Key Points This was an open-label extension (OLE) of a phase 3 trial (GWPCARE6), evaluating the long-term safety and efficacy of cannabidiol (CBD) in patients with tuberous sclerosis complex (TSC)–associated seizures One hundred ninety-nine patients were treated with CBD at a mean modal dose of 27 mg/kg/day for a median of 267 days (range, 18–910) of treatment Most patients (92%) had an adverse event (AE); the most common AEs were diarrhea, seizure, and decreased appetite, and they were mostly mild or moderate in severity The median percentage reduction of 54% in TSC-associated seizures was observed at week 12 and sustained through 48 weeks of treatment More than 80% of patients/caregivers and physicians reported an improvement in the patient's overall condition at 26 weeks 1 INTRODUCTION

Tuberous sclerosis complex (TSC) is a highly variable genetic disorder characterized by benign hamartomas in multiple organ systems, most notably in brain, skin, kidneys, lungs, heart, and eyes.1-4 It is caused primarily by mutations in tumor suppressor genes TSC1 or TSC2, resulting in an increased mechanistic target of rapamycin (mTOR) activation with subsequent excessive cell growth and proliferation.1-3, 5 The incidence of TSC is estimated at 1 in 6000 live births, affecting 40 000–80 000 people in the United States and 1–2 million people worldwide.6, 7 Approximately ~85% of patients with TSC experience epilepsy, with onset usually during the first 2 years of life; it can persist lifelong with multiple seizure types.6, 8-11 Patients with TSC experience infantile spasms and focal seizures as infants and various other seizure types as the disease progresses.9, 12

Current treatments for seizures associated with TSC include antiseizure medications (ASMs), commonly referred to as antiepileptic drugs; the mTOR pathway inhibitor everolimus; surgical procedures; vagus nerve stimulation; and dietary therapy.6, 13-17 Despite these treatment options, more than 60% of patients have treatment-resistant epilepsy,12 which can be associated with various neurodevelopmental disorders.10, 18

Highly purified pharmaceutical formulation of cannabidiol (CBD; Epidiolex® in the United States and Epidyolex® in the United Kingdom, European Union, and Australia) has demonstrated efficacy, with an acceptable safety profile against seizures associated with Dravet syndrome and Lennox-Gastaut syndrome in four randomized, placebo-controlled phase 3 trials.19-22 Results from an expanded-access program demonstrated that CBD may also be an effective and well-tolerated treatment for TSC-associated seizures.23 The effect of CBD on seizures associated with TSC was further evaluated in a 16-week, randomized, double-blind, placebo-controlled, multicenter phase 3 trial (GWPCARE6). In this trial, add-on CBD produced almost 50% reduction in TSC-associated seizures compared with an ≈30% reduction with placebo, and had an acceptable safety profile.24 Based on these results, CBD was approved for treatment of seizures associated with TSC in patients aged 1 year and older in the United States25 and at least 2 years of age in the United Kingdom and in the European Union.26

Patients who completed treatment in the placebo-controlled, double-blind phase of GWPCARE6 were eligible to enroll in the open-label extension (OLE) phase under the same protocol for evaluation of the long-term safety and efficacy of CBD. Herein we present results of an interim analysis (data cutoff, February 26, 2019) of safety, efficacy, and patient/caregiver- and physician-reported outcomes in patients with TSC enrolled in the OLE phase of GWPCARE6.

2 METHODS

This study was an OLE of the 16-week randomized, double-blind, placebo-controlled, multinational phase 3 trial GWPCARE6 (NCT02544763) and enrolled patients who completed treatment in the randomized, controlled, blinded phase. Patients with a definite clinical diagnosis of TSC1 and treatment-resistant epilepsy were eligible if they were 1–65 years of age; had at least eight TSC-associated seizures during the 4-week baseline period, with at least one seizure occurring in at least 3 of the 4 weeks; and were taking at least one ASM at baseline of the randomized, controlled phase. Key exclusion criteria were a history of nonepileptic seizures, clinically significant illness other than epilepsy, surgery for epilepsy in the 6 months before screening, felbamate use for less than 1 year before screening, and a history of alcohol or substance abuse or recreational or medicinal use of cannabis or cannabinoid-based medications. Although not allowed during the randomized phase, on-label use of mTOR inhibitors for the treatment of seizures and tumors was permitted during the OLE. The trial protocol was approved by the relevant institutional review board or ethics committee at each participating site and was conducted in accordance with the principles of the Declaration of Helsinki and the International Conference on Harmonization Tripartite Guideline on Good Clinical Practice. All patients or their caregivers provided written informed consent before any trial procedure was carried out. Patients who were developmentally mature enough to understand the trial provided written assent.

All patients received a pharmaceutical formulation of highly purified CBD derived from Cannabis sativa L. plant (100 mg/ml oral solution; Epidiolex® in the United States and Epidyolex® in the United Kingdom, European Union, and Australia; GW Research Ltd, Cambridge, United Kingdom). After completing treatment in the randomized phase, patients started a 2-week blinded transition period, during which the blinded medication (CBD 25 mg/kg/day, CBD 50 mg/kg/day, or placebo) from the randomized, controlled phase was tapered down to zero while simultaneously CBD was titrated up to 25 mg/kg/day (Table S1). The dose could then be titrated up to 50 mg/kg/day during a 3-week titration period in increments of 2.5 mg/kg/day every 2 days. The daily dose of CBD was maintained throughout the trial; however, the investigator could decrease the dose if a patient experienced intolerance or could increase it up to the maximum dose of 50 mg/kg/day if required for better seizure control, until the optimal dose was found. CBD was taken twice daily in equally divided doses in addition to the patient's current ASM. The dose for concomitant ASMs could also be adjusted to manage side effects associated with their use. Patients could receive treatment for up to 1 year, except in the United States and Poland, where they could continue treatment beyond 1 year. At the end of treatment, patients (outside the United States and Poland) could continue using CBD outside the study. For patients who did not immediately continue using CBD outside the trial or upon decision to withdraw, CBD dose was tapered down by 10% per day for 10 days, unless continued dosing was not possible because of an adverse event (AE). A follow-up visit was performed 4 weeks after the last dose of CBD (including the final taper period dose) in patients completing or withdrawing from the trial. The study design is shown in Figure S1.

The primary objective of this OLE study was to evaluate the long-term safety and tolerability of add-on CBD based on the incidence, type, and severity of treatment-emergent AEs in patients with uncontrolled seizures associated with TSC. Patients used paper diaries for daily recording of any AEs, CBD intake, and the use of concomitant ASMs and rescue medications throughout the study. Blood and urine samples for clinical laboratory assessments were collected at all clinic visits (when possible).

The secondary objective was evaluation of efficacy through assessment of percentage change in the frequency of TSC-associated seizures; the number of patients considered treatment responders with ≥25%, ≥50%, ≥75%, and 100% reduction in TSC-associated seizures; change in overall condition of patients on the Subject/Caregiver Global Impression of Change (S/CGIC) scale and Physician Global Impression of Change (PGIC) scale; and percentage change in the frequency of total seizures. All changes were assessed relative to the pre-randomization baseline of the trial's blinded phase. TSC-associated seizures included countable focal motor seizures without impairment of awareness, focal seizures with impairment of awareness, focal seizures evolving to bilateral generalized convulsive seizures, and generalized seizures (tonic–clonic, tonic, clonic, or atonic); they excluded absence, myoclonic, focal sensory, and infantile/epileptic spasms. This functional definition of TSC-associated seizures was approved by the US Food and Drug Administration (FDA), the European Medicines Agency (EMA), and the Epilepsy Study Consortium's independent committee of experts. Total seizures included TSC-associated seizures, infantile/epileptic spasms, and absence, myoclonic, and focal sensory seizures. Patients or their caregivers used an interactive voice-response system to record the number and type of seizures, severity of focal seizures, and the number of status epilepticus episodes. The S/CGIC and PGIC scales are both 7-point scales that include three categories for improvement (slightly improved, much improved, and very much improved), three for worsening (slightly worse, much worse, and very much worse), and an option to indicate no change.

No formal sample size was calculated for this open-label trial, and all patients who completed the blinded, placebo-controlled phase were eligible to enroll. All patients who received at least one dose of CBD during the OLE were included in the safety evaluation data set. Safety data for the complete OLE phase are reported here. Seizure frequency (average per 28 days) was calculated for each 12-week treatment window and expressed as median percentage reduction from pre-randomization baseline. Seizure outcomes for up to 48 weeks of treatment are presented here. Analyses of seizure frequency and treatment responder rates were repeated using the last observation carried forward (LOCF) imputation step, which is described in the Appendix S1. Seizure outcomes analyses were conducted for patients who completed weeks 37–48 of the treatment period. In addition, we conducted a post hoc analysis to evaluate the safety and efficacy of CBD by modal dose categories ≤25 and >25 mg/kg/day. The S/CGIC and PGIC results at the 26-week visit are reported. There was no formal hypothesis testing and results are presented descriptively.

3 RESULTS 3.1 Patients

Of the 201 patients who completed treatment in the placebo-controlled, double-blind phase of the trial, 199 enrolled in the OLE phase (Figure 1) conducted at 43 sites across six countries (Australia, The Netherlands, Poland, Spain, United Kingdom, and the United States). At the time of this analysis, 10 patients (5%) had completed treatment, 39 (20%) had withdrawn, and 150 (75%) had treatment ongoing. The reasons for withdrawal from the OLE were AEs (11 patients [6%]) and decision to withdraw by the patient (8 [4%]), parent/guardian (5 [3%]), or investigator (3 [2%]); 1 patient (0.5%) met the withdrawal criteria, and 11 patients (6%) had other reasons for withdrawal. The 1-year retention rate (calculated as the number of patients who reached the treatment window at weeks 37–48 divided by the total number of patients who could have reached it at the time of this analysis) was 79%, with 104 patients reaching weeks 37–48 of the 131 patients who could have reached the treatment window (Table S2).

Patient disposition. aOne patient taking cannabidiol (CBD) 25 mg/kg/day in the randomized phase was withdrawn after completing the treatment because of vomiting and did not enter the open label extension (OLE); another patient taking CBD 50 mg/kg/day did not enter the OLE after completing treatment in the randomized phase because of logistical problems and on psychiatrist's recommendation. bWithdrawals are reported by the primary reason reported for each patient. cOne patient met the withdrawal criterion of elevation in ALT/AST levels. dOf patients who had other reasons, nine withdrew due to lack of efficacy, one withdrew because of difficulty maintaining compliance, and one switched to a commercial product. Abbreviations: AE, adverse event; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CBD, cannabidiol; OLE, open-label extension

The median patient age was 10.8 years; 46 (23%) were age 18 and older (Table 1). Patients had previously tried and discontinued a median of four ASMs and were taking a median of three ASMs at the start of the placebo-controlled phase. Valproate (41%), vigabatrin (36%), clobazam (32%), and levetiracetam (29%) were the most commonly used medications during the OLE. Although mTOR inhibitors were not allowed during the randomized phase, 9 patients received everolimus and 25 sirolimus during the OLE. Five patients (3%) were on concomitant ketogenic diet therapy, and 23 (12%) were using vagus nerve stimulation. The median number of TSC-associated seizures during the 4-week pre-randomization baseline was 56.9.

TABLE 1. Demographics and characteristics of patients in the OLE Parameter CBD modal dose

≤25 mg/kg/day

(n = 156)

>25 mg/kg/day

(n = 43)

Total CBD

(N = 199)

Age at entry to randomized, controlled phase, years Mean (SD) 13.5 (10.5) 12.0 (8.0) 13.2 (10.0) Median (range) 11.0 (1.1–56.8) 10.1 (1.7–32.7) 10.8 (1.1–56.8) Age group, years, n (%) <2 3 (2) 5 (12) 8 (4) 2–5 37 (24) 5 (12) 42 (21) 6–11 44 (28) 15 (35) 59 (30) 12–17 36 (23) 8 (19) 44 (22) 18–65 36 (23) 10 (23) 46 (23) Sex, n (%) Male 94 (60) 24 (56) 118 (59) Race, n (%) White 138 (88) 39 (91) 177 (89) Other 18 (12) 4 (9) 22 (11) Geographic region, n (%) United States 77 (49) 26 (60) 103 (52) Rest of the world 79 (51) 17 (40) 96 (48) Number of ASMs at the start of randomized, controlled phase, median (range) Previous 4 (0–15) 4 (0–13) 4 (0–15) Current 3 (0–5) 3 (1–5) 3 (0–5) ASMs at start of randomized, controlled phase (>20% of patients in any group), n (%) Valproate 63 (40) 17 (40) 80 (40) Vigabatrin 54 (35) 15 (35) 69 (35) Levetiracetam 41 (26) 15 (35) 56 (28) Clobazam 43 (28) 11 (26) 54 (27) Most common ASMs during the OLE (>20% of patients in any group), n (%) Valproate 65 (42) 17 (40) 82 (41) Vigabatrin 56 (36) 15 (35) 71 (36) Clobazam 50 (32) 14 (33) 64 (32) Levetiracetam 42 (27) 15 (35) 57 (29) Lamotrigine 30 (19) 15 (35) 45 (23) Lacosamide 32 (21) 12 (28) 44 (22) Oxcarbazepine 25 (16) 9 (21) 34 (17) Number of seizures per 28 days, median (Q1, Q3) TSC-associated 54.7 (27.5, 111.0) 66.0 (38.0, 106.0) 56.9 (28.0, 109.0) Total 57.0 (29.0, 116.9) 66.0 (38.0, 131.0) 58.9 (29.8, 117.0) Seizure subtypes, n (%) Focal seizures without impaired awareness 69 (44) 21 (49) 90 (45) Focal seizures with impaired awareness 108 (69) 26 (60) 134 (67) Focal to bilateral motor seizures 51 (33) 8 (19) 59 (30) Tonic–clonic 31 (20) 11 (26) 42 (21) Tonic 43 (28) 14 (33) 57 (29) Clonic 7 (4) 0 7 (4) Atonic 21 (13) 4 (9) 25 (13) Absence 18 (12) 4 (9) 22 (11) Myoclonic 10 (6) 2 (5) 12 (6) Partial sensory 3 (2) 2 (5) 5 (3) Infantile or epileptic spasms 11 (7) 2 (5) 13 (7) Abbreviations: ASM, antiseizure medication; CBD, cannabidiol; OLE, open-label extension; Q1, first quartile; Q3, third quartile; SD, standard deviation; TSC, tuberous sclerosis complex.

Overall median treatment duration was 267 days (range, 18–910 days) and patient-years on treatment was 152.9, and the mean modal (standard deviation [SD]) dose was 27 mg/kg/day (7.3). The mean modal (SD) dose was 26 mg/kg/day (6.1) for weeks 1–12 and 28 mg/kg/day (7.7–8.8) for the remaining 12-week treatment windows through week 48. Of the total 199 patients, 156 (78%) had a CBD modal dose ≤25 mg/kg/day with a mean (SD) of 24 mg/kg/day (3.3) and 43 (22%) had modal dose >25 mg/kg/day with a mean (SD) of 38 mg/kg/day (7.4). The median treatment duration was 232 days (range, 18–876 days) and patient-years on treatment was 110.5 for patients with a modal dose ≤25 mg/kg/day. For patients with a modal dose >25 mg/kg/day, the median treatment duration was 318 days (range, 40–910 days) and patient-years on treatment was 42.4.

3.2 Safety

Treatment-emergent AEs were reported in 92% of all patients, 91% of patients with modal dose ≤25 mg/kg/day, and 98% of patients with modal dose >25 mg/kg/day (Table 2). Most AEs were of mild or moderate severity, with 36% of patients reporting the severity as mild, 47% as moderate, and 9% as severe. Diarrhea, seizures, and decreased appetite were the most frequently reported AEs, and most of these were of mild or moderate severity (Table 2 and Table S3). Somnolence was reported more frequently in patients taking concomitant clobazam (16 of 66 patients [24%]) than in patients not on clobazam (15 of 133 patients [11%]); the overall incidence of AEs was similar between the two subgroups (94% of patients taking clobazam vs 92% of those not taking clobazam). An AE was listed as one of the reasons for permanent treatment discontinuation in 12 patients (6%). The most common AE leading to permanent treatment discontinuation in >1% of patients was seizures (four patients [2%]; Table S3). Fifty patients (25%) had permanent dose reduction because of an AE, primarily diarrhea (21 patients [11%]; Table S3). Serious AEs were reported by 15% of all patients, with seizures and status epilepticus as the most frequently reported serious AEs in >1% of patients (Table 2). There was one death due to cardiopulmonary failure during the study, which was deemed not treatment related by the investigator.

TABLE 2. Treatment-emergent adverse events during the OLE Adverse event CBD modal dose

≤25 mg/kg/day

(n = 156)

>25 mg/kg/day

(n = 43)

Total CBD

(N = 199)

Patients, n (%) Any AE 142 (91) 42 (98) 184 (92) AEs leading to permanent discontinuationa 12 (8) 0 12 (6) Serious AEs 19 (12) 10 (23) 29 (15) Deaths 1 (0.6) 0 1 (0.5) AEs reported in >10% of patients in any group Diarrhea 63 (40) 20 (47) 83 (42) Seizure 30 (19) 14 (33) 44 (22) Decreased appetite 26 (17) 13 (30) 39 (20) Pyrexia 21 (13) 11 (26) 32 (16) Vomiting 20 (13) 12 (28) 32 (16) Somnolence 19 (12) 12 (28) 31 (16) Nasopharyngitis 22 (14) 6 (14) 28 (14) Upper respiratory tract infection 19 (12) 7 (16) 26 (13) Alanine aminotransferase increasedb 7 (4) 6 (14) 13 (7) Fall 8 (5) 6 (14) 14 (7) Cough 7 (4) 5 (12) 12 (6) Aspartate aminotransferase increasedb 5 (3) 5 (12) 10 (5) Weight decreased 5 (3) 5 (12) 10 (5) Serious AEs reported in >1% of patients in any group Seizure 3 (2) 3 (7) 6 (3) Status epilepticus 4 (3) 1 (2) 5 (3) Influenza 2 (1.3) 0 2 (1) Alanine aminotransferase increasedb 1 (0.6) 1 (2) 2 (1) Aspartate aminotransferase increasedb 1 (0.6) 1 (2) 2 (1) Dehydration 1 (0.6) 1 (2) 2 (1) Mental status changes 0 2 (5) 2 (1) Platelet count decreased 1 (0.6) 1 (2) 2 (1) Transaminase increased 1 (0.6) 1 (2) 2 (1) Blood bilirubin increased 0 1 (2) 1 (0.5) Diarrhea 0 1 (2) 1 (0.5) Gastroenteritis 0 1 (2) 1 (0.5) Gastroenteritis astroviral 0 1 (2) 1 (0.5) Lethargy 0 1 (2) 1 (0.5) Respiratory tract infection 0 1 (2) 1 (0.5) Seizure cluster 0 1 (2) 1 (0.5) Soft tissue inflammation 0 1 (2) 1 (0.5) Abbreviations: AE, adverse event; CBD, cannabidiol; OLE, open-label extension.

Laboratory testing showed elevations in serum alanine aminotransferase (ALT) or aspartate aminotransferase (AST) levels of more than 3 times the upper limit of normal (ULN) in 17 patients (9%), 12 of whom (71%) were taking concomitant valproate. No patient met the standard criteria for severe drug-induced liver injury (Hy's Law) with concurrent bilirubin levels more than 2 times ULN. One patient discontinued treatment because of elevated transaminase levels. Elevations occurred within 1 month of starting treatment in 9 of the 17 patients and between 1 and 3 months of starting treatment in 6 patients; 2 patients had elevations more than 3 months (100 days) after starting treatment. At the time of this analysis, increased ALT or AST levels had resolved in 14 of 17 patients—spontaneously in 6 patients, following treatment discontinuation in 1, and after CBD or ASM dose reduction in 7 (4 patients reduced valproate dose); transaminase levels had not resolved in 3 patients.

3.3 Efficacy

During the first 12 weeks of the OLE, CBD treatment produced a median reduction of 54% from the randomized phase baseline in the monthly TSC-associated seizure frequency (Figure 2A). The effect was maintained throughout the treatment windows at weeks 13–24, 25–36, and 37–48, with 54%–68% reduction from baseline. Median reduction of 52%–73% was observed in patients with modal dose ≤25 mg/kg/day and 49%–61% in those with modal dose >25 mg/kg/day across the 12-week treatment windows up to 48 weeks (Figure S2A). In the LOCF analysis, a 54%–56% reduction from baseline in TSC-associated seizures was observed across the 12-week windows (Figure S3A); reduction was 61%–68% among patients who completed the weeks 37–48 treatment window (Figure S4A).

Reduction in (A) TSC-associated seizure frequency and (B) total seizure frequency. (A) Data from the randomized, controlled phase of the trial are presented here for comparison. Abbreviations: CBD, cannabidiol; Q1, first quartile; Q3, third quartile; TSC, tuberous sclerosis complex; WD, number of patients who withdrew during a treatment window

More than 50% of patients had ≥50% reduction in TSC-associated seizure frequency across the 12-week treatment windows; ≥75% reduction in seizures was observed in >25% of patients (Figure 3A). More than 5% of patients were seizure-free during each 12-week treatment window, with seven patients (4%) remaining seizure-free the entire treatment period. The proportion of patients who had ≥25%, ≥50%, ≥75%, and 100% reduction in TSC-associated seizures remained consistent across the treatment windows when LOCF was used to calculate responder rates (Figure S5A), as well as among patients who completed the weeks 37–48 treatment window (Figure S6A).

Responder rates for (A) TSC-associated seizures and (B) total seizures. Abbreviation: TSC, tuberous sclerosis complex

A median reduction from baseline of 51% in the monthly total seizure frequency was observed at the weeks 1–12 treatment window; reduction in seizures was maintained across the 12-week treatment windows through week 48 (54%–67% reduction from baseline; Figure 2B). Reduction in total seizures ranged from 53%–70% in patients with a modal dose ≤25 mg/kg/day and 48%–55% in those with modal dose >25 mg/kg/day (Figure S2B). In the LOCF analysis, 50%–53% reduction in total seizure frequency was observed (Figure S3B). Among patients who completed weeks 37–48 of treatment, 61%–67% reduction from baseline was observed (Figure S4B). Overall, a ≥50% reduction in total seizure frequency was observed in 51%–60% of patients across the 12-week treatment windows for up to 48 weeks (Figure 3B). Patients reporting ≥25%, ≥50%, ≥75%, and 100% reduction in seizures were consistent across the treatment windows in the LOCF analysis of responder rates (Figure S5B), as well as among patients who completed the treatment window at weeks 37–48 (Figure S6B).

At the 26-week visit, 87% of the 124 patients or caregivers who completed the S/CGIC evaluations reported an improvement from baseline in the overall condition of patients (Figure 4). Improvements were reported by 80% of physicians on the PGIC scale.