Epilepsy is a major manifestation of tuberous sclerosis complex (TSC). Everolimus is an mammalian target of rapamycin complex 1 inhibitor with demonstrated benefit in several aspects of TSC. We report the first prospective human clinical trial to directly assess whether everolimus will also benefit epilepsy in TSC patients.
The effect of everolimus on seizure control was assessed using a prospective, multicenter, open-label, phase I/II clinical trial. Patients ≥2 years of age with confirmed diagnosis of TSC and medically refractory epilepsy were treated for a total of 12 weeks. The primary endpoint was percentage of patients with a ≥50% reduction in seizure frequency over a 4-week period before and after treatment. Secondary endpoints assessed impact on electroencephalography (EEG), behavior, and quality of life.
Twenty-three patients were enrolled, and 20 patients were treated with everolimus. Seizure frequency was reduced by ≥50% in 12 of 20 subjects. Overall, seizures were reduced in 17 of the 20 by a median reduction of 73% (p < 0.001). Seizure frequency was also reduced during 23-hour EEG monitoring (p = 0.007). Significant reductions in seizure duration and improvement in parent-reported behavior and quality of life were also observed. There were 83 reported adverse events that were thought to be treatment-related, all of which were mild or moderate in severity.
Seizure control improved in the majority of TSC patients with medically refractory epilepsy following treatment with everolimus. Everolimus demonstrated additional benefits on behavior and quality of life. Treatment was safe and well tolerated. Everolimus may be a therapeutic option for refractory epilepsy in this population. Ann Neurol 2013;74:679–687
Tuberous sclerosis complex (TSC) is an inheritable, autosomal dominant genetic disorder, with the majority of cases arising through spontaneous mutation. Incidence is estimated at 1 in 6,000 live births, affecting >1 million individuals worldwide. The clinical hallmark of the disease is the occurrence of benign tumors, or hamartomas, which can arise in any organ system from birth through adulthood. Although significant morbidity is associated with cardiac, renal, and pulmonary involvement, it is neurological aspects of TSC that often prove the most challenging to treat. Subependymal giant cell astrocytomas (SEGAs) can arise in as many as 20%, autism spectrum disorder and other neurocognitive, behavioral, and mood disorders in >60%, and epilepsy in 80 to 90%.
There has been limited investigation regarding the optimal treatment of epilepsy in individuals with TSC. Successful seizure control is achievable in some patients with conventional antiepileptic medications (AEDs), but at least ⅓ of TSC patients are refractory to available medical and surgical therapies.[2, 4] In patients with refractory epilepsy, the prevalence of cognitive impairment and neuropsychiatric and developmental disorders is particularly high. There is an acute need to develop new therapeutic approaches for the treatment of epilepsy in TSC, to improve not only seizure control but also cognitive function and patient quality of life.
Understanding of the genetic and molecular basis of TSC has greatly expanded over the past decade. Mutations in either TSC1 (encoding for TSC1, or hamartin) or TSC2 (encoding for TSC2, or tuberin) lead to similar disease manifestations. TSC1 and TSC2 form a complex responsible for the regulation of the mammalian target of rapamycin complex 1 (mTORC1) via the intermediary signaling protein Rheb (ras homolog enriched in brain). Pharmacologic inhibitors of mTORC1, particularly sirolimus and everolimus, have demonstrated efficacy for the treatment of multiple aspects of TSC, including renal angiomyolipomata,[8, 9] SEGAs,[10-12] and lymphangioleiomyomatosis (LAM). Preclinical studies and early human case reports have also demonstrated beneficial effects on seizures.[14-18] In more recent prospective human clinical trials, however, benefit could not be demonstrated when seizure frequency was analyzed as a secondary outcome measure.[11, 12] Here we report the results of the first prospective, multicenter, open-label, phase I/II human clinical trial designed to directly assess the benefit of the mTORC1 inhibitor everolimus on seizure control in patients with TSC.
Patients and Methods
Study participants were recruited from the TSC clinics at Cincinnati Children's Hospital Medical Center (CCHMC) and Texas Children's Hospital. The institutional review board at each institution approved the protocol, and an independent data and safety monitoring board reviewed the progress of the trial twice annually. Written informed consent was obtained from all patients (or their parent or legal guardian, when applicable) prior to enrollment into the study. Able patients under 18 years of age provided oral assent. The clinical trial is listed on clinicaltrials.gov (NCT01070316).
To be eligible for participation, subjects had to be at least 2 years of age, with a confirmed diagnosis of TSC (1998 clinical diagnostic criteria or identified disease-causing mutation in either TSC1 or TSC2). In addition, subjects had to have medically refractory epilepsy, defined here as the patient having been adequately trialed on at least 2 approved AED therapies and having at least 8 reported seizures in the 30 days immediately prior to enrollment. Concomitant treatment with vagus nerve stimulation (VNS) or the ketogenic diet and a history of previous epilepsy surgery was permitted. In addition, subjects had to be medically stable, without evidence of significant infectious, immunologic, or oncologic comorbidity at the time of enrollment, and could not be currently taking or previously treated systemically with an mTOR inhibitor.
A prospective, open-label, phase I/II study design was used. The core study was divided into a 4-week baseline observational period (weeks 1–4), followed by a 4-week titration period (weeks 5–8) and an 8-week maintenance period (weeks 9–16). Clinical examination, safety laboratory studies, Quality of Life for Children with Epilepsy (QOLCE), Nisonger Child Behavioral Rating Form (NCBRF), and a 23-hour electroencephalography with video (vEEG) monitoring study were performed at the beginning of week 1 and the end of week 16. Additional clinical examinations and laboratory assessments were performed at the end of week 8 (end of titration). AED trough levels (weeks 1, 6, 8, and 16) and everolimus trough levels (weeks 6, 8, and 16) were measured. Following enrollment, subjects were maintained on stable doses of concurrent AEDs throughout the baseline, titration, and maintenance phases of the study. If applicable, ketogenic diet parameters and VNS settings also were held constant throughout.
Patients were given a diary to record seizure description, duration, and number of occurrences each day. Subjects with 8 or more seizures recorded during the 4-week baseline phase were eligible to begin study medication; subjects with <8 seizures were discontinued from the study and replaced until the target enrollment was achieved. Everolimus (2.5mg and 5mg tablet formulation) was initiated at the beginning of the fifth week at 5mg/m2/day, dosed once daily in the morning, rounded to the nearest 2.5mg/dose. A serum everolimus trough level was obtained at weeks 6 and 8, and dose was adjusted for a target range between 5 and 15ng/ml. No further adjustments were made during the maintenance phase (weeks 9–16), unless medication-related side effects necessitated a decrease in dose to the next lowest dosing level (−2.5mg/day).
The primary efficacy endpoint was the percentage of patients demonstrating a 50% or greater reduction in seizure frequency at the end of the maintenance phase (weeks 13–16) compared to baseline (weeks 1–4). Patients with 25% or greater reduction were considered partial responders. Responders and partial responders were eligible for continued treatment following the end of maintenance (week 16) if desired. Secondary efficacy endpoints compared baseline (week 1) with end of maintenance (week 16) treatment effects on vEEG, QOLCE assessment, and NCBRF. The vEEGs were performed utilizing the international 10–20 placement of electrodes, and recorded for a minimum of 23 hours at each time point. EEGs were then scored independently by 2 clinical neurophysiologists (K.H.-B. and A.E.A.) for the number of clinical and electrographic seizures, seizure-onset zone, and seizure duration. Interictal epileptiform activity during two 15-minute epochs (awake and asleep) was also determined by region and rated on a continuous scale from 0 (none) to 10 (almost continuous). Throughout the study, reported adverse events were recorded and categorized by Common Terminology Criteria for Adverse Events (CTCAE) version 3.0. Consistent with previous open-label TSC clinical trials with everolimus, all infections were considered to be related to treatment by protocol. Events reported as seizures by parents that were later confirmed by vEEG to be nonepileptic were excluded from the primary and secondary endpoint analyses.
Statistical significance was determined using Wilcoxon signed rank test for before and after repeated measures. For the secondary analysis of response at time intervals other than primary endpoint, Friedman repeated measures analysis of variance on ranks with Tukey all pairwise multiple comparison procedures was used. Pearson correlation coefficient was employed to evaluate the primary efficacy endpoint relationship to everolimus dose and serum trough level. No correction for missing observations was required for seizure frequency or vEEG data. However, infrequent missing observations were noted for other secondary endpoints, including seizure duration, QOLCE, and NCBRF, and when encountered, mean imputation was utilized.
A total of 23 subjects were screened and eligible to begin the baseline observation phase, 20 of whom (10 male, 10 female) qualified for study drug initiation at week 5 (Fig 1). In 2 of the discontinued subjects, insufficient seizures occurred during the baseline observation phase to qualify for study drug initiation. In the other, the subject's parents elected to initiate treatment with a different AED instead of proceeding further in the clinical trial. Of the 20 subjects treated, the median age was 8 years (range = 2–21; Table 1). TSC genotyping was available for 7 subjects, all of whom were mutations of TSC2. The median number of concurrent AEDs was 2 (range = 1–4), with lamotrigine (n = 8), vigabatrin (n = 7), valproic acid (n = 5), and felbamate (n = 5) being most common. Five subjects had VNS, 4 had undergone prior epilepsy surgery, and none was currently on the ketogenic diet.
The median maintenance dose of everolimus was 8.4mg/m2/day (range = 3.4–13.7) or 7.5mg/day (range = 2.5–12.5). Corresponding trough serum levels at the end of the maintenance phase were between 1.6 and 16.1ng/dl, with a median of 6.1ng/dl. During the study, all subjects reported at least 1 adverse event (range = 2–10; Table 2). All were mild or moderate in severity (CTCAE 3.0 grade = 1 or 2), and none was severe or life-threatening. Upper respiratory infections and stomatitis/mucositis were the most common. Only 1 serious adverse event occurred, and this was a subject with a previous history of multiple hospitalizations following prolonged seizures who experienced a similar event during the titration phase, triggered by otitis media and fever. The patient recovered without sequelae and was able to continue treatment with everolimus.
Upper respiratory infection (19); otitis media (5); gastroenteritis (4)
Congestion/rhinorrhea (7); cough (6)
Effect on Seizure Frequency and Duration
After treatment for 12 weeks (4 weeks titration and 8 weeks maintenance), everolimus therapy was associated with a clinically meaningful and statistically significant reduction in seizure frequency (Fig 2A). Seventeen of 20 subjects experienced an improvement in seizure frequency (12 responders, 3 partial responders, and 5 nonresponders). Overall, the median seizure frequency decreased by 73% (31 vs 8.5 seizures per 28-day period, p < 0.001). The median cumulative seizure duration also decreased by 70% (p = 0.020). Four subjects (20%) were free of clinical seizures, and 7 (35%) had at least a 90% reduction in seizure frequency.
Improvement in seizure frequency and duration was also noted on 23-hour vEEG performed before and after treatment with everolimus (see Fig 2B). Four subjects had no seizures at either time point captured on vEEG. Of the remaining 16 subjects, the follow-up vEEG captured fewer seizures in 13. Overall, the median seizure frequency over 23 hours was reduced from 3.5 to 1.5 (range = −33 to +3, p = 0.007). Overall, this response was driven primarily by a reduction in partial onset seizures (p = 0.048), although it is interesting to note that the 2 subjects with the highest number of generalized onset seizures at baseline (40 and 24, respectively) showed the greatest improvement overall (−33 and −22, respectively). The median cumulative seizure duration over the same period also decreased (69%, p = 0.051). Brain regions with abnormal interictal activity, however, did not significantly change during both awake (p = 0.806) or sleep (p = 0.374) epochs.
To determine whether everolimus effect on seizure control occurred early or late in the treatment phase, we analyzed the average number of seizures at each 4-week period after treatment was initiated. Eight subjects reported a reduction in seizure frequency during titration and 11 during the early maintenance period that followed. Overall, the median number of seizures was reduced by 45% during titration and 39% during early maintenance, neither of which was statistically significant compared to baseline (Fig 3). A statistically significant improvement in seizure frequency was only observed in the final maintenance period (p < 0.05). Seizure freedom followed a similar pattern, with none reporting no seizures during titration, 1 during early maintenance, and another 3 at the end of the maintenance phase.
Effect on Behavior and Quality of Life
The NCBRF is a standardized parent rating instrument for assessing behavior in children and adolescents across all ranges of intelligence quotient.[20, 21] Behaviors are divided into positive and negative domains. Everolimus treatment had a greater impact on reducing negative domain behaviors (p = 0.021) than increasing positive domain behaviors (p = 0.083; Table 3), although improvement in some individual components of both domains was observed. These included adaptive social behaviors, conduct problems, and insecurity/anxiety. Overall quality of life, reported by parents through the QOLCE survey instrument, also improved (p < 0.001), driven by positive changes in attention, behavior, other cognitive, social interaction, stigma, physical restrictions, and social activity domains (Table 4).
This is the first prospective, human clinical trial to directly demonstrate efficacy of everolimus in the treatment of seizures for TSC patients with refractory epilepsy. In addition, parent-reported improvements in behavior and quality of life were observed. Although dramatic improvement in seizure frequency was observed in several individuals in as few as 4 weeks while everolimus dosing was being titrated to the target dose range, optimal response overall was not evident until week 12. Had the difference been dose-dependent, we would have expected to see improvement after titration and early maintenance but not between early and late maintenance periods. Instead, difference was detected between the early and late maintenance periods only, suggesting that mTORC1 inhibition is primarily mediated through 1 or more duration-dependent mechanisms.
Which mechanism(s) are responsible for these positive changes is not known, but they are likely to be distinct from those of traditional AEDs, which target ion channels, synaptic receptor function, or neurotransmitter release directly.[23, 24] This may explain why improvement was observed in the majority of patients after many different AED regimens had failed, in addition to VNS and epilepsy surgery. It also supports early evidence that everolimus may act more as a true antiepileptogenic agent and less as a conventional anticonvulsant. Through the inhibition of mTORC1, everolimus likely acts by an indirect mechanism involving protein synthesis and turnover that in turn influences synaptic excitability, long-term depression and potentiation, and neuronal plasticity.[23, 24] Larger-scale changes in neuronal networks could also play a role, supported by the favorable changes that have been observed in large neuronal fiber tracts in TSC patients treated with everolimus. Additional preclinical and clinical investigation is necessary to further elucidate what role each of these mechanisms contributes to the improvement in seizure control associated with everolimus treatment.
Two prior human TSC clinical trials have attempted to assess the effect of everolimus on seizure control and yielded conflicting results. Both trials were performed in patients with SEGAs, where epilepsy was a secondary endpoint assessed by parent report and vEEG. The first study was an open-label, single-center, phase II trial involving 28 patients who showed a statistically significant reduction in seizures on 24-hour vEEG following everolimus treatment for 6 months. There was also an improvement in parent-reported seizure frequency over time. Patients with daily seizures declined from 27% at baseline to 13% at 24 months, and seizure-free patients increased from 39% to 65% over the same interval. However, these findings could not be duplicated in a follow-up randomized, placebo-controlled, double-blind multicenter phase III clinical trial (EXIST-1) involving 117 SEGA patients. Differences in population size and study design could be important in the observed discrepancies. In addition, seizure management was not standardized, and practices among multiple centers can vary significantly, affecting baseline seizure control and concurrent medication use. Parent and clinician blinding and placebo comparison are also important.
The current study does not address all the differences between the 2 prior studies, but does correct 2 fundamental problems involving epilepsy outcomes as a secondary endpoint in both. First, not all subjects in the prior studies had seizures at baseline, and following treatment with everolimus they continued to be seizure-free, making it harder to show a treatment-related effect. This was particularly problematic for the analysis in the EXIST-1 trial, where the majority of patients had no seizures at baseline (median seizure frequency was 0) in both treatment groups. Thus, no change was observed following treatment with everolimus, as no change could be detected from the outset. Second, clinicians in both studies were permitted to adjust patients' epilepsy therapy regimens throughout, as seizure frequency was not the primary objective of either study. This introduced additional confounding variables potentially impacting seizure control independent of everolimus treatment. All patients in the current study were required to have a minimum of 8 seizures per month prior to initiating treatment. The majority were experiencing on average of 1 or more seizures every day at baseline. Also, no changes in any concurrent AEDs or other antiepilepsy therapies were permitted throughout the entire study period.
Nonetheless, our study is not without its own limitations. The sample size is relatively small, could be subject to sampling error bias, and could have inadvertently selected a patient population with an as yet undetermined favorable advantage to respond to everolimus treatment. Our sample size also limits the ability to assess any combinatorial effect of specific AED combinations with everolimus on seizure control. Finally, the open-label design introduces the possibility of ascertainment bias, especially with regard to parent reporting of epileptic events, quality of life, and behavioral ratings. Parents, like clinicians, want to believe the new treatment is helping their child. However, we saw similar improvement in seizure frequency and duration on vEEG, a highly objective measure, suggesting that the observed benefit is real. Definitive proof may be achieved with a follow-up placebo-controlled, randomized, double-blind phase III clinical trial, which is planned to begin later in 2013 (EXIST-3, NCT01713946).
Epilepsy is often a devastating diagnosis even in the absence of TSC and can create significant financial, social, and functional burdens on the affected individual as well as caregivers. Epilepsy, cognitive function, and behavior are also closely linked in TSC.[3, 5, 28, 29] As such, it is not surprising that the observed improvement in seizure control in the current study is accompanied by improvement in parent-reported quality of life on the QOLCE and behavior on the NCBRF. Only 2 previous prospective clinical trials in TSC have included similar assessment on quality of life. TSC patients with SEGA treated with everolimus reported overall improvement, although this did not reach statistical significance. TSC patients with pulmonary LAM treated with sirolimus, an mTORC1 inhibitor similar to everolimus, likewise reported improved quality of life that was significant. Attempts to measure behavioral and cognitive changes in previous TSC clinical trials have been even more problematic as a result of the wide variation in neurocognitive and neuropsychiatric function in any given population with TSC.[5, 11] Limited results in the SEGA trial with everolimus and another angiomyolipoma trial with sirolimus suggest potential benefit in some domains, but results are not conclusive. A randomized, double-blind, placebo-controlled clinical trial is in progress to directly ascertain any effect of everolimus on neurocognition (NCT01289912).
Everolimus was safe and well tolerated in patients with TSC in our study. All adverse events were mild or moderate in severity, and none resulted in discontinuation of treatment. The most commonly encountered treatment-related side effects were upper respiratory infections and stomatitis. Both are known to be associated with the mTORC1 inhibitor class of medications and can be effectively managed to minimize occurrence and severity in patients with TSC. Only 1 serious adverse event occurred (hospitalization) in this study that was disease-related and not attributed to everolimus. No new safety concerns emerged. Our current analysis did not evaluate potential side effects and safety concerns beyond 12 weeks, so long-term safety has yet to be determined for TSC patients treated with everolimus specifically for epilepsy. Long-term safety data are available from 2 separate studies evaluating the use of everolimus for the treatment of SEGA in TSC. No new or unusual treatment-related adverse effects were reported in 28 patients treated for a median duration of 34 months (range = 5–47 months), compared to analysis performed initially when treatment was a median duration of 21 months (range = 5–34 months). A second study, which focused on patients younger than 3 years taken from a larger cohort of 117 patients participating in the EXIST-1 clinical trial, reported that the incidence of adverse events after an average treatment duration of 35 months (range = 33–38 months) was similar to that of older children and adults. In both of these studies, as well as our present study, treatment with everolimus is ongoing, thus providing the needed opportunity for continued surveillance for safety and efficacy of everolimus in this population over even longer durations of treatment.
As already mentioned, aberrant upregulation of mTORC1 is firmly established in TSC pathogenesis and symptomology. Analysis of discrete, hamartomatous lesions such as SEGA, cortical tubers, and subependymal nodules (SEN) confirmed this for cerebral aspects of the disease.[32, 33] More recent studies have demonstrated that mTORC1 plays an important role in brain structure and functions not specifically tied to SEGA, SEN, or tubers, with impact on neurocognition and epilepsy.[34, 35] The results of our study support these earlier findings, and it is interesting to speculate whether the role of mTORC1 in epilepsy and the benefit of everolimus might extend beyond that observed in TSC. mTORC1 overactivation has been implicated in genetic and neurodevelopmental syndromes in which epilepsy is prominent, including Pretzel syndrome, Cowden syndrome,[37-39] and hemimegancephaly. mTORC1 also is increasingly being implicated in more common, nonsyndromic causes and types of epilepsy that include focal cortical dysplasia and other cortical malformations,[41, 42] posttraumatic brain injury, temporal lobe epilepsy,[17, 44] and absence seizures. Additional clinical trials are needed to determine whether everolimus will have benefit in these non-TSC epilepsy syndromes similar to that demonstrated here for TSC. Such evidence would not only justify use outside of TSC, but further implicate mTORC1 as an important pathogenic component that is shared among many different epilepsies regardless of the primary underlying etiology.
In summary, everolimus treatment in this multicenter, open-label, phase II clinical trial reduced seizure frequency and duration in the majority of TSC epilepsy patients whose seizure control was previously refractory to currently available AEDs. This improvement in seizure control was associated with better behavior and quality of life. Treatment-related side effects were limited, providing a favorable risk:benefit ratio. Although follow-up phase III clinical studies are needed to confirm these promising results, everolimus may be a therapeutic option in refractory epilepsy patients with TSC.
Funding for this study was provided by Novartis Pharmaceuticals, the Clack Foundation, and Cincinnati Children's Hospital. Novartis provided medications for the study.
We thank the patients and families for their participation and contributions to this clinical trial; and S. Selk, R. Dosani, S. Bruns, E. Turner, M. Kuhlmann, and B. Fan for assistance contributing to the success of the study.