Epilepsy in TSC: Certain etiology does not mean certain prognosis




Prevalence and long-term outcome of epilepsy in tuberous sclerosis complex (TSC) is reported to be variable, and the reasons for this variability are still controversial.


We reviewed the clinical characteristics of patients with TSC who were regularly followed since 2000 at the San Paolo Multidisciplinary Tuberous Sclerosis Centre in Milan, Italy. From patient charts we collected data about age at epilepsy onset, seizure frequency and seizure type, history of infantile spasms (IS), epileptic syndrome, evolution to refractory epilepsy or to seizure freedom and/or medication freedom, electroencephalography (EEG) features, magnetic resonance imaging (MRI) findings, cognitive outcome, and genetic background.

Key Findings

Among the 160 subjects (120 adults and 40 children), 116 (72.5%) had epilepsy: 57 (35.6%) were seizure-free, and 59 (36.9%) had drug-resistant epilepsy. Most seizure-free patients had a focal epilepsy (89.5%), with 54.4% of them drug resistant for a period of their lives. Epilepsy onset in the first year of life with IS and/or focal seizures was characteristic of the drug-resistant group of patients, as well as cognitive impairment and TSC2 mutation (p < 0.05). A small group of patients (7 patients, 4.4%) experienced a seizure only once; all of them had normal cognition.


Although epilepsy management can be challenging in TSC, more than one third of patients had their seizures controlled: through monotherapy in 56% and by polytherapy in 32%. Moreover, 12% of the patients became seizure-free and were off medication. Identifying predictive features of epilepsy and cognitive outcome can ensure better management for patients with TSC and delineate genotype–phenotype correlations.

Tuberous sclerosis complex (TSC) is an autosomal dominant, multisystemic disease with variable expression.

TSC occurs in at least one in 6,000 individuals (Crino et al., 2006) and is caused by a mutation in either the TSC1 or TSC2 gene. Roughly two thirds of TSC cases reported in mutational and epidemiologic studies are sporadic (simplex), whereas the remaining cases are familial (Jones et al., 1999).

The TSC1/TSC2 complex plays an important role during cortical development and growth control through the hyperactivation of the downstream mammalian target of rapamycin (mTOR) pathway and the associated kinase signaling cascades and translational factors. Therefore, mutations of TSC genes potentially induce abnormal cell morphology, imbalance in excitation and inhibition, as well as alterations in network properties, providing a likely underlying substrate for epilepsy (Moavero et al., 2010).

Neurologic phenotype of TSC is highly variable, ranging from normal to severe, and consisting of refractory epilepsy, intellectual impairment, and psychiatric comorbidity. Although a clear phenotypic/genotypic correlation has not been found as of yet, it is globally accepted that patients with TSC1 gene mutations show a milder phenotype than patients carrying TSC2 gene mutations (Jentarra et al., 2011).

TSC is one of the leading genetic causes of epilepsy, with a high incidence (about 85%) of patients with TSC affected by seizures, often refractory to antiepileptic treatment (Curatolo et al., 2008).

Nevertheless the prevalence and long-term outcome of epilepsy in TSC has not been widely investigated.

A wide population of patients with epilepsy in TSC was investigated retrospectively: the authors reported a wide spectrum of severity for epilepsy in TSC, ranging from seizure-free patients to severe drug-resistant epilepsies (Chu-Shore et al., 2010).

The present study aims to review the clinical characteristics of epilepsy in the population of adult and pediatric patients followed regularly since 2000 at the San Paolo Hospital Tuberous Sclerosis Center in Milan, Italy.


We reviewed the clinical records of all patients with TSC, regularly followed at San Paolo Hospital Tuberous Sclerosis Center, San Paolo Hospital, Milan, Italy, from January 2000 to June 2012. Of 160 individuals with an established diagnosis of TSC, according to the International criteria (Roach et al., 1998), 116 had epilepsy. Only patients with at least 1 year of follow-up participated in this study.

We collected data from patient charts about age at epilepsy onset, seizure frequency and seizure type, history of IS, epileptic syndrome, evolution to refractory epilepsy or to seizure freedom and/or medication freedom, electroencephalography (EEG) features, magnetic resonance imaging (MRI) findings, cognitive outcome, and genetic background. Results of presurgical evaluation for drug-resistant cases were also assessed, when performed.

Seizure types were classified according to clinical semiology provided by description from a witness or when possible by video-EEG recording of the episode. Refractory epilepsy was defined as uncontrolled seizures after at least two first-line anticonvulsant medication trials (Kwan et al., 2009). Patients were considered seizure-free if they were without clinical seizures for at least 1 year, using the last clinical visit documenting seizure status as the end point of follow-up.

Genetic testing of the TSC1 and TSC2 genes was performed on DNA extracted from peripheral lymphocytes using a standard protocol for all patients and their first-degree relatives. Polymerase chain reaction (PCR) was used to amplify all exons of TSC1 and TSC2 genes. Negative patients for the first level of investigation were evaluated with Multiple Ligation-dependent Probe Amplification test. Patients in which genetic analysis was inconclusive were classified as no mutation identified (NMI). Among 160 subjects (120 adults and 40 children), 31 (19.4%) were familial cases (17 TSC1, 13 TSC2, and 1 NMI).

We divided patients into five groups according to intelligence quotient (IQ): (1) normal IQ with IQ > 85; (2) borderline intellectual functioning (BIF) with IQ from 84 to 71; (3) mild intellectual disability (ID) with IQ from 70 to 55; (4) moderate ID with IQ from 54 to 40, and (5) severe ID with IQ < 40 (American Psychiatric Association, 2000). Education was measured in years of schooling. As part of their clinical management, patients were evaluated through a psychiatric interview in order to assess possible Axis I and II disorders (American Psychiatric Association, 2000).

All data collected were transferred into an electronic database and processed using the Statistical Package for the Social Sciences (SPSS, IBM, Chicago, IL, U.S.A.) for Macintosh version 19.0. Continuous variables were presented as mean and standard deviation (SD). Patients (seizure-free vs. drug resistant vs. no epilepsy) were compared on sociodemographic and clinical characteristics, using chi-square test for categorized variables and not normally distributed variables and one-way analyses of variance (ANOVA) with Bonferroni post hoc for variables with a normal distribution. In particularly, significance was determined among three groups of patients simultaneously: seizure-free versus drug-resistant + seizure-free versus no epilepsy + drug-resistant versus no epilepsy.

Variables (seizure-free vs. drug resistant) with a normal distribution were analyzed using independent sample two-tailed Student t-tests. Results were considered statistically significant for p-values < 0.05 (two-sided). The institutional review board of San Paolo Hospital, Milan, approved the study.


Among the 160 patients (64 [40.0%] male, 96 [60.0%] female; mean age 29.6, range 1–74 years) presently followed for TSC at San Paolo Hospital, 116 had epilepsy (72.5%), whereas 7 (4.4%) presented only one isolated seizure. Table 1 shows demographic and clinical characteristics of the entire TSC population, and Table 2 refers to the TSC patients with epilepsy, respectively.

Table 1. Demographic and clinical characteristics
 Seizure-freeDrug-resistantNo epilepsyp-Value
N = 57 (35.6%)N = 59 (36.9%)N = 37 (23.1%)
  1. N.e., not evaluated; SD, standard deviation; BIF, borderline intellectual functioning, ID, intellectual disability; NS, neurosurgery; p-values were obtained from analyses of chi-square test for categorized variables and not normally distributed variables and using one-way analyses of variance (ANOVAs) for variables with a normal distribution.

  2. Significant differences are highlighted in bold (p < 0.05). Numbers in bold indicate statistically significant differences among three groups of patients simultaneously: seizure-free versus drug-resistant + seizure-free versus no epilepsy + drug-resistant versus no epilepsy.

Age (years)    
Mean 31.6 24.1 36.8 <0.05
Level of education (years)    
Mean 8.9 7.6 10.7 <0.05
Genetic analysis (%)    
TSC119 (33.3)11 (18.6)15 (40.5)>0.05
TSC2 30 (52.7) 39 (66.1) 13 (35.1) <0.05
No mutation identified4 (7)3 (5.1)9 (24.4)>0.05
Ongoing4 (7)6 (10.2)  
Familial cases (%)10 (17.5)7 (11.9)11 (29.7)>0.05
Cognitive profiles (%)    
Normal IQ 25 (43.9) 12 (20.3) 35 (94.6) <0.05
BIF10 (17.5)5 (8.5) >0.05
Mild ID 7 (12.3) 12 (20.3) 2 (5.4) <0.05
Moderate ID8 (14) 9 (15.3) >0.05
Severe ID 7 (12.3) 21 (35.6)   <0.05
Neurologic examination (%)    
Normal50 (87.7)42 (71.2)36 (97.3) 
Hemiparesis3 (5.3)7 (11.9)  
Cranial nerves1 (1.7)1 (1.7)  
Tetraparesis 3 (5.1)  
Extrapyramidal 1 (1.7)  
N.e.3 (5.3)5 (8.5)1 (2.7) 
Cortical tubers (%)    
Yes56 (98.2)59 (100)33 (89.2) 
No004 (10.8)  
N.e.1 (1.8)   
No. of tuber from 1 to 69 (15.8)4 (6.8)14 (42.4)>0.05
 No. of tuber >6 43 (75.4) 52 (88.1) 17 (51.5) <0.05
N.e.5 (8.8)3 (5.1)2 (6.1) 
Subependymal nodules (%)    
 Yes 43 (75.4) 51 (86.4) 25 (67.6) <0.05
SEGA (%) 7 (12.3) 13 (22) 1 (2.7) <0.05
NS for SEGA (%)2 (3.5)5 (8.5)1 (2.7) 
Mean age at NCH13.516.244>0.05
Range (years) 6–212–34  
EEG (%)    
Normal20 (35.1)6 (10.2)23 (62.2) 
Focal16 (28.1)14 (23.7)4 (10.8) 
Focal + slow4 (7)14 (23.7  
Slow activity12 (21.1)4 (6.8)3 (8.1) 
Multifocal anomalies1 (1.7)10 (16.9)  
Multifocal + slow 2 (3.4)  
Diffuse1 (1.7)4 (6.8)  
Generalized anomalies 4 (6.8)  
N.e.3 (5.3)1 (1.7)7 (18.9) 
Psychiatric disorders (%) 15 (26.3) 23 (38.9) 8 (21.6) <0.05
Age at TSC diagnosis    
 Mean 13.2 7.6 23.8 <0.05
Range (years)prenatal-74prenatal-451–48 
Table 2. Clinical characteristics of population with epilepsy
N = 57 (35.6%)N = 59 (36.9%)
  1. IS, infantile spasms; CBZ, carbamazepine; VPA, valproic acid; PB, phenobarbital; OXC, oxcarbazepine; PHT, phenytoin; PRM, primidone; GBP, gabapentin; LEV, levetiracetam; OXA, oxcarbazepine; ZNS, zonisamide; N.e., not evaluated.

  2. Statistically significant for p-values < 0.05 (bold).

Type of epilepsy (%)   
Focal epilepsies51 (89.5)36 (61) 
IS1 (1.7)  
Lennox-Gastaut syndrome 9 (15.3) 
Multifocal 8 (13.5) 
Undetermined1 (1.7)6 (10.2) 
Not classified4 (7.1)  
Infantile spasms (%)20 (35.1)32 (54.2)>0.05
Epilepsy onset (%)   
 Mean 5.3 3.3 <0.05
Range2 months–41 years2 days–33 years 
<1 year27 (47.4)39 (66.1)>0.05
1–3 years13 (22.8)11 (18.6)>0.05
3–17 years11 (19.3)5 (8.5)>0.05
≥18 years6 (10.5)4 (6.8)>0.05
Seizures type at onset (%)   
Focal34 (59.7)32 (54.2)>0.05
Infantile spasms12 (21.1)22 (37.3)>0.05
Focal + infantile spasms5 (8.8)4 (6.8)>0.05
Absence1 (1.7)  
Tonic–clonic1 (1.7)  
N.e.4 (7)1 (1.7)>0.05
Number of AEDs (%)   
 Monotherapy 32 (56.1) 9 (15.3) <0.05
 Polytherapy 18 (31.6) 50 (84.7) <0.05
No therapy7 (12.3)0 
CBZ12 3 
Febrile seizures (%)6 (10.5)8 (13.6)>0.05
Status (not febrile) (%) 1 (1.7) 7 (11.9) <0.05
Febrile status (%)04 (6.8) 
History of pharmacoresistant epilepsy (%)   
  Yes 30 (52.6) 47 (79.6) <0.05
  No 22 (38.6) 7 (11.9) <0.05
N.e.5 (8.8)5 (8.5) 
Recorded seizures (%)1 (1.7)7 (11.9) 
Seizure frequency (%)   
Yearly 3 (5.1) 
Monthly 16 (27.1) 
Weekly 14 (23.7) 
Daily 21 (35.6) 
Sporadic4 (7)5 (8.5) 
Epilepsy surgery    
Engel class   
Proposed lost to follow-up1  
Vagus nerve stimulation3  

The mean follow-up period was 3.5 years (range 1–29 years). In relation to epilepsy, patients were divided into four subgroups: seizure-free (57 patients, 35.6%), drug-resistant epilepsy (59 patients, 36.9%), patients with only one seizure in their lives (7 patients, 4.4%), and no epilepsy (37 patients, 23.1%).

Seizure-free patients

Seizure-free patients were in the great majority adults (45/57), with a mean age of 31.6 years (SD 16.7). Mean age at TSC diagnosis was 13.2 years (SD 15.9). Mutation analysis demonstrated a TSC2 gene mutation in most cases (52.7%) and a TSC1 gene mutation in 33.3%, and NMI in 7.0%. In 10 cases (17.5%) there was a family history of TSC (7 TSC1; 3 TSC2).

Cognitive profile revealed normal IQ in 43.9% (p < 0.05), BIF in 17.5%, mild ID in 12.3%, moderate ID in 14%, and severe ID in 12.3%. Moreover the mean level of education was 8.9 (SD 3.7) years. Neurologic examination was considered normal in 87.7% of cases, whereas a psychiatric disorder was diagnosed in 26.3% of patients.

MRI findings demonstrated cortical tubers in 98.2% (with more than six tubers in 75.4%), subependymal nodules (SENs) in 75.4%, and subependymal giant astrocytoma (SEGA) in 12.3% of patients.

Most patients in this group had a focal epilepsy (89.5%), and 35.1% of these patients experienced IS in their early life. A history of drug-resistant epilepsy was experienced by 52.6% of patients, who successively went on to become seizure-free. Most of them were on monotherapy (56.1%), mainly on carbamazepine (21.0%) and valproic acid (12.3%). No patients were treated with vigabatrin.

Patients were seizure-free for a mean period of 12.9 years (range from 1 to 49 years). The mean age at seizure control was 19.4 years (range from 7 months to 71 years).

Three patients were seizure-free after epilepsy surgery (Engel class IA) (follow-up 7.5 years; range 1–14.5 years).

Ten patients withdrew from therapy after a seizure-free period and seven of these patients are still seizure-free (follow-up 20–49 years). Considering epilepsy diagnosis, six of seven had focal epilepsy and one of seven had IS. Most (six of seven) had seizure onset in the first year of life (range 3 months to 9 years); five of seven subjects experienced IS in their epilepsy history. Six patients carried a TSC2 mutation, and NMI in one patient NMI.

All the patients without relapse had more than six cortical tubers on magnetic resonance imaging (MRI), whereas the cognitive profiles were normal in three patients and BIF in two; two patients had ID.

The three patients who relapsed had focal epilepsy (onset 18 months to 8 years), and none experienced IS; the seizure-free period was from 18 to 43 years. Two patients had TSC2 mutation; in one patient these data are not yet available. One patient had ID; two patients had normal IQ. Two patients had more than six tuberous lesions on MRI; one patient had two lesions only.

Patients with drug-resistant epilepsy

Among patients with drug-resistant seizures, 38 were adults and 21 were children; therefore, the mean age of this group of patients was 24.1 (SD 16.2) years. Mean age at TSC diagnosis was 7.6 years (SD 11.6), considerably lower than in seizure-free patients (p < 0.05), and in patients without epilepsy (p < 0.05).

The majority of patients with drug-resistant seizures had TSC2 mutations (66.1%, p < 0.05), as opposed to TSC1 mutations (18.6%); 11.9% of patients were familial cases (two TSC1, four TSC2, one NMI) (Fig. 1).

Figure 1.

Genetic analysis (%) divided into three groups (seizure-free; drug-resistant; no epilepsy). *p < 0.05.

Considering cognition, severe ID was evident in 35.6% (p < 0.05), moderate ID in 15.3%, mild ID in 20.3% (p < 0.05), BIF in 8.5% and normal IQ in 20.3% (Fig. 2). Educational level (mean 7.6, SD 2.9 years) was lower than in seizure-free patients (p < 0.05) and patients without epilepsy (p < 0.05).

Figure 2.

Cognitive profiles (%) divided into three groups (seizure-free; drug-resistant; no epilepsy). *p < 0.05; BIF, borderline intellectual functioning; ID, intellectual disability.

Cortical tubers were evident in 100% of cases, with more than six tubers in 88.1% (p < 0.05); SENs and SEGA were present in 86.4% and 22.0%, respectively (p < 0.05).

Even if the neurologic examination did not demonstrate significant alterations in most of the cases (71.2%), a psychiatric disorder was evident in 38.9% (p < 0.05).

Most of the patients had a focal epilepsy (61.0%) or multifocal epilepsy (13.5%), and 15.3% had Lennox-Gastaut syndrome (LGS). A history of IS was present in 54.2% of cases (p > 0.05) (Fig. 3).

Figure 3.

Type of seizures between the two groups (seizure-free vs. drug-resistant).

Epilepsy onset in the first year of life was characteristic of this group of patients (66.1%, p > 0.05), either with focal seizures (54.2%) or IS (37.3%). Eleven patients experienced status epilepticus, associated with fever in four patients (6.8%), and without fever in seven cases (11.9%, p < 0.05).

Due to drug resistance, polytherapy was prescribed in most cases (84.7%). In the drug-refractory group, the mean number of medications was 2.3 (range 1–4). Four patients underwent epilepsy surgery (Engel class IVB); mean follow-up was 11 years (range 9–16 years).

Patients without epilepsy

Among patients without epilepsy, most of these (32/37) were adults (mean age 36.8, SD 15.5 years) with a normal IQ (94.6%, p < 0.05) and higher educational level compared to patients with epilepsy (10.7 years, p < 0.05); the age at TSC diagnosis was also relatively higher in patients without epilepsy (23.8 years; p < 0.05)

Genetic analysis in this group demonstrated mutations in the TSC1 gene in 40.5% and in TSC2 gene in 35.1%, respectively, and NMI in 24.4% of cases. In about 30% of cases TSC was familial (seven TSC1, four TSC2).

Nevertheless, MRI findings showed cortical tubers in 89.2% of cases and SENs in 67.7%, but only one patient had a SEGA, diagnosed due to intracranial hypertension symptoms, and treated with neurosurgery at 44 years.

EEG findings were unremarkable in 62.2% of these patients, even if in a small proportion of cases focal (10.8%) and/or slow activity (8.1%) were detected. Most of these had a normal neurologic evaluation (97.3%), but a psychiatric disorder was diagnosed in 8/37 (21.6%).

Patients with only one isolated seizure

Only seven patients in our TSC group experienced one isolated seizure (4.4%); five were adults—mean age was 21.6 years (SD 18). Mean age at TSC diagnosis was 17.0 years (SD 18.8) (range from 1 to 46 years), but mean age at epilepsy diagnosis was 7.7 years (SD 7.2) (range from 1 to 20 years). Mean follow-up was 2.8 years. Two patients experienced a febrile seizure, whereas one patient aged 5 years had status epileptic (not febrile), making her the only patient of this group on antiepileptic drug (AED) therapy.

All patients who experienced only one seizure had a normal cognitive IQ and a mean education level of 8.2 (SD 3.4) years. TSC1 and TSC2 mutations were equally distributed in these patients (three TSC1, three TSC2, one NMI).

Cortical tubers and SENs were detected on MRI in 100% and 71.4%, respectively, and in most of the cases (57.1%) EEG showed focal or slow (28.6%) abnormalities.

Regarding medication, one patient was on monotherapy, whereas six patients are not on any therapy. None of the patients had neurologic disturbances. Two (28.6%) of seven patients presented an anxiety disorder. Familial cases were 42.9% (one TSC1, two TSC2). Patients were seizure-free for 13.0 years (range from 2 to 44 years).


The prevalence of epilepsy in patients with TSC is reported to be between 80% and 90% (Jóźwiak et al., 2000; O'Callaghan et al., 2004; Chu-Shore et al., 2010).

Most of the patients with TSC (>90%) have seizures at least once, and in 75% of them epilepsy starts before 1 year of age (Roach et al., 1998).

In our group of patients, the prevalence of epilepsy is slightly reduced with a prevalence of 72.5%, whereas an additional 4.4% experienced just one seizure in their lives. These data could be ascribed to the fact that the recruitment of our patients is made through a multidisciplinary tuberous sclerosis center, whereas most frequently epilepsy in TSC patients is studied in highly specialized epilepsy centers (Sparagana et al., 2003; Curatolo et al., 2008; Chu-Shore et al., 2010). Moreover, some of the patients included in this study were diagnosed as parents or relatives of the patients, and often presented with milder or subclinical forms of the disease.

Even if we cannot extrapolate certain indications, owing to the small sample of cases with one isolated seizure, the possibility that a patient with TSC experiences a seizure once is supported in our series by the mean follow-up of 13 years (range 2–49 years), and the prescription for medication in this group of patients was limited to young age at onset and long seizure duration. Careful long-term surveillance should be dedicated to patients with TSC who are presenting a first seizure, but we do not support the indication of treating with anticonvulsant medication after the first episode all patients with TSC (Chu-Shore et al., 2010).

The most striking feature associated with poor seizure outcome in our group of patients with TSC is the age at seizure onset, so we therefore can surmise that early age at epilepsy onset is one of the major risk factors for subsequent drug resistance, as recently confirmed (Fig. 4) (Curatolo et al., 2012).

Figure 4.

Epilepsy onset mean age and cognitive level for drug-resistant and seizure-free patients. ID, intellectual disability.

Adult patients with TSC have better seizure control, even if in their epilepsy history half of them have experienced periods of drug resistance. These data can be attributed to the age dependency of epilepsy, even in patients with a well-defined symptomatic syndrome.

There were 10 patients (8.6%) with adulthood-onset epilepsy in our group. These data are consistent with the current literature, which reports that >12% of adult patients without a previous history of seizures have been shown to subsequently develop epilepsy (Chu-Shore et al., 2010).

With respect to genetic analysis, it is noteworthy that both in the group of patients without epilepsy and in the group of patients with drug responsive epilepsy TSC1 mutations were prevalent, as well as absence of mutation in the group without epilepsy. This finding supports previous data on genotype–phenotype correlation in TSC, demonstrating that patients carrying TSC1 gene mutation or without mutation identified (NMI) have a milder neurologic phenotype (reduced risk of epilepsy and cognitive impairment) (Camposano et al., 2009; Moavero et al., 2010). Moreover, relative to sporadic cases, the familial cases in our series, demonstrated a more favorable outcome (regarding epilepsy and cognitive outcome), even if this difference did not reach statistical significance (Moavero et al., 2010).

On the other hand, we confirmed previous data consistent with a poor epilepsy prognosis in the group with drug-resistance epilepsy associated with TSC2 mutations (Chu-Shore et al., 2010), although a relation with the mutation type needs to be pursued (Van Eeghen et al., 2013). Regarding seizure type, IS were diagnosed in many patients in this series (52), but only 32 of them became drug refractory at long-term follow-up.

As previously reported by Goh et al. (2005), a history of IS does not correlate invariably with mental retardation and/or drug-resistant epilepsy. In our group of patients, IS associated or not with focal seizures at onset were not significantly associated with drug resistance, but there is a significant correlation with ID (Fig. 5).

Figure 5.

Cognitive profiles in patients with a history of infantile spasms. BIF, borderline intellectual functioning; ID, intellectual disability.

Moreover, in our group of TSC patients, those with focal epilepsy from the onset have a favorable trend toward epilepsy freedom, as opposed to patients that develop multifocal EEG abnormalities, multiple seizure types, and LGS pattern. These data confirm previous reports (Curatolo et al., 2005; Cusmai et al., 2011) and enhance the necessity of seizure control as early as possible (Bombardieri et al., 2010), even before seizure onset (Jóźwiak et al., 2011), in order to prevent future drug-resistance.

The relationship between the number and location of cortical tubers detected by MRI and the severity of seizures is still controversial (Chou et al., 2008; Soo et al., 2011). In our series, MRI revealed multiple lesions in most of the patients, but a significant correlation was found between the number of cortical tubers, as well as the incidence of SEN and SEGA, and unfavorable epilepsy outcome. On the other hand, it is also true that patients who obtained seizure freedom and discontinued AEDs showed more than six tubers and SEN at MRI. Therefore, in our opinion, MRI findings should be considered as an adjunctive factor that is not always crucial for determining epilepsy prognosis.

A limitation of our study is the retrospective analysis of the cases, so that we were not able to verify longitudinal data that could be helpful to allow more precise prognostic indicators. In this perspective we were not able to evaluate the EEG history of all our cases and subsequently verify if initial EEG showing epileptiform discharges could be crucial for epilepsy prognosis as previously reported (Soo et al., 2011). However, we emphasize that EEG findings were unremarkable in the majority of patients who did not develop epilepsy, so this evaluation should be considered a good indicator of favorable prognosis.

Also in our series of patients, the early age at epilepsy onset and ID (mild to severe) are prominent in the group with drug-resistant epilepsy, as well as low level of education. It is noteworthy that the incidence of psychiatric comorbidity is obviously higher in the drug-resistant group, but psychiatric disorders are present both in the seizure-free group and in patients without epilepsy.

Although epilepsy management can be challenging in TSC, 35.6 % of patients had their seizures controlled: by monotherapy in 56% and by polytherapy in 32%. Moreover, 12% of the patients became seizure-free and were off medication.

The most effective AEDs in our group of patients appeared to be carbamazepine (37.5%) and valproic acid (21.9%). No patient was in therapy with vigabatrin, since, probably because of concern over long-term visual field deficits, most of the patients were switched to other AEDs.

AED discontinuation was possible in a small sample of patients with a favorable outcome, independent from early seizure onset, IS history, MRI findings, and intellectual functioning. With the limitation of the number of the patients, we can argue the hypothesis that AED discontinuation can be faced in a proportion of TSC patients without relapse, as reported before (Sparagana et al., 2003). In addition to drug treatment, epilepsy surgery should be considered in patients with TSC.

Even if the majority of patients (>60%) operated on remain seizure-free after surgery, improvement is strongly related to focal stereotypical seizures, single EEG focus, and/or focal lesion. A number of novel noninvasive methods (positron emission tomography [PET], single-photon emission computerized tomography [SPECT], and magnetoencephalography [MEG]), as well as invasive monitoring with intracranial electrodes, can be used in preoperative evaluation in order to better define the area for resection and expand the number of surgical candidates (Evans et al., 2012).

In medically refractory patients who are not suitable for surgery, vagus nerve stimulation has proven efficacy in significantly reducing seizure frequency in >50% of cases (Elliott et al., 2009; Zamponi et al., 2010; ). Few patients in our series underwent resective or palliative surgery for their drug-resistant seizures, and we cannot extrapolate certain data from these cases.

New evidence suggests that mammalian targets of rapamycin (mTOR) inhibitors may be helpful in the management of intractable seizures for individuals with TSC, but this treatment may need to be continuous and long term to maintain seizure control, and the benefit and the safety of this alternative therapeutic option is still unknown (Curatolo et al., 2012).


Wider knowledge about epilepsy characteristics and long-term prognosis in patients with TSC is needed to ensure better management for patients and their families. Age-dependent characteristics of epilepsy in TSC should be identified to develop strategic therapeutic options tailored to the single patient.

Moreover, delineating different categories of epileptic patients with TSC will further allow identification of specific biomarkers that could be fundamental for delineation of genotype–phenotype correlations, and to address early preventive treatment when indicated.


We are grateful to all patients and families who participated in the study.


None of the Authors have any conflicts of interest to report. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.


  • Image of creator

    Aglaia Vignoli, MD is a Researcher in Child and Adolescent Neurology and Psychiatry at Epilepsy Center, San Paolo Hospital, University of Milan, Italy.