB-Hôpital Erasme Ethics Committee gave approval for conducting this retrospective study. The ULB-Hôpital Erasme Ethics Committee waived the requirement for obtaining parents’ written informed consent in the context of this retrospective study.
A group of 18 pseudo-control children (pediatric pseudocontrols, 12 girls and 6 boys, mean age: 9.6 years, age range: 6—11 years) was selected in a pediatric pseudocontrol group previously described by the Service Hospitalier Frédéric Joliot (Archambaud et al., 2013) (Orsay, France) to match as close as possible age range and mean of the group of CSWS patients. This pediatric pseudo-control group consisted of 18 children selected among 24 patients with non-CSWS refractory focal epilepsy with negative findings on MRI and FDG-PET. The use of these FDG-PET data was approved by the institutional ethical standards committee on human experimentation, and written informed consent was obtained from all guardians of patients participating in the study. A group of 26 healthy adult volunteers (adult controls, 16 females and 10 males, mean age 28 years, 18—42 years) studied with FDG-PET at the ULB-Hôpital Erasme was used as an adult control group for FDG-PET data analyses. This population of adult controls has been used in previous FDGPET studies from our group (De Tiège et al., 2004, 2006, 2008, 2013). The ULB-Hôpital Erasme Ethics Committee gave approval for the PET investigations in adult control subjects. Written informed consent was obtained from all adult control subjects.
PET data acquisitions in CSWS patients and adult controls PET scans were obtained using a CTI-Siemens ECAT 962 (HR+, Siemens Medical Solutions, Munich, Germany) tomograph
864 installed at the ULB-Hôpital Erasme (Brussels, Belgium). All participants were awake at the time of FDG injection, placed with the eye-closed in a quiet dark room, fasted for at least 4 h and received an intravenous bolus injection of 2—3 mCi (74—111 MBq) of FDG before PET acquisition in 3D mode. They were not sedated for PET data acquisition, which started 30 min after FDG injection. The current antiepileptic treatment was unchanged. All patients were free of clinical seizures for at least 72 h. EEG was monitored during the PET procedure.
PET data acquisitions in pediatric pseudo-controls Patients were examined using ECAT 962 (HR+, Siemens Medical Solutions, Munich, Germany) installed at the Institute of Biomedical Imaging (I2BM, CEA, Service Hospitalier F. Joliot, Orsay, France). The patients were investigated in a fasting and resting state, in a quiet, dimly lit environment. The last seizure had occurred more than 6 h before PET examination except for 3 patients (between 1 and 6 h in 3). All patients received an intravenous bolus injection of 3.7 MBq/kg (maximum 180 MBq) of FDG while lying in the scanner. Patients were not sedated for PET data acquisition, which started 30 min after FDG injection. EEG was not monitored during the procedure. The fact that patients had their last seizure more than 6 h before FDG-PET (except for 3 patients) and that EEG was not monitored during FDG incorporation is not an issue since all patients had negative FDG-PET findings.
PET data analyses All FDG-PET data were analyzed using SPM8 software (http://www.fil.ion.ucl.ac.uk/spm, Wellcome Department of Imaging Neuroscience, London, UK) implemented in Matlab 7.1 (Mathworks, Sherborn, MA). The PET images were first spatially normalized into the MNI space and subsequently smoothed using a 16-mm full width at half-maximum isotropic kernel. Global activity normalization was performed using proportional scaling. Then, subtractive analyses were conducted both at the individual and group levels using the same methodology that was previously described (De Tiège et al., 2004, 2008). Briefly, individual-level analyses first compared PET data of each CSWS patient with those of the pediatric pseudo-control group. For these analyses, a design matrix containing PET data of each CSWS patient and those of the pediatric pseudo-controls taken as a group was built. Then, t-contrasts identified brain regions where glucose metabolism was significantly lower or higher in each CSWS patient than in pediatric pseudo-controls. Group-level analyses were subsequently performed to compare (1) PET data of CSWS patients taken as a group with those of pediatric pseudo-controls, (2) PET data of CSWS patients taken as a group with those of adult controls, and (3) PET data of pediatric pseudo-controls taken as a group with those of adult controls. For these analyses, a design matrix containing PET data of CSWS patients, pediatric pseudocontrols and adult controls taken as separated groups was built. Using this design matrix, t-contrasts first identified brain regions where glucose metabolism was significantly lower or higher in CSWS patients than in the pediatric
N. Ligot et al. pseudo-control group. For this analysis, age and sex were used as covariates of no interest to avoid the potential confounding effects of these variables on the results. Then, t-contrasts identified brain regions where glucose metabolism was significantly lower or higher in CSWS patients or pediatric pseudo-controls than in the adult control group. Subsequently, conjunction analyses (Friston et al., 1999) between CSWS patients versus adult controls t-contrasts and pediatric pseudo-controls versus adult controls t-contrasts were performed to identify metabolic changes characterizing both pediatric populations. Finally, we searched for pathophysiological interactions (PathoPI) that we previously defined as disease-related changes in the contribution of a brain area to the level of metabolic activity in another brain area in the group of CSWS patients compared with the pediatric pseudocontrol group (De Tiège et al., 2004, 2008). PathoPI analyses were conducted based on the a priori hypothesis of altered effective connectivity between hyper- and hypometabolic brain regions (De Tiège et al., 2004). In practice, the peak voxel values of the most significant hypo and hypermetabolic areas found in the group-level analysis comparing CSWS patients with pediatric pseudo-controls were used as a covariate of interest centered around condition means and interacting with each condition. These covariates of interest were introduced in separate design matrices that included CSWS patients’ scans and those of the pediatric pseudo-controls taken as separated groups. PathoPI analyses identified throughout the brain the regions that showed CSWS-related differences in modulation with the considered hypo- and hypermetabolic areas in the group of CSWS patients compared with pediatric pseudo-controls. For these analyses, age and sex were used as covariates of no interest to avoid the potential confounding effect of these variables on the results. All results of SPM analyses were considered significant at p
