- Email: [email protected]
Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring
Journal of Clinical Neuroscience xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Case report
Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring Shan Wang 1, Bo Jin 1, Linglin Yang, Cong Chen, Yao Ding, Yi Guo, Zhongjin Wang, Wenjie Ming, Yelei Tang, Shuang Wang, Meiping Ding ⇑ Epilepsy Center, Department of Neurology, Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, China
a r t i c l e
i n f o
Article history: Received 25 April 2017 Accepted 21 June 2017 Available online xxxx Keywords: Subclinical seizure Temporal lobe epilepsy Video-electroencephalographic monitoring
a b s t r a c t The aim of this study was to determine the clinical importance and predictors of SCSs in a large population of patients with temporal epilepsy (TLE) undergoing video electroencephalographic (VEEG) monitoring. We reviewed the VEEG data of 327 consecutive patients with TLE admitted to our epilepsy center between August 2012 and January 2017. Demographic, electro-clinical, and neuroimaging data were recorded and re-analyzed. To our knowledge, this is the first study assessing SCSs recorded by longterm VEEG monitoring in patients with TLE. Twenty-seven of 327 (8.3%) patients exhibited SCSs during VEEG monitoring. Of these patients, 24 had both SCSs and clinical seizures. The mean duration of the SCSs was 23.18 s (range: 5–1307 s). Of the 27 patients with SCSs, 24 (88.9%) showed localizing value during the diagnostic process. Seventeen patients exhibited colocalization with clinical seizures, 4 showed useless localization related to clinical seizures, and 3 did not have clinical seizures. Sixteen patients (59.3%) experienced their first SCSs within the first 24 h of monitoring and one had the first SCSs within 20 min. Multivariate logistic regression analysis showed that age <18 years at VEEG monitoring (OR = 3.272, 95% CI = 1.283–8.343, p = 0.013) and bilateral IEDs (OR = 4.558, 95% CI = 1.982–10.477, p < 0.001) were independently associated with the presence of SCSs. Thus, SCSs are not uncommon in patients with TLE, particularly those with age <18 years or bilateral IEDs, and should be considered of significant clinical relevance during the diagnostic process. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Long-term video-electroencephalography (VEEG) is an important diagnostic tool for planning epilepsy surgery, since it can be used for recording clinical seizures, interictal epileptiform discharges (IEDs), and subclinical seizures (SCSs) [1,2]. Clinical seizures and IEDs have been widely studied and have district localizing value and suitability in guiding treatment. However, the clinical characteristics and localizing value of SCSs have rarely been studied, owing to small sample sizes or the limited detection rate of routine electroencephalographic (EEG) recordings in most in-patients [2,3]. SCSs are electrographic seizures lacking subjective or obvious objective neurological and somatic symptoms, and are likely to be ignored in clinical practice [4,5]. However, SCSs may be accom⇑ Corresponding author at: Epilepsy Center, Department of Neurology, Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou 310009, China. E-mail address: [email protected] (M. Ding). 1 These authors contributed equally to this paper.
panied by subtle clinical and imaging manifestations, including unnoticed cognitive disturbances, memory deficits and hypermetabolism in 18F-FDG PET images [6,7]. There have been suggestions that SCSs could be as significant to the surgical decision process as clinical seizures, especially in TLE [4,5,8,9]. Patients with SCSs (incidence >80%) were more likely to become seizure-free than those without SCSs (incidence <29%) were, after temporal lobectomy [4]. Recent studies showed that surgical outcome might be influenced by the colocalization of SCSs and clinical seizures, rather than the prevalence of SCSs [5]. SCSs could also be used to improve the accuracy of localization by reducing disturbing movement artifacts in EEG monitoring. Thus, it is important to investigate the prevalence of SCSs and their colocalization rate with clinical seizures. Previous studies on SCSs in patients with epilepsy included many biases, especially in patient selection, and often focused on patients with refractory epilepsy or pediatric focal epilepsy, without epilepsy classification [2,8,10]. The prevalence of SCSs in the general population remains unclear [1]. TLE is the most common focal epilepsy, and accounts for up to 30% of refractory patients
http://dx.doi.org/10.1016/j.jocn.2017.06.071 0967-5868/Ó 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Wang S et al. Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring. J Clin Neurosci (2017), http://dx.doi.org/10.1016/j.jocn.2017.06.071
2
S. Wang et al. / Journal of Clinical Neuroscience xxx (2017) xxx–xxx
[11]. Persistent SCSs in temporal regions lead to predominant memory impairment [7]. Normal temporal lobe networks or structures are sensitive to IEDs, which might disrupt cognition or memory performance [7]. However, there is no data available on the exact prevalence and risk factors of SCSs in TLE based on a larger cohort. We therefore reviewed the demographic, electroclinical, and neuroimaging data of a large cohort of patients with TLE in order to elucidate the prevalence and localizing value of SCSs, and the associated risk factors in patients with TLE.
3. Results Of the 327 patients with TLE included in the study, 144 (44.0%) had refractory epilepsy and 160 (49.2%) were females. The mean duration of VEEG monitoring was 2.4 days (range: 1–15 days). A total of 175 SCSs were detected in 27 (8.3%) of the 327 patients who underwent VEEG. The mean duration of SCSs was 23.18 s (range, 5–1307 s). Of these, 24 patients (88.9%) had both SCSs and clinical seizures. 3.1. Clinical characteristics of SCSs
2. Materials and methods 2.1. Patients We retrospectively reviewed the medical charts and VEEG reports of patients consecutively admitted to our epilepsy center at the Second Affiliated Hospital, School of Medicine, Zhejiang University, for seizure diagnosis, medication and adjustment, and presurgical evaluation between August 2012 and January 2017. Patients with acute symptomatic seizures caused by brain insults resulting from conditions such as trauma, encephalitis, intracranial infection, stroke, and hypoxic-ischemic encephalopathy, with previous history of brain surgery were excluded. A total of 327 patients were chosen for analysis. Only patients who underwent scalp VEEG monitoring and 3.0 T MRI scan, and were finally diagnosed TLE were included. 2.2. VEEG procedure VEEG was performed using digital VEEG systems (Nicolet, VIASYS, USA and Biologic, NATUS, USA) with scalp electrodes placed according to the international 10–20 systems. Additional scalp electrodes or sphenoidal electrodes were used if necessary. Electromyography (EMG) in the bilateral upper extremities and electrocardiogram (EKG) were routinely recorded. All patients were monitored for at least 24 h. Medications were not adjusted during VEEG in patients admitted for diagnostic clarification. In patients admitted for presurgical evaluation, medications were not changed during the first two days of VEEG monitoring. Clinical and EEG data of these patients were re-analyzed. Clinical seizure frequencies within 6 months before admission were categorized as daily (a least once a day), persistent (at least once a week), or rare (less than once a week). Ictal EEG patterns were classified as either regional (appearing exclusively within a single lobe or in two contiguous regions, such as frontotemporal discharges) or non-regional (e.g. multi-lobar, hemispheric, or generalized). Interictal epileptiform discharges (IEDs) were classified as lateral or bilateral discharges. SCSs were defined as abnormal, paroxysmal electroencephalographic events differing from the background, without corresponding clinical manifestation, with a temporal-spatial evolution in morphology, frequency, and amplitude, and with a plausible electrographic field [1]. For each patient, SCSs were determined by at least two reviewers (Y. Tang and Z. Wang). If reviewers disagreed, an additional reviewer (S. Wang) was consulted, and a consensus was reached. 2.3. Statistical analysis Data are presented as median (Percentiles 25–75). Statistical analyses were carried out using the Mann-Whitney U test, Pearson Chi-square, or Fisher exact tests for comparing TLE patients with and without SCSs. Variables with a significance level of 5% on univariate analysis were then entered in a multivariable logistic regression model with statistical significance set at 5%.
SCSs showed regional or hemispheric onset in 23 patients (temporal in 17, hemispheric in 6), and generalized onset in 4 patients (14.8%). In the 24 patients with both SCSs and clinical seizures, 17 (70.8%) had SCSs and clinical seizures arising from the same brain area. In 24 (88.9%) patients, the presence of SCSs had diagnostic value for localizing the epileptogenic focus; 17 patients showed SCS colocalization with clinical seizures, 4 patients had clinical seizures that did not localize or lateralize, and 3 patients did not have clinical seizures. The median duration of VEEG recordings was 48 h (24–96 h) in patients without SCSs and 24 h (24–72 h) in those with SCSs. Sixteen patients (59.3%) had their first subclinical seizure during the first 24 h of VEEG monitoring, 21 (77.7%) had their first subclinical seizure during the first 48 h of VEEG monitoring, and 25 (92.6%) had their first subclinical seizure during the first 72 h of VEEG monitoring. In 3 patients with only SCSs, all SCSs were detected during the first 24 h of monitoring. The age was <18 years in 8 of 27 patients (29.6%) with SCSs and 41 of 259 patients (13.7%) without SCSs. Of the 144 patients with pharmacoresistant epilepsy, 17 (11.8%) had subclinical seizures. 3.2. Risk factors for SCSs In univariate analysis, we found that SCSs were associated with pharmacoresistant epilepsy, age at VEEG monitoring, and the presence of bilateral IEDs (Table 1). To avoid collinearity or redundancy, logistic regression analysis was performed. Logistic regression analysis showed that younger age (<18 years) at VEEG monitoring (p = 0.013) and the presence of bilateral IEDs (p < 0.001) were independently associated with SCSs (Table 2). 4. Discussion In the present study, SCSs were recorded in 8.3% of patients with TLE admitted for VEEG monitoring. In 24 patients (88.9%), the presence of SCSs had localizing value during the diagnostic process. Age <18 years and the presence of bilateral IEDs were independently associated with SCSs. To the best of our knowledge, no prior data is available on the frequency of SCSs in patients with TLE. The prevalence of SCSs in this study was lower than that reported from previous studies focusing on refractory epilepsy or pediatric patients, especially those using intracranial EEG, where SCS occurrence ranged from 58% to 64% [4,5,8]. Variance of group population could contribute to such a difference. Our study only included patients with TLE covering all ages and was more representative to analyze [2,8,10]. In addition, the attenuating property of the skull might decrease the sensitivity of scalp EEG to detect spikes compared to intracranial EEG [4,5]. Intracranial spikes with <6 cm2 of synchronized cortical activity would likely not produce scalp EEG signals, while 20–30 cm2 gyrus areas are typically preferred to record prominent scalp spikes [12]. Seizures could be recorded on clinical macroelectrodes after a sufficient population of synchronously firing cells is recruited and detectable on the microelectrodes [13].
Please cite this article in press as: Wang S et al. Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring. J Clin Neurosci (2017), http://dx.doi.org/10.1016/j.jocn.2017.06.071
3
S. Wang et al. / Journal of Clinical Neuroscience xxx (2017) xxx–xxx Table 1 The demographic and clinical characteristic of patients with TLE with or without subclinical seizures. Variables a
Pharmacoresistant epilepsy Malea, N (%) Age at epilepsy onsetb (years) Age at VEEG monitoringa (years) 18 years <18 years Epilepsy durationb (months) Number of AEDsb Left side of epileptogenic focusa Epigastric aurasa Childhood febrile seizurea Trauma-hypoxiaa CNS-infectiona Family historya History of status epilepticusa Number of seizure typesb Seizure frequencya Daily (a least once a day) Persistent (at least once a week) Rare (less than once a week) Reducing AEDs during monitoringa Duration of VEEG monitoring (h)b Habitual seizuresa History of sGTCSa Non Recent Remote PET characteristicsa Normal Abnormal Hippocampal sclerosis on MRIa Presence of IEDsa Lateral Bilateral
Subclinical seizure (n = 27)
Non-subclinical seizure (299)
P value
17 (63%) 11 (40.7%) 12.5 (4–20)
127 (42.3%) 155 (51.7%) 19 (12–30.75)
0.039* 0.277 0.067 0.043*
19 (70.4%) 8 (29.6%) 84 (12–240) 2 (1–3) 14 (51.9%) 4 (14.8%) 7 (25.9%) 1 (3.7%) 2 (7.4%) 2 (7.4%) 2 (7.4%) 2 (1–3)
259 (86.3%) 41 (13.7%) 72 (24–180) 1 (1–2) 152 (50.7%) 36 (12%) 41 (13.7%) 35 (11.7%) 17 (5.7%) 6 (2.0%) 4 (1.3%) 2 (1–2)
5 (18.5%) 11 (40.7%) 11 (40.7%) 14 (51.9%) 48 (24–96) 17 (85%)
46 (15.3%) 105 (35.0%) 149 (49.7%) 115 (38.3%) 24 (24–72) 137 (79.2%)
10 (37.0%) 3 (11.1%) 14 (51.9%)
149 (49.7%) 21 (7.0%) 130 (43.3%)
4 (28.6%) 10 (71.4%) 10 (37%)
9 (12.2%) 65 (87.8%) 110 (36.7%)
11 (40.7%) 16 (59.3%)
221 (73.7%) 79 (26.3%)
0.853 0.293 0.906 0.757 0.092 0.335 0.663 0.135 0.080 0.313 0.672
0.169 0.950 0.770 0.408
0.210
0.969 <0.001***
Note: VEEG = video-electroencephalography; AEDs = antiepileptic drugs; CNS = central nervous system; sGTCS = secondary generalized tonic-clonic seizure; PET = Positron Emission Tomography; MRI = magnetic resonance imaging; IEDs = interictal epileptiform discharges. a The Chi-square test was used for comparison. b The Mann-Whitney U test was used. * P < 0.05. *** P < 0.001.
Table 2 Logistic regression analysis of factors associated with subclinical seizures. Variables
OR (95% CI)
P-value
Pharmacoresistant epilepsy Age at VEEG monitoring (years) 18 years <18 years Presence of IEDs Lateral Bilateral
1.932 (0.806–4.631)
0.140
1 3.272 (1.283–8.343)
0.013*
1 4.558 (1.982–10.477)
<0.001***
Note: OR = odds ratio; CI = confidence interval; VEEG = video electroencephalography; IEDs = interictal epileptiform discharges. * P < 0.05. *** P < 0.001.
Cortical epileptogenic generators of scalp EEG potentials are larger than initially considered and simple scalp EEG could miss certain SCSs. SCSs lack obvious clinical symptoms because of inadequate neuron recruitment. Only 7% of neurons had synchronous firing during SCSs without any clinical signals [10]. Compared to auras with certain altered consciousness and only 14% neurons fired, clinical seizures had more firing of neurons (approximately 36%) [10]. The occurrence of clinical symptoms depended on the percentage of fired neurons recruited in an epileptic focus [10]. Although lacking distinct semiology, SCSs could have important localizing significance in the clinical decision-making process.
Our results showed that SCSs had a high probability of originating from the same epileptogenic site as clinical seizures and provided useful localizing information in 24 of 27 patients (88.9%). These findings were concordant with those of a previous study wherein SCSs had supplementary value in 85% of patients during the diagnostic process [8]. Sperling et al. hypothesized that the presence of SCSs represented the preservation of the brain’s ability to restrict seizure development and restrain propagation [4]. Therefore, patients with SCSs were more likely to become seizure-free after temporal lobectomy than those without SCSs [4]. Another study suggested that patients with complete colocalization between SCSs and clinical seizures had more than twice the seizure-free outcome rate than patients with partial or contradicted colocalization in TLE and extra-TLE [5]. The impact of SCS localization on surgical outcome highlights the conformity between SCSs and clinical seizures. A recent study showed that all patients with SCSs that propagated to different regions beyond their onset zone had poor surgical outcomes [9]. The ability and the tendency of SCSs to spread were also correlated with surgical outcome since they might provide additional distinct epileptic networks with the potential to affect surgical outcome [9]. All the aforementioned studies emphasized the potential importance of SCSs, despite the lack of distinct symptoms, for more accurate prediction of prognosis when evaluating epilepsy surgery. In this study, we found that patients aged <18 years were more likely to exhibit SCSs during VEEG monitoring compared to those aged 18 years. This finding was in accordance with previous evidence of a higher occurrence of SCSs in young subjects with epi-
Please cite this article in press as: Wang S et al. Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring. J Clin Neurosci (2017), http://dx.doi.org/10.1016/j.jocn.2017.06.071
4
S. Wang et al. / Journal of Clinical Neuroscience xxx (2017) xxx–xxx
lepsy [2,8]. SCSs were reported in 32 children (17.0%) among 188 patients with epilepsy, who were younger than the patients without SCSs [2]. One study found SCSs in 59 of 322 (18%) children with focal epilepsy, which was also higher than the results of our study, which included adults and children [8]. Additionally, a higher frequency of SCSs was described in newborns or children with critical illness [2]. SCSs in young children might sometimes help locating the epileptogenic focus when patients and physicians have difficulties in recognizing the performance associated with seizures or interpreting a polluted EEG manifestation due to severe movement artifacts [14]. SCSs were also more common in patients with bilateral IEDs [3]. IEDs induce long-term changes in synaptic connections between neurons and guide the development of neuronal circuits related to initiation of spontaneous seizures [15]. The management of an unprovoked first seizure in adults demonstrated that IEDs were associated with a relative increase in seizure recurrence rate at 1–5 years after the first seizure [15]. Therefore, it was unsurprising that IEDs could induce seizures. Bilateral IEDs and SCSs in mesial temporal lobes have a tendency to occur simultaneously according to case reports [7]. Published data has shown that bilateral IEDs in lateral TLE appeared in only 14% of patients compared, with mesial TLE where bilateral IEDs were present in 42% of patients [16]. Most SCSs in TLE emanated from mesial temporal structures to which they remained confined and seldom spread to another neocortex [4]. Thus, SCSs and bilateral IEDs tend to appear in regions of the mesial temporal lobe, a restricted circuit for epileptogenesis poorly connected with other brain regions [4,17]. It appears likely that seizure activity emanating from specific regions of the mesial TLE may not produce any obvious behavioral effects [17]. However, we did not differentiate neocortical from mesial TLE in this study because a majority of the patient population did not undergo surgery to identify the focus more accurately. Further studies are therefore needed in this context. Several limitations apply to this study. It is a retrospective study and our results may over- or underestimate the frequency of SCSs in patients with TLE, since this analysis was not performed on a random population-based sample of patients with TLE, and time of VEEG monitoring may have been inadequate to discover all SCSs. Moreover, we diagnosed TLE based on complex partial seizures, VEEG, and imaging data, whereas the gold standard of TLE diagnosis is based on the pathology or long-term surgical outcome. 5. Conclusion In summary, SCSs are not uncommon in patients with TLE, particularly those aged <18 years and those with bilateral IEDs, during VEEG monitoring, and merit as much consideration as clinical seizures for guiding the surgical decision and advising patients about prognosis. Since not all SCSs could be discovered within the first 24 h of VEEG monitoring, it is appropriate to extend the monitoring time for their detection to improve the accuracy of seizure frequency estimation in the future.
Conflict of interest statement None. Acknowledgments This work was supported by grants from the National Natural Science Foundation of China [grant numbers 81401071, 81300991, 81671283], and the National Natural Science Foundation of Zhejiang Province [grant number Y16H090030]. References [1] Jin B, Wang S, Yang L, Shen C, Ding Y, Guo Y, Wang Z, Zhu J, Wang S, Ding M. Prevalence and predictors of subclinical seizures during scalp video-EEG monitoring in patients with epilepsy. Int J Neurosci 2016:1–8. [2] Akman CI, Montenegro MA, Jacob S, Eck K, McBrian D, Chiriboga CA, Patterson MC. Subclinical seizures in children diagnosed with localization-related epilepsy: clinical and EEG characteristics. Epilepsy Behav 2009;16:86–98. [3] Kanazawa K, Matsumoto R, Shimotake A, Kinoshita M, Otsuka A, Watanabe O, Tanaka K, Takahashi R, Ikeda A. Persistent frequent subclinical seizures and memory impairment after clinical remission in smoldering limbic encephalitis. Epileptic Disord 2014;16:312–7. [4] Sperling MR, O’Connor MJ. Auras and subclinical seizures: characteristics and prognostic significance. Ann Neurol 1990;28:320–8. [5] Zangaladze A, Nei M, Liporace JD, Sperling MR. Characteristics and clinical significance of subclinical seizures. Epilepsia 2008;49:2016–21. [6] Bridgman PA, Malamut BL, Sperling MR, Saykin AJ, O’Connor MJ. Memory during subclinical hippocampal seizures. Neurology 1989;39:853–6. [7] Tafti BA, Mandelkern M, Berenji GR. Subclinical seizures as a pitfall in 18F-FDG PET imaging of temporal lobe epilepsy. Clin Nucl Med 2014;39:819–21. [8] Velkey A, Siegler Z, Janszky J, Duray B, Fogarasi A. Clinical value of subclinical seizures in children with focal epilepsy. Epilepsy Res 2011;95:82–5. [9] Farooque P, Duckrow R. Subclinical seizures during intracranial EEG recording: are they clinically significant? Epilepsy Res 2014;108:1790–6. [10] Babb TL, Wilson CL, Isokawa-Akesson M. Firing patterns of human limbic neurons during stereoencephalography (SEEG) and clinical temporal lobe seizures. Electroencephalogr Clin Neurophysiol 1987;66:467–82. [11] Schmeiser B, Hammen T, Steinhoff BJ, Zentner J, Schulze-Bonhage A. Long-term outcome characteristics in mesial temporal lobe epilepsy with and without associated cortical dysplasia. Epilepsy Res 2016;126:147–56. [12] Tao JX, Ray A, Hawes-Ebersole S, Ebersole JS. Intracranial EEG substrates of scalp EEG interictal spikes. Epilepsia 2005;46:669–76. [13] Stead M, Bower M, Brinkmann BH, Lee K, Marsh WR, Meyer FB, Litt B, Van Gompel J, Worrell GA. Microseizures and the spatiotemporal scales of human partial epilepsy. Brain 2010;133:2789–97. [14] Nordli Jr DR, Bazil CW, Scheuer ML, Pedley TA. Recognition and classification of seizures in infants. Epilepsia 1997;38:553–60. [15] Krumholz A, Shinnar S, French J, Gronseth G, Wiebe S. Evidence-based guideline: management of an unprovoked first seizure in adults: report of the Guideline Development Subcommittee of the American Academy of Neurology and the American Epilepsy Society. Neurology 2015;85:1526–7. [16] Janszky J, Rasonyi G, Clemens Z, Schulz R, Hoppe M, Barsi P, Fogarasi A, Halasz P, Ebner A. Clinical differences in patients with unilateral hippocampal sclerosis and unitemporal or bitemporal epileptiform discharges. Seizure 2003;12:550–4. [17] Asadi-Pooya AA, Rostami C, Rabiei AH, Sperling MR. Factors associated with tonic-clonic seizures in patients with drug-resistant mesial temporal epilepsy. J Neurol Sci 2015;359:452–4.
Please cite this article in press as: Wang S et al. Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring. J Clin Neurosci (2017), http://dx.doi.org/10.1016/j.jocn.2017.06.071
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Case report
Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring Shan Wang 1, Bo Jin 1, Linglin Yang, Cong Chen, Yao Ding, Yi Guo, Zhongjin Wang, Wenjie Ming, Yelei Tang, Shuang Wang, Meiping Ding ⇑ Epilepsy Center, Department of Neurology, Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou, China
a r t i c l e
i n f o
Article history: Received 25 April 2017 Accepted 21 June 2017 Available online xxxx Keywords: Subclinical seizure Temporal lobe epilepsy Video-electroencephalographic monitoring
a b s t r a c t The aim of this study was to determine the clinical importance and predictors of SCSs in a large population of patients with temporal epilepsy (TLE) undergoing video electroencephalographic (VEEG) monitoring. We reviewed the VEEG data of 327 consecutive patients with TLE admitted to our epilepsy center between August 2012 and January 2017. Demographic, electro-clinical, and neuroimaging data were recorded and re-analyzed. To our knowledge, this is the first study assessing SCSs recorded by longterm VEEG monitoring in patients with TLE. Twenty-seven of 327 (8.3%) patients exhibited SCSs during VEEG monitoring. Of these patients, 24 had both SCSs and clinical seizures. The mean duration of the SCSs was 23.18 s (range: 5–1307 s). Of the 27 patients with SCSs, 24 (88.9%) showed localizing value during the diagnostic process. Seventeen patients exhibited colocalization with clinical seizures, 4 showed useless localization related to clinical seizures, and 3 did not have clinical seizures. Sixteen patients (59.3%) experienced their first SCSs within the first 24 h of monitoring and one had the first SCSs within 20 min. Multivariate logistic regression analysis showed that age <18 years at VEEG monitoring (OR = 3.272, 95% CI = 1.283–8.343, p = 0.013) and bilateral IEDs (OR = 4.558, 95% CI = 1.982–10.477, p < 0.001) were independently associated with the presence of SCSs. Thus, SCSs are not uncommon in patients with TLE, particularly those with age <18 years or bilateral IEDs, and should be considered of significant clinical relevance during the diagnostic process. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Long-term video-electroencephalography (VEEG) is an important diagnostic tool for planning epilepsy surgery, since it can be used for recording clinical seizures, interictal epileptiform discharges (IEDs), and subclinical seizures (SCSs) [1,2]. Clinical seizures and IEDs have been widely studied and have district localizing value and suitability in guiding treatment. However, the clinical characteristics and localizing value of SCSs have rarely been studied, owing to small sample sizes or the limited detection rate of routine electroencephalographic (EEG) recordings in most in-patients [2,3]. SCSs are electrographic seizures lacking subjective or obvious objective neurological and somatic symptoms, and are likely to be ignored in clinical practice [4,5]. However, SCSs may be accom⇑ Corresponding author at: Epilepsy Center, Department of Neurology, Second Affiliated Hospital, School of Medicine, Zhejiang University, Hangzhou 310009, China. E-mail address: [email protected] (M. Ding). 1 These authors contributed equally to this paper.
panied by subtle clinical and imaging manifestations, including unnoticed cognitive disturbances, memory deficits and hypermetabolism in 18F-FDG PET images [6,7]. There have been suggestions that SCSs could be as significant to the surgical decision process as clinical seizures, especially in TLE [4,5,8,9]. Patients with SCSs (incidence >80%) were more likely to become seizure-free than those without SCSs (incidence <29%) were, after temporal lobectomy [4]. Recent studies showed that surgical outcome might be influenced by the colocalization of SCSs and clinical seizures, rather than the prevalence of SCSs [5]. SCSs could also be used to improve the accuracy of localization by reducing disturbing movement artifacts in EEG monitoring. Thus, it is important to investigate the prevalence of SCSs and their colocalization rate with clinical seizures. Previous studies on SCSs in patients with epilepsy included many biases, especially in patient selection, and often focused on patients with refractory epilepsy or pediatric focal epilepsy, without epilepsy classification [2,8,10]. The prevalence of SCSs in the general population remains unclear [1]. TLE is the most common focal epilepsy, and accounts for up to 30% of refractory patients
http://dx.doi.org/10.1016/j.jocn.2017.06.071 0967-5868/Ó 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Wang S et al. Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring. J Clin Neurosci (2017), http://dx.doi.org/10.1016/j.jocn.2017.06.071
2
S. Wang et al. / Journal of Clinical Neuroscience xxx (2017) xxx–xxx
[11]. Persistent SCSs in temporal regions lead to predominant memory impairment [7]. Normal temporal lobe networks or structures are sensitive to IEDs, which might disrupt cognition or memory performance [7]. However, there is no data available on the exact prevalence and risk factors of SCSs in TLE based on a larger cohort. We therefore reviewed the demographic, electroclinical, and neuroimaging data of a large cohort of patients with TLE in order to elucidate the prevalence and localizing value of SCSs, and the associated risk factors in patients with TLE.
3. Results Of the 327 patients with TLE included in the study, 144 (44.0%) had refractory epilepsy and 160 (49.2%) were females. The mean duration of VEEG monitoring was 2.4 days (range: 1–15 days). A total of 175 SCSs were detected in 27 (8.3%) of the 327 patients who underwent VEEG. The mean duration of SCSs was 23.18 s (range, 5–1307 s). Of these, 24 patients (88.9%) had both SCSs and clinical seizures. 3.1. Clinical characteristics of SCSs
2. Materials and methods 2.1. Patients We retrospectively reviewed the medical charts and VEEG reports of patients consecutively admitted to our epilepsy center at the Second Affiliated Hospital, School of Medicine, Zhejiang University, for seizure diagnosis, medication and adjustment, and presurgical evaluation between August 2012 and January 2017. Patients with acute symptomatic seizures caused by brain insults resulting from conditions such as trauma, encephalitis, intracranial infection, stroke, and hypoxic-ischemic encephalopathy, with previous history of brain surgery were excluded. A total of 327 patients were chosen for analysis. Only patients who underwent scalp VEEG monitoring and 3.0 T MRI scan, and were finally diagnosed TLE were included. 2.2. VEEG procedure VEEG was performed using digital VEEG systems (Nicolet, VIASYS, USA and Biologic, NATUS, USA) with scalp electrodes placed according to the international 10–20 systems. Additional scalp electrodes or sphenoidal electrodes were used if necessary. Electromyography (EMG) in the bilateral upper extremities and electrocardiogram (EKG) were routinely recorded. All patients were monitored for at least 24 h. Medications were not adjusted during VEEG in patients admitted for diagnostic clarification. In patients admitted for presurgical evaluation, medications were not changed during the first two days of VEEG monitoring. Clinical and EEG data of these patients were re-analyzed. Clinical seizure frequencies within 6 months before admission were categorized as daily (a least once a day), persistent (at least once a week), or rare (less than once a week). Ictal EEG patterns were classified as either regional (appearing exclusively within a single lobe or in two contiguous regions, such as frontotemporal discharges) or non-regional (e.g. multi-lobar, hemispheric, or generalized). Interictal epileptiform discharges (IEDs) were classified as lateral or bilateral discharges. SCSs were defined as abnormal, paroxysmal electroencephalographic events differing from the background, without corresponding clinical manifestation, with a temporal-spatial evolution in morphology, frequency, and amplitude, and with a plausible electrographic field [1]. For each patient, SCSs were determined by at least two reviewers (Y. Tang and Z. Wang). If reviewers disagreed, an additional reviewer (S. Wang) was consulted, and a consensus was reached. 2.3. Statistical analysis Data are presented as median (Percentiles 25–75). Statistical analyses were carried out using the Mann-Whitney U test, Pearson Chi-square, or Fisher exact tests for comparing TLE patients with and without SCSs. Variables with a significance level of 5% on univariate analysis were then entered in a multivariable logistic regression model with statistical significance set at 5%.
SCSs showed regional or hemispheric onset in 23 patients (temporal in 17, hemispheric in 6), and generalized onset in 4 patients (14.8%). In the 24 patients with both SCSs and clinical seizures, 17 (70.8%) had SCSs and clinical seizures arising from the same brain area. In 24 (88.9%) patients, the presence of SCSs had diagnostic value for localizing the epileptogenic focus; 17 patients showed SCS colocalization with clinical seizures, 4 patients had clinical seizures that did not localize or lateralize, and 3 patients did not have clinical seizures. The median duration of VEEG recordings was 48 h (24–96 h) in patients without SCSs and 24 h (24–72 h) in those with SCSs. Sixteen patients (59.3%) had their first subclinical seizure during the first 24 h of VEEG monitoring, 21 (77.7%) had their first subclinical seizure during the first 48 h of VEEG monitoring, and 25 (92.6%) had their first subclinical seizure during the first 72 h of VEEG monitoring. In 3 patients with only SCSs, all SCSs were detected during the first 24 h of monitoring. The age was <18 years in 8 of 27 patients (29.6%) with SCSs and 41 of 259 patients (13.7%) without SCSs. Of the 144 patients with pharmacoresistant epilepsy, 17 (11.8%) had subclinical seizures. 3.2. Risk factors for SCSs In univariate analysis, we found that SCSs were associated with pharmacoresistant epilepsy, age at VEEG monitoring, and the presence of bilateral IEDs (Table 1). To avoid collinearity or redundancy, logistic regression analysis was performed. Logistic regression analysis showed that younger age (<18 years) at VEEG monitoring (p = 0.013) and the presence of bilateral IEDs (p < 0.001) were independently associated with SCSs (Table 2). 4. Discussion In the present study, SCSs were recorded in 8.3% of patients with TLE admitted for VEEG monitoring. In 24 patients (88.9%), the presence of SCSs had localizing value during the diagnostic process. Age <18 years and the presence of bilateral IEDs were independently associated with SCSs. To the best of our knowledge, no prior data is available on the frequency of SCSs in patients with TLE. The prevalence of SCSs in this study was lower than that reported from previous studies focusing on refractory epilepsy or pediatric patients, especially those using intracranial EEG, where SCS occurrence ranged from 58% to 64% [4,5,8]. Variance of group population could contribute to such a difference. Our study only included patients with TLE covering all ages and was more representative to analyze [2,8,10]. In addition, the attenuating property of the skull might decrease the sensitivity of scalp EEG to detect spikes compared to intracranial EEG [4,5]. Intracranial spikes with <6 cm2 of synchronized cortical activity would likely not produce scalp EEG signals, while 20–30 cm2 gyrus areas are typically preferred to record prominent scalp spikes [12]. Seizures could be recorded on clinical macroelectrodes after a sufficient population of synchronously firing cells is recruited and detectable on the microelectrodes [13].
Please cite this article in press as: Wang S et al. Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring. J Clin Neurosci (2017), http://dx.doi.org/10.1016/j.jocn.2017.06.071
3
S. Wang et al. / Journal of Clinical Neuroscience xxx (2017) xxx–xxx Table 1 The demographic and clinical characteristic of patients with TLE with or without subclinical seizures. Variables a
Pharmacoresistant epilepsy Malea, N (%) Age at epilepsy onsetb (years) Age at VEEG monitoringa (years) 18 years <18 years Epilepsy durationb (months) Number of AEDsb Left side of epileptogenic focusa Epigastric aurasa Childhood febrile seizurea Trauma-hypoxiaa CNS-infectiona Family historya History of status epilepticusa Number of seizure typesb Seizure frequencya Daily (a least once a day) Persistent (at least once a week) Rare (less than once a week) Reducing AEDs during monitoringa Duration of VEEG monitoring (h)b Habitual seizuresa History of sGTCSa Non Recent Remote PET characteristicsa Normal Abnormal Hippocampal sclerosis on MRIa Presence of IEDsa Lateral Bilateral
Subclinical seizure (n = 27)
Non-subclinical seizure (299)
P value
17 (63%) 11 (40.7%) 12.5 (4–20)
127 (42.3%) 155 (51.7%) 19 (12–30.75)
0.039* 0.277 0.067 0.043*
19 (70.4%) 8 (29.6%) 84 (12–240) 2 (1–3) 14 (51.9%) 4 (14.8%) 7 (25.9%) 1 (3.7%) 2 (7.4%) 2 (7.4%) 2 (7.4%) 2 (1–3)
259 (86.3%) 41 (13.7%) 72 (24–180) 1 (1–2) 152 (50.7%) 36 (12%) 41 (13.7%) 35 (11.7%) 17 (5.7%) 6 (2.0%) 4 (1.3%) 2 (1–2)
5 (18.5%) 11 (40.7%) 11 (40.7%) 14 (51.9%) 48 (24–96) 17 (85%)
46 (15.3%) 105 (35.0%) 149 (49.7%) 115 (38.3%) 24 (24–72) 137 (79.2%)
10 (37.0%) 3 (11.1%) 14 (51.9%)
149 (49.7%) 21 (7.0%) 130 (43.3%)
4 (28.6%) 10 (71.4%) 10 (37%)
9 (12.2%) 65 (87.8%) 110 (36.7%)
11 (40.7%) 16 (59.3%)
221 (73.7%) 79 (26.3%)
0.853 0.293 0.906 0.757 0.092 0.335 0.663 0.135 0.080 0.313 0.672
0.169 0.950 0.770 0.408
0.210
0.969 <0.001***
Note: VEEG = video-electroencephalography; AEDs = antiepileptic drugs; CNS = central nervous system; sGTCS = secondary generalized tonic-clonic seizure; PET = Positron Emission Tomography; MRI = magnetic resonance imaging; IEDs = interictal epileptiform discharges. a The Chi-square test was used for comparison. b The Mann-Whitney U test was used. * P < 0.05. *** P < 0.001.
Table 2 Logistic regression analysis of factors associated with subclinical seizures. Variables
OR (95% CI)
P-value
Pharmacoresistant epilepsy Age at VEEG monitoring (years) 18 years <18 years Presence of IEDs Lateral Bilateral
1.932 (0.806–4.631)
0.140
1 3.272 (1.283–8.343)
0.013*
1 4.558 (1.982–10.477)
<0.001***
Note: OR = odds ratio; CI = confidence interval; VEEG = video electroencephalography; IEDs = interictal epileptiform discharges. * P < 0.05. *** P < 0.001.
Cortical epileptogenic generators of scalp EEG potentials are larger than initially considered and simple scalp EEG could miss certain SCSs. SCSs lack obvious clinical symptoms because of inadequate neuron recruitment. Only 7% of neurons had synchronous firing during SCSs without any clinical signals [10]. Compared to auras with certain altered consciousness and only 14% neurons fired, clinical seizures had more firing of neurons (approximately 36%) [10]. The occurrence of clinical symptoms depended on the percentage of fired neurons recruited in an epileptic focus [10]. Although lacking distinct semiology, SCSs could have important localizing significance in the clinical decision-making process.
Our results showed that SCSs had a high probability of originating from the same epileptogenic site as clinical seizures and provided useful localizing information in 24 of 27 patients (88.9%). These findings were concordant with those of a previous study wherein SCSs had supplementary value in 85% of patients during the diagnostic process [8]. Sperling et al. hypothesized that the presence of SCSs represented the preservation of the brain’s ability to restrict seizure development and restrain propagation [4]. Therefore, patients with SCSs were more likely to become seizure-free after temporal lobectomy than those without SCSs [4]. Another study suggested that patients with complete colocalization between SCSs and clinical seizures had more than twice the seizure-free outcome rate than patients with partial or contradicted colocalization in TLE and extra-TLE [5]. The impact of SCS localization on surgical outcome highlights the conformity between SCSs and clinical seizures. A recent study showed that all patients with SCSs that propagated to different regions beyond their onset zone had poor surgical outcomes [9]. The ability and the tendency of SCSs to spread were also correlated with surgical outcome since they might provide additional distinct epileptic networks with the potential to affect surgical outcome [9]. All the aforementioned studies emphasized the potential importance of SCSs, despite the lack of distinct symptoms, for more accurate prediction of prognosis when evaluating epilepsy surgery. In this study, we found that patients aged <18 years were more likely to exhibit SCSs during VEEG monitoring compared to those aged 18 years. This finding was in accordance with previous evidence of a higher occurrence of SCSs in young subjects with epi-
Please cite this article in press as: Wang S et al. Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring. J Clin Neurosci (2017), http://dx.doi.org/10.1016/j.jocn.2017.06.071
4
S. Wang et al. / Journal of Clinical Neuroscience xxx (2017) xxx–xxx
lepsy [2,8]. SCSs were reported in 32 children (17.0%) among 188 patients with epilepsy, who were younger than the patients without SCSs [2]. One study found SCSs in 59 of 322 (18%) children with focal epilepsy, which was also higher than the results of our study, which included adults and children [8]. Additionally, a higher frequency of SCSs was described in newborns or children with critical illness [2]. SCSs in young children might sometimes help locating the epileptogenic focus when patients and physicians have difficulties in recognizing the performance associated with seizures or interpreting a polluted EEG manifestation due to severe movement artifacts [14]. SCSs were also more common in patients with bilateral IEDs [3]. IEDs induce long-term changes in synaptic connections between neurons and guide the development of neuronal circuits related to initiation of spontaneous seizures [15]. The management of an unprovoked first seizure in adults demonstrated that IEDs were associated with a relative increase in seizure recurrence rate at 1–5 years after the first seizure [15]. Therefore, it was unsurprising that IEDs could induce seizures. Bilateral IEDs and SCSs in mesial temporal lobes have a tendency to occur simultaneously according to case reports [7]. Published data has shown that bilateral IEDs in lateral TLE appeared in only 14% of patients compared, with mesial TLE where bilateral IEDs were present in 42% of patients [16]. Most SCSs in TLE emanated from mesial temporal structures to which they remained confined and seldom spread to another neocortex [4]. Thus, SCSs and bilateral IEDs tend to appear in regions of the mesial temporal lobe, a restricted circuit for epileptogenesis poorly connected with other brain regions [4,17]. It appears likely that seizure activity emanating from specific regions of the mesial TLE may not produce any obvious behavioral effects [17]. However, we did not differentiate neocortical from mesial TLE in this study because a majority of the patient population did not undergo surgery to identify the focus more accurately. Further studies are therefore needed in this context. Several limitations apply to this study. It is a retrospective study and our results may over- or underestimate the frequency of SCSs in patients with TLE, since this analysis was not performed on a random population-based sample of patients with TLE, and time of VEEG monitoring may have been inadequate to discover all SCSs. Moreover, we diagnosed TLE based on complex partial seizures, VEEG, and imaging data, whereas the gold standard of TLE diagnosis is based on the pathology or long-term surgical outcome. 5. Conclusion In summary, SCSs are not uncommon in patients with TLE, particularly those aged <18 years and those with bilateral IEDs, during VEEG monitoring, and merit as much consideration as clinical seizures for guiding the surgical decision and advising patients about prognosis. Since not all SCSs could be discovered within the first 24 h of VEEG monitoring, it is appropriate to extend the monitoring time for their detection to improve the accuracy of seizure frequency estimation in the future.
Conflict of interest statement None. Acknowledgments This work was supported by grants from the National Natural Science Foundation of China [grant numbers 81401071, 81300991, 81671283], and the National Natural Science Foundation of Zhejiang Province [grant number Y16H090030]. References [1] Jin B, Wang S, Yang L, Shen C, Ding Y, Guo Y, Wang Z, Zhu J, Wang S, Ding M. Prevalence and predictors of subclinical seizures during scalp video-EEG monitoring in patients with epilepsy. Int J Neurosci 2016:1–8. [2] Akman CI, Montenegro MA, Jacob S, Eck K, McBrian D, Chiriboga CA, Patterson MC. Subclinical seizures in children diagnosed with localization-related epilepsy: clinical and EEG characteristics. Epilepsy Behav 2009;16:86–98. [3] Kanazawa K, Matsumoto R, Shimotake A, Kinoshita M, Otsuka A, Watanabe O, Tanaka K, Takahashi R, Ikeda A. Persistent frequent subclinical seizures and memory impairment after clinical remission in smoldering limbic encephalitis. Epileptic Disord 2014;16:312–7. [4] Sperling MR, O’Connor MJ. Auras and subclinical seizures: characteristics and prognostic significance. Ann Neurol 1990;28:320–8. [5] Zangaladze A, Nei M, Liporace JD, Sperling MR. Characteristics and clinical significance of subclinical seizures. Epilepsia 2008;49:2016–21. [6] Bridgman PA, Malamut BL, Sperling MR, Saykin AJ, O’Connor MJ. Memory during subclinical hippocampal seizures. Neurology 1989;39:853–6. [7] Tafti BA, Mandelkern M, Berenji GR. Subclinical seizures as a pitfall in 18F-FDG PET imaging of temporal lobe epilepsy. Clin Nucl Med 2014;39:819–21. [8] Velkey A, Siegler Z, Janszky J, Duray B, Fogarasi A. Clinical value of subclinical seizures in children with focal epilepsy. Epilepsy Res 2011;95:82–5. [9] Farooque P, Duckrow R. Subclinical seizures during intracranial EEG recording: are they clinically significant? Epilepsy Res 2014;108:1790–6. [10] Babb TL, Wilson CL, Isokawa-Akesson M. Firing patterns of human limbic neurons during stereoencephalography (SEEG) and clinical temporal lobe seizures. Electroencephalogr Clin Neurophysiol 1987;66:467–82. [11] Schmeiser B, Hammen T, Steinhoff BJ, Zentner J, Schulze-Bonhage A. Long-term outcome characteristics in mesial temporal lobe epilepsy with and without associated cortical dysplasia. Epilepsy Res 2016;126:147–56. [12] Tao JX, Ray A, Hawes-Ebersole S, Ebersole JS. Intracranial EEG substrates of scalp EEG interictal spikes. Epilepsia 2005;46:669–76. [13] Stead M, Bower M, Brinkmann BH, Lee K, Marsh WR, Meyer FB, Litt B, Van Gompel J, Worrell GA. Microseizures and the spatiotemporal scales of human partial epilepsy. Brain 2010;133:2789–97. [14] Nordli Jr DR, Bazil CW, Scheuer ML, Pedley TA. Recognition and classification of seizures in infants. Epilepsia 1997;38:553–60. [15] Krumholz A, Shinnar S, French J, Gronseth G, Wiebe S. Evidence-based guideline: management of an unprovoked first seizure in adults: report of the Guideline Development Subcommittee of the American Academy of Neurology and the American Epilepsy Society. Neurology 2015;85:1526–7. [16] Janszky J, Rasonyi G, Clemens Z, Schulz R, Hoppe M, Barsi P, Fogarasi A, Halasz P, Ebner A. Clinical differences in patients with unilateral hippocampal sclerosis and unitemporal or bitemporal epileptiform discharges. Seizure 2003;12:550–4. [17] Asadi-Pooya AA, Rostami C, Rabiei AH, Sperling MR. Factors associated with tonic-clonic seizures in patients with drug-resistant mesial temporal epilepsy. J Neurol Sci 2015;359:452–4.
Please cite this article in press as: Wang S et al. Clinical value and predictors of subclinical seizures in patients with temporal lobe epilepsy undergoing scalp video-EEG monitoring. J Clin Neurosci (2017), http://dx.doi.org/10.1016/j.jocn.2017.06.071