R. Madhavan Nayar Center for Comprehensive Epilepsy Care, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, Kerala, India
Address correspondence to Kurupath Radhakrishnan, Department of Neurology, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Trivandrum, Kerala 695 011, India. E-mail: firstname.lastname@example.org
In contrast to the well-recognized association between acute symptomatic seizures and neurocysticercosis, the association between antiepileptic drug (AED)–resistant epilepsy and calcified neurocysticercosis lesions (CNLs) is poorly understood. We studied the association between AED-resistant epilepsy and CNLs, including the feasibility and outcome of resective surgery.
From the prospective database maintained at our epilepsy center, we reviewed the data of all patients with AED-resistant epilepsy who underwent presurgical evaluation from January 2001 to July 2010 and had CNL on imaging. We used clinical, neuroimaging, and interictal, ictal, and intracranial electroencephalography (EEG) findings to determine the association between CNL and epilepsy. Suitable candidates underwent resective surgery.
Forty-five patients fulfilled the inclusion criteria. In 17 patients, CNL was proven to be the causative lesion for AED-resistant epilepsy (group 1); in 18 patients, CNL was associated with unilateral hippocampal sclerosis (HS; group 2); and in 10 patients, CNLs were considered as incidental lesions (group 3). In group 1 patients, CNLs were more common in frontal lobes (12/17), whereas in group 2 patients, CNLs were more commonly located in temporal lobes (11/18; p = 0.002). Group 2 patients were of a younger age at epilepsy onset than those in group 1 (8.9 ± 7.3 vs. 12.6 ± 6.8 years, p = 0.003). Perilesional gliosis was more common among patients in group 1 when compared to group 3 patients (12/17 vs. 1/10; p = 0.006). Fifteen patients underwent resective surgery. Among group 1 patients, four of five became seizure-free following lesionectomy alone. In group 2, four patients underwent anterior temporal lobectomy (ATL) alone, of whom one became seizure-free; five underwent ATL combined with removal of CNL (two of them after intracranial EEG and all of them became seizure-free, whereas one patient underwent lesionectomy alone and did not become seizure-free.
In endemic regions, although rare, CNLs are potential cause for AED-resistant and surgically remediable epilepsy, as well as dual pathology. Presence of perilesional gliosis contributes to epileptogenicity of these lesions. For those patients with CNL and HS, resection of both lesions favors better chance of seizure-free outcome.
In contrast to the well-recognized association between acute symptomatic seizures and neurocysticercosis, the association between antiepileptic drug (AED)–resistant epilepsy and neurocysticercosis is controversial (Nash et al., 2004). A majority of the patients with acute symptomatic seizures during the active stage of the neurocysticercosis experience remission in the next 3–6 months, along with the disappearance of the lesions (Rajshekhar, 2001; Singh et al., 2001; Rajshekhar & Jeyaseelan, 2004). However, few neurocysticercosis lesions heal with calcifications, and these patients are at a higher risk of subsequent seizures (Murthy & Reddy, 1998; Rajshekhar & Jeyaseelan, 2004). Previous studies have shown that majority of the patients with calcified neurocysticercosis lesions (CNLs) have well-controlled epilepsy (Cukiert et al., 1994; Murthy & Reddy, 1998; Singh et al., 2000). Similarly, few studies from Latin American countries have shown that most, if not all, CNLs in patients with AED-resistant epilepsy are incidental lesions (Leite et al., 2000; Velasco et al., 2006). On the basis of these findings, the CNLs are considered to have a low epileptogenic potential and their causative role in AED-resistant epilepsy is doubted (Caprio et al., 1998; Singh et al., 2013). We encountered several patients with AED-resistant epilepsy and CNL, where the CNL was found to be responsible for AED-resistant epilepsy. Establishing or refuting a cause–effect relationship between AED-resistant epilepsy and CNL is important, especially in regions that are endemic for neurocysticercosis, as CNLs are commonly encountered in patients who are undergoing presurgical evaluation and pose a diagnostic and therapeutic dilemma. To clarify the role of CNL in AED-resistant epilepsy, we studied consecutive patients with AED-resistant epilepsy and CNL on imaging. The objectives of our study were the following: (1) to assess the causative role of CNL in AED-resistant epilepsy; (2) to study the factors that contribute to AED-resistant epilepsy in patients with CNL; and (3) to define the role of resective surgery in patients with CNL and AED-resistant epilepsy.
Patients and Methods
We have recently described our methods for the selection and evaluation of the patients with AED-resistant epilepsy and CNL (Rathore et al., 2012). Briefly, all the patients evaluated for AED-resistant epilepsy from January 2001 to July 2010 at our center and found to have calcified lesion/s on imaging studies were selected for this study. Patients with calcified lesions caused by tumors and vascular malformations were excluded. We defined AED-resistant epilepsy as persistent seizures (one or more seizures per month) despite two adequate and tolerated AED monotherapy trials and at least one duotherapy trial. We diagnosed CNL on the basis of previously described diagnostic criteria: the presence of solid, dense, supratentorial calcification of ≤10 mm in diameter in a resident of an endemic region (Del Brutto et al., 2001). All patients underwent detailed clinical evaluation, brain imaging, and long-term video-electroencephalography (VEEG) monitoring as described by us previously, whereas additional tests were undertaken when indicated as a part of presurgical evaluation (Sylaja et al., 2004; Chemmanam et al., 2009; Rathore et al., 2011).
From the prospective data base maintained at our center, we specifically gathered the following clinical information: age at onset of seizures, age at initial precipitating injury (IPI) and its nature, duration of epilepsy, clinical semiology of all seizure types, and any change in seizure semiology during the course of illness. We defined IPI as any illness leading to seizure(s) and/or impaired consciousness (Rathore et al., 2009). We also noted the details of previous treatment for acute symptomatic seizures.
All patients underwent long-term VEEG monitoring utilizing standard 10-20 system of electrode placement with additional anterior temporal (T1, T2) electrodes (Chemmanam et al., 2009). Sphenoidal electrodes were inserted in patients with temporal lobe epilepsy and nonlateralizing or discordant electroclinical data (Cherian et al., 2012). At least two habitual seizures were recorded for every patient. The presence and distribution of interictal epileptiform discharges (IEDs) during wakefulness and sleep was assessed by visual analysis of a minimum of 10 min of data every hour during a 24-h period by two study epileptologists (CR and KR). Only definite spikes or sharp waves were considered as IEDs. The IEDs in anterior and midtemporal electrodes (F7, F8, T1, T2, T3, and T4) were defined as temporal IEDs.
CT and MRI evaluation
All patients underwent plain and contrast enhanced computed tomography (CT) scan and a 1.5 T magnetic response imaging (MRI) utilizing a standard protocol as described by us in detail previously (Sylaja et al., 2004). Two neuroradiologists (BT and CK), attached to the epilepsy program and blinded to the clinical data, independently reviewed all the images. Any discordance was cleared by consensus opinion. The following details were specifically assessed: number and site of calcified lesion(s); size of lesions; presence of surrounding gliosis as determined on T2-weighted and fluid-attenuated inversion recovery (FLAIR) sequences; and presence of any other potentially epileptogenic lesion including hippocampal sclerosis (HS). We defined HS as the presence of unequivocal atrophy of the hippocampus with a corresponding increase in the signal on T2-weighted and FLAIR sequences.
Additional tests and interpretation of data
Initial diagnosis of the epilepsy syndrome and localization of the ictal onset was based on the information obtained from the clinical evaluation, interictal and ictal EEG, and imaging studies. If the initial evaluation provided unequivocal localizing and concordant information, enabling a hypothesis regarding the probable epileptogenic zone, then further evaluation was not performed. In the event of nonlocalizing or discordant data, further investigations including interictal and ictal single photon emission computed tomography (SPECT), and intracranial monitoring were undertaken. Final decisions about the localization and surgical resection were made during the comprehensive patient management conference, by a team comprising epileptologists, neuroradiologists, and neurosurgeons.
The information obtained from presurgical evaluation was used to establish the relationship between CNL and epilepsy: whether the CNL was the cause of epilepsy or simply an incidental lesion or whether the association between the two was uncertain. If a clear hypothesis could be generated about the possible epileptogenic zone, then these patients were subjected to resective surgery. For anterior temporal lobectomy, all patients underwent the same technique of resecting the lateral temporal neocortex and mesial temporal structures under general anesthesia by the same neurosurgeon (MA), as described elsewhere in detail (Sylaja et al., 2004). For calcified lesions, surgery consisted of removal of the lesion along with one centimeter of surrounding margin, also guided by the intraoperative electrocorticography. All surgical specimens were histopathologically examined by the pathologist, the results of which were reviewed for this study. We defined HS as the loss of neuronal cell population of 30% or more in the CA1 sector of the hippocampal formation with or without neuronal loss and gliosis involving other mesial temporal structures (Radhakrishnan et al., 2007). For patients who could not be selected for surgery, the reasons for the same were noted.
Follow-up and seizure outcome
Patients who underwent resective surgery were followed up at 3 months and at 1 year following surgery and then yearly afterwards. Seizure outcome was assessed at each year and was classified as seizure-free (free of all seizures and auras) or not seizure-free. For each patient, we classified the seizure outcome at 2 years after surgery as well as for the entire postoperative period. All the operated patients had a minimum follow-up of 2 years. None of our patients, who remained seizure-free during the first two postoperative years, had a seizure recurrence subsequently.
We summarized the quantitative data as mean ± standard deviation (SD) and qualitative data as percentages. Different clinical and radiologic characteristics between the groups were compared using Student's t-test and Fisher's exact test. A p-value of <0.05 was considered as statistically significant.
From January 2001 to July 2010, a total of 3,895 patients with AED-resistant epilepsy underwent presurgical evaluation. Of these, 51 patients had AED-resistant epilepsy and calcified granulomatous lesions on imaging studies. Six patients, diagnosed to have calcified lesions related to previous central nervous system tuberculosis, were excluded from this study. The remaining 45 patients (24 male, 21 female) fulfilled the inclusion criteria for CNL.
Based on the presurgical evaluation data and postsurgery outcome in surgically treated patients, we divided the cohort into three groups: patients in whom CNL was the only imaging abnormality and it was considered as the causative lesion for AED-resistant epilepsy (group 1, n = 17); CNL associated with unilateral HS (group 2, n = 18); and patients in whom CNLs were considered as incidental lesions (group 3, n = 10).
Group 1: CNL as the causative lesion for AED-resistant epilepsy
Among the 17 patients in this group, 16 had a single CNL, whereas one patient had two lesions. Mean age at the time of evaluation was 23.9 ± 7.9 years, mean age at epilepsy onset was 12.6 ± 6.8 years, and mean duration of epilepsy was 11.5 ± 5.8 years. Twelve lesions were in frontal lobes, four were in the parietooccipital region, and one each was noted in posterior superior temporal and fusiform gyri. Twelve (66.7%) of these lesions had surrounding gliosis on T2-weighted and FLAIR MRI sequences. Eleven patients in this group were selected for resective surgery: five patients (two frontal, two parietal, and one parietooccipital junction lesions) underwent surgery (Table 1), whereas the other six are awaiting surgery. Three patients with lesions close to eloquent cortex (one over hand motor area and two near to the language areas) were advised against surgery, and the other three patients opted for medical management. One patient with right parietal CNL was selected for surgery after intracranial monitoring as initial noninvasive data was nonlocalizing (patient 4, Table 1). This patient and the other three patients were completely seizure-free and aura-free during the entire postoperative follow-up period (range 2–8 years). One patient had a significant reduction in seizure frequency but did not become seizure-free during 5 years of postoperative follow-up. We have illustrated the MRI and EEG data of one patient in this group (patient 1, Table 1) through Fig. 1.
Table 1. Characteristics of patients with AED-resistant epilepsy and calcified neurocysticercal lesion, who underwent surgery
Seizure-free following intracranial EEG and lesionectomy
Left parietal, Left temporal
Seizure-free following lesionectomy
Left ATL; not seizure-free
Right ATL; not seizure-free
Right ATL; not seizure-free
Right frontal, right occipital and left temporal
Right ATL; seizure-free
Right ATL with lesionectomy; seizure-free
Left AH with lesionectomy; seizure-free
Right fusiform gyrus
Right ATL with lesionectomy; seizure-free
Right occipital and bilateral temporal
Intracranial EEG showed seizure origin from the lesion and hippocampus; seizure-free following temporal resection and lesionectomy
Bilateral temporal and left parietal
Intracranial EEG showed seizure origin from the lesion and hippocampus; seizure-free following temporal resection and lesionectomy
Right parietal, Right temporal
Lesionectomy alone; not seizure-free
Group 2: CNL associated with unilateral HS
We have previously described the characteristics of this group of patients (Rathore et al., 2012). Their mean age at the time of evaluation was 32.1 ± 10.7 years, mean age at epilepsy onset was 15.8 ± 6.7 years, and mean duration of epilepsy was 16.1 ± 7.9 years. Four patients in group 2 had a history of typical febrile seizures. Of the 18 patients in this group, 12 had a single calcified lesion, 4 had two lesions, and 2 patients had three lesions. All the patients with single lesion and three patients with two unilateral lesions had ipsilateral HS. Eleven of these 26 lesions were in the anterior temporal lobe (six within the hippocampus), and location within the temporal lobe was more common in this group as compared to group 1 (11/18 vs. 1/16, p = 0.002). First seizure occurred at a younger age in group 2 when compared to group 1 patients (8.9 ± 7.3 vs. 12.6 ± 6.8 years, p = 0.003).
Nine patients were advised surgery based on the noninvasive data, whereas eight patients were advised to undergo intracranial monitoring because an unequivocal ictal onset zone could not be determined. One patient opted for medical management. Two patients with CNL within the hippocampus (patients 10 and 11, Table 1) and one in the anterior fusiform gyrus (patient 12, Table 1) underwent anterior temporal lobectomy along with lesion removal. All three have remained seizure-free throughout the postoperative period (range 4–6 years). Four patients with HS and ipsilateral extratemporal CNL, in whom noninvasive data was suggestive of anteromesial epileptogenic zone in the form of strictly unilateral temporal IEDs and localized temporal ictal-onset zone, were subjected to anterior temporal lobectomy alone (patients 6–9, Table 1). Only one of these became seizure-free (patient 9, Table 1), whereas other patients continued to have postoperative seizures. One patient with right parietal CNL and right HS underwent lesionectomy alone based on noninvasive data and did not become seizure-free (patient 15, Table 1). Two patients, one with right occipital CNL and HS and the other with left parietal CNL and HS (patients 13 and 14, Table 1), who underwent intracranial monitoring, had ictal onsets from both the CNL and HS. They subsequently underwent anterior temporal lobectomy and lesionectomy and had remained seizure-free at 2 years of follow-up. The MRI and EEG data of one patient in group 2 (patient 11, Table 1) is illustrated in Fig. 2. In Figure 3, we have depicted data of one of our patients with right posterior temporal CNL along with right HS, who had independent right anterior and posterior temporal IEDs and documented ictal onsets from both the lesions by scalp EEG. This patient has been advised to undergo an extended right temporal lobe resection to include both the HS and CNL.
Group 3: CNL as the incidental lesion associated with AED-resistant epilepsy
Of the 10 patients in this group, 3 patients had malformations of cortical development as causative lesions for seizures, 2 had predominantly nonepileptic events after the initial generalized seizures, and 2 other patients had mental retardation and seizures from early childhood. One patient was diagnosed with benign rolandic epilepsy and the other two had cryptogenic extratemporal epilepsy unrelated to CNL. As compared to patients in group 1, patients in group 3 had a lower incidence of gliosis around the lesion (12/17 vs. 1/10; p = 0.006). Group 3 patients were managed on an individual basis.
Intraoperative electrocorticography showed spikes in relation to the lesion in all the five group 1 patients who underwent lesionectomy alone. In three of these patients and the other two patients who underwent extraoperative electrocorticography, focal rhythmic spikes discharges (≥1/s) and focal bursts of polyspikes (spikes at frequency of 10 Hz lasting for minimum of 0.4 s) were noted in relation to the lesion. In two patients (patients 1 and 3, Table 1), spikes were also noted distant to the lesion, which persisted after the lesionectomy.
Histopathologic examination of the lesions showed dense calcifications surrounded by chronic inflammatory infiltrates and reactive gliosis in surrounding brain parenchyma. All the hippocampal specimens showed features of classical hippocampal sclerosis without any features of inflammation (Radhakrishnan et al., 2007).
The evidence for the epileptogenicity of CNL is mostly circumstantial. In endemic countries, prevalence of calcified lesions is higher in patients with epilepsy than in controls (Garcia-Noval et al., 2001; Fleury et al., 2003; Garcia et al., 2003). Similarly, the edema around the CNL following a seizure is considered to be one form of evidence that the lesion is responsible for seizures (Antoniuk et al., 2001; Nash et al., 2008).
There is also a relative dearth of data regarding the causative role of CNL in AED-resistant epilepsy. In a cross-sectional study of 512 patients with AED-resistant epilepsy from Brazil, calcifications were found in 27% patients, mostly in association with other epileptogenic lesions (Velasco et al., 2006). Isolated calcified lesions were found in only eight patients (1.6%) and were considered to be the cause of refractory epilepsy in only two of them (Velasco et al., 2006). Another study from Brazil also did not find any difference in the clinical presentation and posttemporal lobectomy seizure outcome in patients of HS with and without CNL, indicating that the majority of the CNLs in these patients were incidental (Leite et al., 2000). Similarly, a study from India reported CNL in 29 (0.7%) of 4,452 CT scans performed for nonseizure causes (Singh et al., 2000). These studies cast doubts on the epileptogenicity of CNL.
Our results indicate that CNLs are a potential cause of AED-resistant epilepsy. In 17 of our patients, CNLs were found to be directly responsible for seizures and were associated with HS in another 18 patients. The direct evidence of epileptogenicity was further provided by the complete seizure freedom following lesion resection alone in four of five patients. Two additional patients had evidence of seizure origin in relation to CNL on intracranial EEG. Our data also show that CNL are indeed a very rare cause for AED-resistant epilepsy, as these lesions were found to be associated with AED-resistant epilepsy in only 35 patients (35/3,895; 0.9%) who had undergone presurgical evaluation at our epilepsy center. However, it needs to be mentioned that more than two thirds of all the patients evaluated for AED-resistant epilepsy at our center were from the state of Kerala, where neurocysticercosis is almost nonexistent (Kannoth et al., 2009). On the other hand, only 4 of the 45 patients with CNL were native residents of Kerala (who had in the past resided in endemic regions), and others were from regions in India known to be endemic for neurocysticercosis. This suggests that prevalence of AED-resistant epilepsy associated with CNL may be higher in endemic areas, the fact that needs to be further studied in endemic regions.
Similar to many previous studies, our results also confirm that not all the CNLs are epileptogenic. Presence of the gliosis around CNLs as demonstrated on T2-weighted and FLAIR MRI sequences was associated with a higher chance of CNL being the cause for epilepsy. In a previous study, presence of perilesional gliosis on magnetization transfer MRI has been shown to be associated with a higher risk of seizure recurrence on AED withdrawal (Pradhan et al., 2000). Perilesional gliosis probably contributes to the epileptogenicity of CNL, and its presence can help in identifying a causative rather than incidental lesion. We did not find any association between the lesion location and AED-resistant epilepsy. The majority of the single CNLs were located in the frontal or parietal lobes, a finding similar to that of the previous studies (Cukiert et al., 1994; Murthy & Reddy, 1998; Singh et al., 2000).
The pathogenesis of seizures in CNL is uncertain. A recent study has found that the patients with CNL and seizures had a higher degree of blood–brain barrier breakdown and inflammation surrounding the CNL as compared to those without seizures (Gupta et al., 2012). This was further correlated with significant increase in serum levels of matrix metalloproteinase-9 (MMP-9) and MMP-9 (R279Q) gene polymorphisms in symptomatic subjects. Taken together, these data indicate that genetic, parasitic, and environmental factors contribute to the epileptogenesis in CNL, with more intense inflammation and subsequent perilesional gliosis associated with a higher risk of seizures and epilepsy. Genetic predisposition and differences in parasitic factors may also explain the higher incidence of symptomatic CNL on the Indian subcontinent than in Latin American countries (Singh, 1997).
About 40% of our patients with CNL had associated HS. As we have shown earlier, presence of CNL in patients with HS should be considered as potential dual pathology, especially if the CNLs are located within the temporal lobes, and due care should be taken during presurgical evaluation to localize the ictal-onset zone(s) (Rathore et al., 2012).
There is a paucity of reports detailing the surgical treatment of AED-resistant epilepsy caused by CNL. In a study from New Delhi, India, six patients with CNL who underwent surgical resection had Engel class I outcome (Chandra et al., 2010). Four of these patients with lesions in the medial temporal lobe had additional anterior temporal lobectomy. In a report of six patients from South Africa, four of whom underwent lesionectomy alone had seizure-free outcome (Butler, 2005). Our series of 15 surgically treated patients appears to be the largest reported yet (Table 1). These included 11 patients who underwent lesionectomy alone (n = 6; one with associated HS) or lesionectomy along with temporal resection (n = 5). Nine of these 11 patients were seizure-free following surgery, whereas one patient had a significant improvement. All these results taken together indicate that drug-resistant epilepsy caused by CNLs is a potentially surgically remediable syndrome with a very good chance of seizure freedom following focal resection. Epileptogenicity in these patients is usually restricted to the perilesional tissue. However, as our data indicate, these patients pose unique challenges during presurgical evaluation, particularly if associated with HS. Moreover, small calcified lesions can be easily missed on MRI and it is advisable to include either susceptibility-weighted MRI images or CT scan as a part of imaging protocol in endemic regions (Saini et al., 2009).
Our data has certain limitations. Our results apply to a highly selected group of patients who underwent presurgical evaluation for drug-resistant epilepsy, and these results cannot be generalized. Mere presence of CNL should not be taken as the cause of epilepsy and due care should be exercised to establish the definite cause–effect relationship. Our results merely indicate that AED-resistant epilepsy can occur as a result of CNL and do not provide the true estimate of AED-resistant epilepsy caused by CNL. Longitudinal long-term follow-up studies of patients with acute neurocysticercosis in endemic areas are required to gauge the true estimate of the AED-resistant epilepsy caused by CNL. Many of our patients did not undergo surgery or invasive monitoring to define the epileptogenic zone. However, in patients with CNL alone, there was enough evidence to establish the cause–effect relationship and select the patients for surgery. This may not hold true for patients with associated HS, but our data does indicate a more than incidental association between CNL and HS, which needs to be studied further.
None of the authors has any conflict of interest to disclose. We confirm that we have read the Journal's position on issues involved in ethical publications and affirm that this report is consistent with those guidelines.