Epilepsy Research (2014) 108, 953—962
journal homepage: www.elsevier.com/locate/epilepsyres
Drug-resistant focal sleep related epilepsy: Results and predictors of surgical outcome Anna Losurdo a, Paola Proserpio b, Francesco Cardinale b, Francesca Gozzo b, Laura Tassi b, Roberto Mai b, Stefano Francione b, Laura Castana b, Giorgio Lo Russo b, Giuseppe Casaceli b, Ivana Sartori b, Giacomo Della Marca a, Massimo Cossu b, Lino Nobili b,c,∗ a
Institute of Neurology, Catholic University, Policlinico Universitario A. Gemelli, Rome, Italy ‘‘C. Munari’’ Center of Epilepsy Surgery, Niguarda Hospital, Milan, Italy c Institute of Bioimaging and Molecular Physiology, Genoa Unit, National Research Council, Genoa, Italy b
Received 25 July 2013; received in revised form 3 February 2014; accepted 28 February 2014 Available online 12 March 2014
KEYWORDS Epilepsy; Epilepsy surgery; Sleep; Sleep related epilepsy
Summary In this study we report the results of surgery in a large population of patients affected by drug-resistant focal sleep related epilepsy (SRE) and the identified prognostic factors. We conducted a retrospective analysis of a case series of 955 patients operated on for drug-resistant focal epilepsy from 1997 to 2009. Ninety-five patients with focal SRE and a followup of at least 2 years were identified. Presurgical, surgical and histopathological variables were analyzed. Risk of seizures recurrence was assessed by univariate and multivariate analysis. Mean age at epilepsy onset was 5.6 ± 4.9 years. MRI revealed a focal abnormality in 78.9% of cases. Sixty-two percent of patients required a Stereo-EEG investigation. The cortical resection involved the frontal lobe in 61.1% of cases, while in 38.9% an extrafrontal resection was performed. Focal cortical dysplasia (FCD) type II was the most frequent histopathological finding. Mean postoperative follow-up was 82.3 months. Seventy-three patients (76.8%) were in Engel’s class I. At univariate analysis, variables associated with a favorable outcome were: absence of Stereo-EEG investigation; positive MRI; complete removal of the epileptogenic zone (EZ);
∗ Corresponding author at: Centre of Epilepsy Surgery ‘‘C. Munari’’, Center of Sleep Medicine, Niguarda Hospital, Piazza Ospedale Maggiore 2, 20162 Milan, Italy. Tel.: +39 0264447323; fax: +39 0264442874. E-mail address: [email protected] (L. Nobili).
http://dx.doi.org/10.1016/j.eplepsyres.2014.02.016 0920-1211/© 2014 Elsevier B.V. All rights reserved.
954
A. Losurdo et al. presence of FCD type II and FCD type IIb. A diagnosis of FCD type I was associated with postoperative recurrence of seizures. Multivariate analysis identified the complete removal of the EZ and FCD type I as independent predictors of a favorable and unfavorable outcome respectively. SRE can frequently originate outside the frontal lobe and a favorable surgical outcome is achieved in three-fourths of cases independently from the location of the EZ. © 2014 Elsevier B.V. All rights reserved.
Introduction In patients with sleep related epilepsy (SRE), more than 90% of ictal events arise from sleep (Iber et al., 2007). The percentage of patients with epilepsy affected by focal SRE ranges from 7.5 to 45%, depending on the method of case ascertainment (Thomas et al., 2010). Focal SRE is considered a relatively benign clinical condition (D’Alessandro et al., 1983) because seizures occur almost exclusively during nocturnal sleep and in the majority of patients a good seizure control is achieved with pharmacological treatment. Nevertheless, a relatively high percentage of patients, mainly with a focal frontal epilepsy, is drug resistant (Bernasconi et al., 1998; Nobili et al., 2007; Provini et al., 1999). Sleep related seizures are a typical manifestation of Nocturnal Frontal Lobe Epilepsy (NFLE) (Nobili et al., 2007; Oldani et al., 1998; Provini et al., 1999; Scheffer et al., 1995); however, many studies have shown that, among drug resistant patients, sleep related seizures can originate in the temporal lobe (Bernasconi et al., 1998; Nobili et al., 2004; Tao et al., 2010), in the insular lobe (Dobesberger et al., 2008; Kaido et al., 2006; Proserpio et al., 2011b; Ryvlin et al., 2006; Zhang et al., 2008) and in the posterior cortical regions (Proserpio et al., 2011a). Epilepsy surgery in NFLE has shown to be an effective treatment (Nobili et al., 2007); moreover, although limited to single cases or small groups of patients, good results seem to be achieved also in extra-frontal focal nocturnal seizures (Dobesberger et al., 2008; Elsharkawy et al., 2009; Mai et al., 2005; Proserpio et al., 2011a). The aim of our study is to report the results of surgery in a large population of patients affected by drug resistant focal SRE epilepsy and to point out possible presurgical and surgical prognostic factors.
Materials and methods We have evaluated retrospectively a series of 955 patients operated on for drug-resistant focal epilepsy at ‘‘C. Munari’’ Epilepsy Surgery Centre from 1997 to 2009. The selection criteria were: 1) Presence of sleep-related seizures. Patients were considered to be affected by sleep-related seizures if more than 90% of ictal events arose from sleep, referred to questioning of patients and their relatives. This distribution of seizures was confirmed by seizure diaries filled over a period for at least one year, and by subsequent long-term video-EEG recordings conducted both during sleep and wakefulness. 2) Post-operative follow-up period of at least 24 months.
3) Availability of a post-surgical MRI study (necessary for assessing the complete or incomplete removal of the epileptogenic zone; see ‘‘surgery’’). The study was approved by the Ethic Committee of the Niguarda Hospital, Milan.
Presurgical evaluation All patients underwent a presurgical investigation based on: 1. accurate analysis of personal and epileptic history; 2. scalp video-EEG (VEEG) monitoring, including at least one video-polysomnographic recording of nocturnal sleep; 3. MRI studies were performed according to the protocol proposed by Colombo et al. (2009, 2003), and they were customized with employment of appropriate sequences according to main electroclinical information (1.5-tesla ACS-NT unit; Philips Medical Systems, Best, The Netherlands). Intravenous contrast was injected when a neoplasm was suspected. 4. When non-invasive investigations failed to localize the epileptogenic zone (EZ; the brain region considered essential for inducing and maintaining the epileptic seizures), a stereo-electro-encephalography (SEEG) with stereotactically placed intracerebral electrodes was performed (Cossu et al., 2005). The arrangement of electrodes was tailored according to a predefined localization hypothesis based on non-invasive findings.
Surgery Surgery aimed at resection of the EZ, whose identification was based on the anatomo-electro-clinical correlations. On these basis the area to be removed could be limited to a possible discrete lesion detected by MRI or be extended also to extralesional areas. In all the MRI-negative cases the resection area was defined on the basis of a SEEG investigation. Surgical procedures were classified as complete or incomplete resection of the EZ. Resections were considered incomplete when: (1) the EZ was partly spared because it involved eloquent cortex (2) postoperative MRI revealed that surgical resection did not completely match the preoperative plan (3) the indications to surgery were based on data with residual uncertainties as to the actual limits of the EZ. In particular, the presurgical evaluation allowed the identification of a definite anatomical region to be removed; however it was not able to exclude with certainty the
Drug-resistant SRE and surgical outcome
955
Surgical specimens were processed for histopathological and immunohistochemical evaluation. For histopathological categorization, the revised WHO classification for tumors of the central nervous system were adopted (Kleihues and Cavenee, 2000), while for focal cortical dysplasia (FCD) we referred to the classification proposed by Blumcke et al. (2011). Engel’s scoring system (Engel et al., 1993) was employed to assess postoperative seizure outcome. For patients who underwent more than one operation, the outcome referred to the last procedure.
years, the age at epilepsy onset was 5.6 ± 4.9, with a mean duration of epilepsy of 17.3 ± 10.4 years. A high seizure frequency (from 1/week to pluri-daily) was present in 83 patients (87.4%); the remaining 12 patients (12.6%) had a low frequency of seizures (less than 4/month). According to patients’ seizure diaries, 64 subjects presented up to 10% of their seizures during wakefulness which, in some cases, were also detected during video-EEG monitoring. In general, seizures during wakefulness were characterized by less complex motor behaviors with respect to seizures occurring during sleep and in most cases they were limited to a subjective manifestation. In 75 cases (78.9%) brain MRI revealed anatomical unifocal abnormalities. In 20 patients (21.1%) MRI was unrevealing. In 36 patients (37.9%) scalp video-EEG recording of habitual seizures, along with MRI, was sufficient to define the surgical strategy (Fig. 1). Fifty-nine (62.1%) patients required a SEEG investigation, which was monolateral in 48 patients (right-sided in 31 procedures and left-sided in 17) and bilateral in 11 patients (Fig. 2).
Statistical analysis
Surgery
The statistical analysis was performed in successive steps. As a first step, distribution of numerical variables was tested by means of the Shapiro—Wilk normality test. The subsequent univariate analysis included: the Student’s t-test to compare normal numerical variables, the non-parametric Mann—Whitney rank sum test to analyze variables with non-normal distribution, and the Fisher’s two-tailed exact test to analyze categorical (binomial or multinomial) variables. The dependent variable was the seizure-free outcome (Engels class I); the independent variables tested were: gender, duration of epilepsy, age at surgery, seizure frequency (defined as daily, weekly, monthly or sporadic), presence of seizures also during wakefulness, presence of MRI abnormalities, SEEG (performed or not performed), site of surgery, side of surgery (right and left), type of surgical intervention, histology of removed tissue. The site of surgery was classified as: frontal, temporal, fronto-temporal, insularopercular, posterior (including parietal, temporal-occipital and parietal-occipital). Resection of the EZ was classified as complete or incomplete. Finally, a multivariate analysis was performed in order to evaluate the predictive value of each independent variable. All statistics were performed by means of the SYSTAT 12 software version 12.02.00 for Windows® (copyright SYSTAT® Software Inc. 2007).
In 58 patients (61.1%) the cortical resection involved the frontal lobe, while in 37 patients (38.9%) the extrafrontal resections were performed. Sixteen patients (16.8%) (13 frontal, 1 temporal and one parietal) had received a previous cortical resection that did not result in seizure control. Seventy patients (73.7%) had a complete resection of the EZ, whereas in 25 (26.3%) resection of EZ was incomplete. Of these latter, the EZ was partly spared because it involved eloquent cortex in 5 patients, postoperative MRI revealed insufficient resection of the EZ in 9 patients and the indications to surgery were based on data with residual uncertainties as to the actual limits of the EZ in 11 patients. Surgical complications were observed in 3 cases (one case of pulmonary embolism, one of pneumothorax, one of depressed skull fracture). Transient neurological morbidity (deficit lasting no more than three weeks) was observed in 20 patients, with various combinations: 15 hemiparesis, 6 central facial palsy (isolated or associated with hemiparesis), 7 language impairment. Permanent contralateral superior quadrantanopia was documented in 6 out of 19 temporal lobe resections.
primary involvement of other cerebral regions not included in the surgical plan. For patients with hippocampal sclerosis, surgical resection was defined as complete when hippocampal removal was extended posteriorly as far as the aqueductal plane.
Postsurgical evaluation
Results Among 955 cases operated on for drug-resistant epilepsy we identified 103 patients (10.8%) with sleep related seizures. All these patients had a follow up of at least 24 months. Eight cases were excluded because postoperative MRI was not available.
Presurgical evaluation The study population consisted of 95 patients (48 men and 47 women). The mean age at surgery was 22.9 ± 11.6
Histological results (HS) Histological examination of resected specimens disclosed a focal cortical dysplasia (FCD) in 72 patients (75.8%). Seventeen patients had a FCD type I and 55 a FCD type II (14 FCD type IIa; 41 FCD type IIb). In 6 cases the FCD type I was associated to another pathology (FCD type III): a hippocampal sclerosis (HS) in one (FCD type IIIa) and a tumor (2 dysembryoplastic neuroepithelial tumors and 3 gangliogliomas) in 5 cases (FCD type IIIb). In 6 patients an isolated HS was found, while in 7 patients the histopathological evaluation disclosed a tumor not associated to a FCD type I (2 gangliogliomas, 3 dysembryoplastic neuroepithelial tumors, 1 hamartoma, 1 neurocytoma). In 4 cases (4.2%) the pathologic findings consisted in ectopic neurons in the white matter (2 cases), polymicrogyria and periventricular
956
A. Losurdo et al.
Figure 1 Patient with sleep related seizures originating from the left fronto-mesial region. Interictal EEG activity during wakefulness (A) and during NREM sleep (B). Ictal EEG of a seizure occurring during NREM sleep (C). Left fronto-mesial cortical dysplasia (D). Post-operative MRI (E).
heterotopia. In the remaining 6 (6.3%) patients histopathological analysis was unremarkable.
Postoperative outcome on seizures Mean postoperative follow-up was 82.3 months (s.d. ±37.5; range 24—168). Seventy-three patients (76.8%) were in Engel’s class I at last follow-up visit (75.7% at 24 months after surgery), while in 22 cases (23.2%) presented recurrence of disabling seizures (6 patients in Engel’s class II, 7 in Engel’s class III, and 9 in Engel’s class IV). In all the patients
with a follow-up longer than 2 years, the surgical outcome at 2 years was identical to the one at last clinical evaluation. Only in 1 not seizure-free case, recurrent seizures became prevalent during wakefulness. Anticonvulsive drugs had been withdrawn in 48 patients (50.5%), tapered in 25 (26.3%) and they were unchanged in 22 (23.1%).
Statistical analysis As concerns the numerical variables, duration of epilepsy had a non-normal distribution (Shapiro—Wilk 0.962,
Drug-resistant SRE and surgical outcome
957
Figure 2 Patient with sleep related seizures originating from the right insulo-opercular region. Lateral view of the Stereo EEG exploration scheme (A). Blended-image of 3D T1-weighted MRI and co-registered post-implantation volumetric CT, in which single contacts of intracerebral electrodes are easily recognizable (B). Post-operative MRI (C). Stereo-EEG recording of a spontaneous seizure, showing a low voltage fast activity involving the insulo-opercular derivations (D). The arrow indicates the clinical ictal onset. T (temporal region), O (occipital region), aI (anterior insula), P (parietal region), pI (posterior insula), fOp (frontal operculum), cOp (central operculum), F (frontal region).
Table 1 Demographic and presurgical clinical data of patients vs. seizure outcome after surgery. The seizure outcome has been classified according to Engel’s class I and Engel’s class II—IV. M, males; F, females. n.s., not statistically significant (p > 0.05). Duration of epilepsy and age at operation are expressed in years (means ± standard deviation). All data are expressed as number of patients for each category. The correspondent percentage of patients for each variable is between round brackets.
Gender Duration of epilepsy (yrs) Age at operation (yrs) Seizure frequency Daily seizures Weekly seizures Monthly seizures Sporadic seizures Seizures during wakefulness Lesion at MRI SEEG not performed
Total (n. 95)
Engel’s class I (n. 73)
Engel’s class II—IV (n. 22)
p Value
48 M/47 F 17.3 ± 10.4 22.9 ± 11.6
37 M/36 F 17.4 ± 10.4 23.1 ± 11.6
11 M/11 F 17.3 ± 10.4 22.9 ± 11.6
n.s. n.s. n.s.
56 27 8 4 64 75 36
43 (76.8) 21 (77.7) 5 (62.5) 4 (100) 48 (75) 63 (84) 33 (91.7)
13 (23.2) 6 (22.3) 3 (37.5) 0 (0) 16 (25) 12 (16) 3 (8.3)
n.s. n.s. n.s. n.s. n.s. p < 0.001 p = 0.015
958
A. Losurdo et al.
Table 2 Site and type of surgical resections seizure outcome after surgery. All data are expressed as number of patients for each category. The correspondent percentage of patients for each variable is between round brackets. EZ: epileptogenic zone. n.s.: not statistically significant (p > 0.05).
Side of surgical resections Right/left Sites of surgical resections Frontal Temporal Fronto-temporal Insular-opercular Posterior EZ complete removal
Total (n. 95)
Engel’s class I (n. 73)
Engel’s class II—IV (n. 22)
p Value
59/36
44/29
15/7
n.s.
58 19 1 10 7 70
45 (77.6) 16 (84.2) 0 (0) 8 (80.0) 4 (57.1) 62 (88.6)
13 (22.4) 3 (15.8) 1 (100) 2 (20.0) 3 (42.9) 8 (11.4)
n.s. n.s. n.s. n.s. n.s. p < 0.001
Table 3 Results of histological examination vs. seizure outcome after surgery. All data are expressed as number of patients for each category. The correspondent percentage of patients for each variable is between round brackets. FCD, focal cortical dysplasia; HS, hippocampal sclerosis; n.s., not statistically significant (p > 0.05).
FCD type I FCD type Ia FCD type II FCD type IIb FCD type III HS HSIa Tumor Tumora Cryptogenetic Other a
Total (n. 95)
Engel’s class I (n. 73)
Engel’s class II—IV (n. 22)
p Value
17 11 55 42 6 7 6 12 7 6 4
8 (47.0) 3 (27.2) 50 (90.1) 41 (97.6) 5 (83.3) 4 (57.1) 4 (66.7) 11 (91.7) 6 (85.7) 3 (50) 2 (50)
9 (53.0) 8 (72.7) 5 (9.9) 1 (2.4) 1 (16.7) 3 (42.9) 2 (33.3) 1 (8.3) 1 (14.3) 3 (50) 2 (50)
0.003
journal homepage: www.elsevier.com/locate/epilepsyres
Drug-resistant focal sleep related epilepsy: Results and predictors of surgical outcome Anna Losurdo a, Paola Proserpio b, Francesco Cardinale b, Francesca Gozzo b, Laura Tassi b, Roberto Mai b, Stefano Francione b, Laura Castana b, Giorgio Lo Russo b, Giuseppe Casaceli b, Ivana Sartori b, Giacomo Della Marca a, Massimo Cossu b, Lino Nobili b,c,∗ a
Institute of Neurology, Catholic University, Policlinico Universitario A. Gemelli, Rome, Italy ‘‘C. Munari’’ Center of Epilepsy Surgery, Niguarda Hospital, Milan, Italy c Institute of Bioimaging and Molecular Physiology, Genoa Unit, National Research Council, Genoa, Italy b
Received 25 July 2013; received in revised form 3 February 2014; accepted 28 February 2014 Available online 12 March 2014
KEYWORDS Epilepsy; Epilepsy surgery; Sleep; Sleep related epilepsy
Summary In this study we report the results of surgery in a large population of patients affected by drug-resistant focal sleep related epilepsy (SRE) and the identified prognostic factors. We conducted a retrospective analysis of a case series of 955 patients operated on for drug-resistant focal epilepsy from 1997 to 2009. Ninety-five patients with focal SRE and a followup of at least 2 years were identified. Presurgical, surgical and histopathological variables were analyzed. Risk of seizures recurrence was assessed by univariate and multivariate analysis. Mean age at epilepsy onset was 5.6 ± 4.9 years. MRI revealed a focal abnormality in 78.9% of cases. Sixty-two percent of patients required a Stereo-EEG investigation. The cortical resection involved the frontal lobe in 61.1% of cases, while in 38.9% an extrafrontal resection was performed. Focal cortical dysplasia (FCD) type II was the most frequent histopathological finding. Mean postoperative follow-up was 82.3 months. Seventy-three patients (76.8%) were in Engel’s class I. At univariate analysis, variables associated with a favorable outcome were: absence of Stereo-EEG investigation; positive MRI; complete removal of the epileptogenic zone (EZ);
∗ Corresponding author at: Centre of Epilepsy Surgery ‘‘C. Munari’’, Center of Sleep Medicine, Niguarda Hospital, Piazza Ospedale Maggiore 2, 20162 Milan, Italy. Tel.: +39 0264447323; fax: +39 0264442874. E-mail address: [email protected] (L. Nobili).
http://dx.doi.org/10.1016/j.eplepsyres.2014.02.016 0920-1211/© 2014 Elsevier B.V. All rights reserved.
954
A. Losurdo et al. presence of FCD type II and FCD type IIb. A diagnosis of FCD type I was associated with postoperative recurrence of seizures. Multivariate analysis identified the complete removal of the EZ and FCD type I as independent predictors of a favorable and unfavorable outcome respectively. SRE can frequently originate outside the frontal lobe and a favorable surgical outcome is achieved in three-fourths of cases independently from the location of the EZ. © 2014 Elsevier B.V. All rights reserved.
Introduction In patients with sleep related epilepsy (SRE), more than 90% of ictal events arise from sleep (Iber et al., 2007). The percentage of patients with epilepsy affected by focal SRE ranges from 7.5 to 45%, depending on the method of case ascertainment (Thomas et al., 2010). Focal SRE is considered a relatively benign clinical condition (D’Alessandro et al., 1983) because seizures occur almost exclusively during nocturnal sleep and in the majority of patients a good seizure control is achieved with pharmacological treatment. Nevertheless, a relatively high percentage of patients, mainly with a focal frontal epilepsy, is drug resistant (Bernasconi et al., 1998; Nobili et al., 2007; Provini et al., 1999). Sleep related seizures are a typical manifestation of Nocturnal Frontal Lobe Epilepsy (NFLE) (Nobili et al., 2007; Oldani et al., 1998; Provini et al., 1999; Scheffer et al., 1995); however, many studies have shown that, among drug resistant patients, sleep related seizures can originate in the temporal lobe (Bernasconi et al., 1998; Nobili et al., 2004; Tao et al., 2010), in the insular lobe (Dobesberger et al., 2008; Kaido et al., 2006; Proserpio et al., 2011b; Ryvlin et al., 2006; Zhang et al., 2008) and in the posterior cortical regions (Proserpio et al., 2011a). Epilepsy surgery in NFLE has shown to be an effective treatment (Nobili et al., 2007); moreover, although limited to single cases or small groups of patients, good results seem to be achieved also in extra-frontal focal nocturnal seizures (Dobesberger et al., 2008; Elsharkawy et al., 2009; Mai et al., 2005; Proserpio et al., 2011a). The aim of our study is to report the results of surgery in a large population of patients affected by drug resistant focal SRE epilepsy and to point out possible presurgical and surgical prognostic factors.
Materials and methods We have evaluated retrospectively a series of 955 patients operated on for drug-resistant focal epilepsy at ‘‘C. Munari’’ Epilepsy Surgery Centre from 1997 to 2009. The selection criteria were: 1) Presence of sleep-related seizures. Patients were considered to be affected by sleep-related seizures if more than 90% of ictal events arose from sleep, referred to questioning of patients and their relatives. This distribution of seizures was confirmed by seizure diaries filled over a period for at least one year, and by subsequent long-term video-EEG recordings conducted both during sleep and wakefulness. 2) Post-operative follow-up period of at least 24 months.
3) Availability of a post-surgical MRI study (necessary for assessing the complete or incomplete removal of the epileptogenic zone; see ‘‘surgery’’). The study was approved by the Ethic Committee of the Niguarda Hospital, Milan.
Presurgical evaluation All patients underwent a presurgical investigation based on: 1. accurate analysis of personal and epileptic history; 2. scalp video-EEG (VEEG) monitoring, including at least one video-polysomnographic recording of nocturnal sleep; 3. MRI studies were performed according to the protocol proposed by Colombo et al. (2009, 2003), and they were customized with employment of appropriate sequences according to main electroclinical information (1.5-tesla ACS-NT unit; Philips Medical Systems, Best, The Netherlands). Intravenous contrast was injected when a neoplasm was suspected. 4. When non-invasive investigations failed to localize the epileptogenic zone (EZ; the brain region considered essential for inducing and maintaining the epileptic seizures), a stereo-electro-encephalography (SEEG) with stereotactically placed intracerebral electrodes was performed (Cossu et al., 2005). The arrangement of electrodes was tailored according to a predefined localization hypothesis based on non-invasive findings.
Surgery Surgery aimed at resection of the EZ, whose identification was based on the anatomo-electro-clinical correlations. On these basis the area to be removed could be limited to a possible discrete lesion detected by MRI or be extended also to extralesional areas. In all the MRI-negative cases the resection area was defined on the basis of a SEEG investigation. Surgical procedures were classified as complete or incomplete resection of the EZ. Resections were considered incomplete when: (1) the EZ was partly spared because it involved eloquent cortex (2) postoperative MRI revealed that surgical resection did not completely match the preoperative plan (3) the indications to surgery were based on data with residual uncertainties as to the actual limits of the EZ. In particular, the presurgical evaluation allowed the identification of a definite anatomical region to be removed; however it was not able to exclude with certainty the
Drug-resistant SRE and surgical outcome
955
Surgical specimens were processed for histopathological and immunohistochemical evaluation. For histopathological categorization, the revised WHO classification for tumors of the central nervous system were adopted (Kleihues and Cavenee, 2000), while for focal cortical dysplasia (FCD) we referred to the classification proposed by Blumcke et al. (2011). Engel’s scoring system (Engel et al., 1993) was employed to assess postoperative seizure outcome. For patients who underwent more than one operation, the outcome referred to the last procedure.
years, the age at epilepsy onset was 5.6 ± 4.9, with a mean duration of epilepsy of 17.3 ± 10.4 years. A high seizure frequency (from 1/week to pluri-daily) was present in 83 patients (87.4%); the remaining 12 patients (12.6%) had a low frequency of seizures (less than 4/month). According to patients’ seizure diaries, 64 subjects presented up to 10% of their seizures during wakefulness which, in some cases, were also detected during video-EEG monitoring. In general, seizures during wakefulness were characterized by less complex motor behaviors with respect to seizures occurring during sleep and in most cases they were limited to a subjective manifestation. In 75 cases (78.9%) brain MRI revealed anatomical unifocal abnormalities. In 20 patients (21.1%) MRI was unrevealing. In 36 patients (37.9%) scalp video-EEG recording of habitual seizures, along with MRI, was sufficient to define the surgical strategy (Fig. 1). Fifty-nine (62.1%) patients required a SEEG investigation, which was monolateral in 48 patients (right-sided in 31 procedures and left-sided in 17) and bilateral in 11 patients (Fig. 2).
Statistical analysis
Surgery
The statistical analysis was performed in successive steps. As a first step, distribution of numerical variables was tested by means of the Shapiro—Wilk normality test. The subsequent univariate analysis included: the Student’s t-test to compare normal numerical variables, the non-parametric Mann—Whitney rank sum test to analyze variables with non-normal distribution, and the Fisher’s two-tailed exact test to analyze categorical (binomial or multinomial) variables. The dependent variable was the seizure-free outcome (Engels class I); the independent variables tested were: gender, duration of epilepsy, age at surgery, seizure frequency (defined as daily, weekly, monthly or sporadic), presence of seizures also during wakefulness, presence of MRI abnormalities, SEEG (performed or not performed), site of surgery, side of surgery (right and left), type of surgical intervention, histology of removed tissue. The site of surgery was classified as: frontal, temporal, fronto-temporal, insularopercular, posterior (including parietal, temporal-occipital and parietal-occipital). Resection of the EZ was classified as complete or incomplete. Finally, a multivariate analysis was performed in order to evaluate the predictive value of each independent variable. All statistics were performed by means of the SYSTAT 12 software version 12.02.00 for Windows® (copyright SYSTAT® Software Inc. 2007).
In 58 patients (61.1%) the cortical resection involved the frontal lobe, while in 37 patients (38.9%) the extrafrontal resections were performed. Sixteen patients (16.8%) (13 frontal, 1 temporal and one parietal) had received a previous cortical resection that did not result in seizure control. Seventy patients (73.7%) had a complete resection of the EZ, whereas in 25 (26.3%) resection of EZ was incomplete. Of these latter, the EZ was partly spared because it involved eloquent cortex in 5 patients, postoperative MRI revealed insufficient resection of the EZ in 9 patients and the indications to surgery were based on data with residual uncertainties as to the actual limits of the EZ in 11 patients. Surgical complications were observed in 3 cases (one case of pulmonary embolism, one of pneumothorax, one of depressed skull fracture). Transient neurological morbidity (deficit lasting no more than three weeks) was observed in 20 patients, with various combinations: 15 hemiparesis, 6 central facial palsy (isolated or associated with hemiparesis), 7 language impairment. Permanent contralateral superior quadrantanopia was documented in 6 out of 19 temporal lobe resections.
primary involvement of other cerebral regions not included in the surgical plan. For patients with hippocampal sclerosis, surgical resection was defined as complete when hippocampal removal was extended posteriorly as far as the aqueductal plane.
Postsurgical evaluation
Results Among 955 cases operated on for drug-resistant epilepsy we identified 103 patients (10.8%) with sleep related seizures. All these patients had a follow up of at least 24 months. Eight cases were excluded because postoperative MRI was not available.
Presurgical evaluation The study population consisted of 95 patients (48 men and 47 women). The mean age at surgery was 22.9 ± 11.6
Histological results (HS) Histological examination of resected specimens disclosed a focal cortical dysplasia (FCD) in 72 patients (75.8%). Seventeen patients had a FCD type I and 55 a FCD type II (14 FCD type IIa; 41 FCD type IIb). In 6 cases the FCD type I was associated to another pathology (FCD type III): a hippocampal sclerosis (HS) in one (FCD type IIIa) and a tumor (2 dysembryoplastic neuroepithelial tumors and 3 gangliogliomas) in 5 cases (FCD type IIIb). In 6 patients an isolated HS was found, while in 7 patients the histopathological evaluation disclosed a tumor not associated to a FCD type I (2 gangliogliomas, 3 dysembryoplastic neuroepithelial tumors, 1 hamartoma, 1 neurocytoma). In 4 cases (4.2%) the pathologic findings consisted in ectopic neurons in the white matter (2 cases), polymicrogyria and periventricular
956
A. Losurdo et al.
Figure 1 Patient with sleep related seizures originating from the left fronto-mesial region. Interictal EEG activity during wakefulness (A) and during NREM sleep (B). Ictal EEG of a seizure occurring during NREM sleep (C). Left fronto-mesial cortical dysplasia (D). Post-operative MRI (E).
heterotopia. In the remaining 6 (6.3%) patients histopathological analysis was unremarkable.
Postoperative outcome on seizures Mean postoperative follow-up was 82.3 months (s.d. ±37.5; range 24—168). Seventy-three patients (76.8%) were in Engel’s class I at last follow-up visit (75.7% at 24 months after surgery), while in 22 cases (23.2%) presented recurrence of disabling seizures (6 patients in Engel’s class II, 7 in Engel’s class III, and 9 in Engel’s class IV). In all the patients
with a follow-up longer than 2 years, the surgical outcome at 2 years was identical to the one at last clinical evaluation. Only in 1 not seizure-free case, recurrent seizures became prevalent during wakefulness. Anticonvulsive drugs had been withdrawn in 48 patients (50.5%), tapered in 25 (26.3%) and they were unchanged in 22 (23.1%).
Statistical analysis As concerns the numerical variables, duration of epilepsy had a non-normal distribution (Shapiro—Wilk 0.962,
Drug-resistant SRE and surgical outcome
957
Figure 2 Patient with sleep related seizures originating from the right insulo-opercular region. Lateral view of the Stereo EEG exploration scheme (A). Blended-image of 3D T1-weighted MRI and co-registered post-implantation volumetric CT, in which single contacts of intracerebral electrodes are easily recognizable (B). Post-operative MRI (C). Stereo-EEG recording of a spontaneous seizure, showing a low voltage fast activity involving the insulo-opercular derivations (D). The arrow indicates the clinical ictal onset. T (temporal region), O (occipital region), aI (anterior insula), P (parietal region), pI (posterior insula), fOp (frontal operculum), cOp (central operculum), F (frontal region).
Table 1 Demographic and presurgical clinical data of patients vs. seizure outcome after surgery. The seizure outcome has been classified according to Engel’s class I and Engel’s class II—IV. M, males; F, females. n.s., not statistically significant (p > 0.05). Duration of epilepsy and age at operation are expressed in years (means ± standard deviation). All data are expressed as number of patients for each category. The correspondent percentage of patients for each variable is between round brackets.
Gender Duration of epilepsy (yrs) Age at operation (yrs) Seizure frequency Daily seizures Weekly seizures Monthly seizures Sporadic seizures Seizures during wakefulness Lesion at MRI SEEG not performed
Total (n. 95)
Engel’s class I (n. 73)
Engel’s class II—IV (n. 22)
p Value
48 M/47 F 17.3 ± 10.4 22.9 ± 11.6
37 M/36 F 17.4 ± 10.4 23.1 ± 11.6
11 M/11 F 17.3 ± 10.4 22.9 ± 11.6
n.s. n.s. n.s.
56 27 8 4 64 75 36
43 (76.8) 21 (77.7) 5 (62.5) 4 (100) 48 (75) 63 (84) 33 (91.7)
13 (23.2) 6 (22.3) 3 (37.5) 0 (0) 16 (25) 12 (16) 3 (8.3)
n.s. n.s. n.s. n.s. n.s. p < 0.001 p = 0.015
958
A. Losurdo et al.
Table 2 Site and type of surgical resections seizure outcome after surgery. All data are expressed as number of patients for each category. The correspondent percentage of patients for each variable is between round brackets. EZ: epileptogenic zone. n.s.: not statistically significant (p > 0.05).
Side of surgical resections Right/left Sites of surgical resections Frontal Temporal Fronto-temporal Insular-opercular Posterior EZ complete removal
Total (n. 95)
Engel’s class I (n. 73)
Engel’s class II—IV (n. 22)
p Value
59/36
44/29
15/7
n.s.
58 19 1 10 7 70
45 (77.6) 16 (84.2) 0 (0) 8 (80.0) 4 (57.1) 62 (88.6)
13 (22.4) 3 (15.8) 1 (100) 2 (20.0) 3 (42.9) 8 (11.4)
n.s. n.s. n.s. n.s. n.s. p < 0.001
Table 3 Results of histological examination vs. seizure outcome after surgery. All data are expressed as number of patients for each category. The correspondent percentage of patients for each variable is between round brackets. FCD, focal cortical dysplasia; HS, hippocampal sclerosis; n.s., not statistically significant (p > 0.05).
FCD type I FCD type Ia FCD type II FCD type IIb FCD type III HS HSIa Tumor Tumora Cryptogenetic Other a
Total (n. 95)
Engel’s class I (n. 73)
Engel’s class II—IV (n. 22)
p Value
17 11 55 42 6 7 6 12 7 6 4
8 (47.0) 3 (27.2) 50 (90.1) 41 (97.6) 5 (83.3) 4 (57.1) 4 (66.7) 11 (91.7) 6 (85.7) 3 (50) 2 (50)
9 (53.0) 8 (72.7) 5 (9.9) 1 (2.4) 1 (16.7) 3 (42.9) 2 (33.3) 1 (8.3) 1 (14.3) 3 (50) 2 (50)
0.003
