CADTH Reimbursement Review

Cannabidiol (Epidiolex)

Sponsor: Jazz Pharmaceuticals Canada, Inc.

Therapeutic area: Lennox-Gastaut syndrome

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Clinical Review

Executive Summary

An overview of the submission details for the drug under review is provided in Table 1.

Table 1: Background Information of Application Submitted for Review

DS = Dravet syndrome; LGS = Lennox-Gastaut syndrome; NOC = Notice of Compliance; TSC = tuberous sclerosis complex.

^aDeviation requests for this reimbursement request were accepted on June 21, 2023, and September 15, 2023.

Introduction

Lennox-Gastaut syndrome (LGS) is a lifelong, complex epilepsy syndrome associated with multiple seizure types that vary across patients. LGS presents in the first 4 years of life (peak age of onset: 3 to 5 years) in children who had otherwise experienced age-appropriate development and is associated with refractory and multiple treatment-resistant seizure types, cognitive and behavioural impairments, and poor outcomes into adulthood.^1-4 LGS has various etiologies, and patients experience multiple seizure types with distinctive electroencephalogram (EEG) features, which include tonic seizures (stiffening of the body, upward eye gaze, dilated pupils, and altered breathing patterns) and atypical absence seizures (staring spells), myoclonic jerks (sudden muscle jerks), tonic or atonic “drops” (brief loss of muscle tone), generalized tonic-clonic seizures (muscle stiffness and rhythmic jerking), and focal seizures.³ Atonic and tonic seizures can be accompanied by dangerous falls or “drop seizures” that often lead to injury.⁵ Indeed, LGS is considered a life-threatening condition associated with high rates of sudden unexpected death in epilepsy (SUDEP) and a risk of death among children 14 times higher than among the US general population.^5-7 Currently, LGS is diagnosed using clinical criteria; there is no specific diagnostic test and no specific biologic markers for the diagnosis of LGS.^6,8 LGS affects between 3% and 10% of children with epilepsy, more commonly males. The peak age for onset is 3 to 5 years, with extreme incidence occurring in the first and 10th years of life.⁹ The prevalence of LGS in Canada, as estimated by the sponsor, is 12 in 100,000 people.¹⁰

The goal of treatment is to achieve seizure freedom.^3,6,8,11 In Canada, the only drugs currently indicated specifically for LGS are rufinamide¹² and lamotrigine,¹³ both as add-on treatments to other antiseizure medications (ASMs). While lamotrigine is available through the Ontario Drug Benefit program, rufinamide must be obtained through the Exceptional Access Program.¹⁴ In addition to ASMs, most patients use enteral medications; dietary therapies such as ketogenic, modified Atkins, or low–glycemic index diets; neuromodulation with vagus nerve stimulation (VNS) or deep brain stimulation; and nonresective surgeries such as corpus callosotomy. Patients can benefit from surgical resection when the source of seizure activity can be identified. The clinical experts consulted by CADTH noted that purified cannabidiol from a licensed producer (“artisanal CBD” or medical cannabis) is available in Canada and has been used extensively for the treatment of drug-resistant LGS in children and adults, albeit at lower doses than were used in clinical trials of cannabidiol.

Cannabidiol received a Notice of Compliance from Health Canada on November 15, 2023, through the standard review pathway. The Health Canada indication for cannabidiol is for use as adjunctive therapy for the treatment of seizures associated with LGS, Dravet syndrome, or tuberous sclerosis complex in patients aged 2 years and older. The sponsor requested that CADTH review each neurologic disorder as a separate submission.

The product monograph–recommended starting dosage of cannabidiol is 2.5 mg/kg twice daily.¹⁵ After 1 week, the dosage should be increased to a maintenance dosage of 5 mg/kg twice daily. Based on individual clinical response and tolerability, each dose can be increased in weekly increments of 2.5 mg/kg, administered twice daily, up to a maximum recommended dosage of 10 mg/kg twice daily (20 mg/kg/day). For patients in whom a more rapid titration from 10 mg/kg/day to 20 mg/kg/day is warranted, the dose may be increased no more frequently than every other day. The product monograph also provides a dosing adjustment for patients with moderate or severe hepatic impairment, due to the risk of hepatocellular injury related to cannabidiol.¹⁵

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of cannabidiol 100 mg/mL oral solution for use as adjunctive therapy for the treatment of seizures associated with LGS in patients aged 2 years and older.

Stakeholder Perspectives

The information in this section is a summary of input provided by the patient and clinician groups who responded to CADTH’s call for input and from the clinical experts consulted by CADTH for the purpose of this review.

Patient Input

Patient group input was provided by the Canadian Epilepsy Alliance (CEA). The CEA is a network of organizations supporting and advocating for people living with epilepsy and their families. Patient input was sourced from the knowledge and experiences of patients, caregivers, clinicians, volunteers, donors, and funders, collected from 24 member associations. According to input from the CEA, current treatments may fail in 30% of patients with epilepsy. According to the input, people living with uncontrolled epilepsy are often socially isolated due to stigma and to fear of rejection in social, work, and educational settings. Patients often experience depression and anxiety upon initial diagnosis and often continuously experience these conditions if ASMs stop working. Caregivers and family members are also impacted by epilepsy as their lives revolve around the seizures experienced by patients under their care. Anxiety among caregivers is common as they worry about when the next seizure will occur, the consequences of the epilepsy, and how to navigate social gatherings (e.g., a young person with epilepsy gets invited to a birthday party). In addition, caregivers often experience compassion fatigue since they cannot leave the person with epilepsy alone, are often sleep deprived due to sleep interruptions or anxiety, and can be affected by side effects the patient may experience from medications (e.g., mood swings, sexual dysfunction, suicidal thoughts, memory loss, fatigue, exhaustion). Both patients and caregivers emphasized the importance of treatment that results in seizure freedom. However, patients and caregivers noted that they would accept a treatment that resulted in a reduction in the absolute number of seizures, as even a reduction in seizure frequency could improve overall quality of life. Because people with intractable epilepsy are very often unemployed or underemployed, not covered under employer-funded insurance plans, and/or have restricted income, most drugs to treat their epilepsy are inaccessible. Accordingly, the CEA stressed the importance of new medications being placed on the provincial formulary so that patients with intractable epilepsy have access to novel treatments. Input provided by the CEA did not include experience with cannabidiol.

Clinician Input

Input From Clinical Experts Consulted by CADTH

The following input was provided by 2 clinical specialists with expertise in the diagnosis and management of LGS.

The clinical experts consulted by CADTH for the purpose of this review noted that, despite the multiple treatment options available, there are currently no treatments available to reverse the course of LGS. Moreover, the overall prognosis of LGS remains unfavourable, and not all patients experience response to the available treatments. The clinical experts added that there is a need for treatment that can meaningfully improve quality of life for patients and their caregivers. The clinical experts noted that, despite its novel mechanism of action compared to available therapies, cannabidiol does not address the underlying disease process any more than other available treatments. Accordingly, the clinical experts suggest that cannabidiol would complement other available treatments as a symptom management treatment. The clinical experts opined that cannabidiol could be combined with 1 or 2 first-line ASM. The clinical experts also felt that it would be reasonable to require adequate trials of 1 or 2 other ASMs before the use of cannabidiol. The clinical experts expected that the approval of cannabidiol would lead to a shift away from the use of medical cannabidiol in jurisdictions where cannabidiol is reimbursed by either public or private drug insurance plans. According to input from the clinical experts, it is difficult to predict which patients with LGS are most likely to benefit from cannabidiol. Patients who have not experienced response with multiple ASMs are generally less likely to experience response to the next ASM; however, these are the patients who are most in need of novel therapies. The clinical experts suggested that patients should be screened for treatment according to clinician judgment based on seizure characterization and frequency, etiology investigation, and previous ASMs trialled, along with EEG interpretation. According to the clinical experts’ input, a clinically meaningful response to treatment in epilepsy is assessed in terms of the median reduction in seizure frequency over 28-day periods, a reduction in the number of seizures of 50% or more, and seizure freedom rates (i.e., reduction of total seizures per day and seizure-free days per month). The clinical experts added that seizure frequency should be assessed every 4 weeks. The clinical experts also noted that improved quality of life for both patients and caregivers is an important secondary outcome. According to the clinical experts’ input, treatment with cannabidiol should be discontinued if patients develop persistent and progressive elevation of transaminases or recurrent vomiting and diarrhea, which would compromise the absorption of ASMs. In addition, treatment with cannabidiol should be reassessed if a patient develops status epilepticus with no other reasonable explanation. The clinical experts noted that the prescribing and monitoring of cannabidiol for LGS should be limited to neurologists who have special expertise in the management of epilepsy in pediatric or adult patients.

Clinician Group Input

No clinician group input was received with this submission.

Drug Program Input

Input was obtained from the drug programs that participate in the CADTH reimbursement review process. The following were identified as key factors that could potentially impact the implementation of a CADTH recommendation for cannabidiol:

• relevant comparators

• considerations for initiation of therapy

• considerations for continuation or renewal of therapy

• considerations for discontinuation of therapy

• considerations for prescribing of therapy

• generalizability

• care provision issues

• system and economic issues.

Clinical Evidence

Systematic Review

Description of Studies

Two studies were included in the sponsor-conducted systematic review: the CARE 3 and CARE 4 trials.^16,17

Both the CARE 3 and CARE 4 trials were multicentre, double-blind, phase III randomized controlled trials (RCTs) evaluating the efficacy of cannabidiol as adjunctive treatment in reducing drop seizure frequency in patients aged 2 to 55 years with a clinical diagnosis of LGS.

In the CARE 3 trial, 225 patients across 29 sites in 4 countries (US, Spain, France, and UK) were randomized 1:1:1 to receive treatment with cannabidiol 20 mg/kg/day (n = 76), cannabidiol 10 mg/kg/day (n = 73), or volume-matched placebo (n = 76). Patients in the placebo group were split into 2 equivalent cohorts: half receiving 10 mg/kg/day (n = 38), and half receiving 20 mg/kg/day (n = 38). In the CARE 4 trial, 171 patients across 24 sites in 3 countries (US, Netherlands, and Poland) were randomized to receive treatment with either cannabidiol 20 mg/kg/day (n = 86) or volume-matched placebo (n = 86). The randomization in both the CARE 3 and CARE 4 trials was stratified by age group (2 to 5 years, 6 to 11 years, 12 to 17 years, and 18 to 55 years). Patients were titrated from a starting dosage of 2.5 mg/kg/day up to 10 mg/kg/day over 7 days, or up to 20 mg/kg/day over 11 days, and remained at this dosage level for the duration of the treatment period. Assigned treatments were add-ons to 1 or more background ASMs.

The primary efficacy end point for both the CARE 3 and CARE 4 trials was reduction in the number of drop seizures (per 28 days), compared with placebo, in patients with LGS. A “drop seizure” was defined as an attack or spell (atonic, tonic, or tonic-clonic) involving the entire body, trunk, or head that led or could have led to a fall, an injury, slumping in a chair, or the patient’s head hitting a surface. The key secondary outcome of interest was the number of patients who experienced a 50% reduction in the number of drop seizures at the end of the treatment period. Other outcomes that were assessed in the CARE 3 and CARE 4 trials included the following: proportions of patients who experienced at least a 25%, at least a 75%, or a 100% reduction in the number of drop seizures at the end of the treatment period; number of inpatient hospitalizations due to epilepsy; and health-related quality of life (HRQoL) as assessed by the Quality of Life in Childhood Epilepsy (QOLCE) questionnaire, for patients aged 2 to 18 years, or the Quality of Life in Epilepsy, version 2 (QOLIE-31-P) questionnaire, for patients aged 19 years and older, and by the number of drop seizure–free days.

In the CARE 3 trial, the mean age of the patients enrolled in the trial was 16.01 years (standard deviation [SD] = 10.77 years) in the cannabidiol 20 mg/kg/day group, 15.43 years (SD = 9.48 years) in the cannabidiol 10 mg/kg/day group, and 15.29 years (SD = 9.26 years) in the pooled placebo group. In the CARE 3 trial, the median number of drop seizures per 28 days at baseline was higher in the cannabidiol 10 mg/kg/day group (median = 86.90; interquartile range [IQR], 14.0 to 7,494.0) than in the cannabidiol 20 mg/kg/day group (median = 85.53; IQR, 13.0 to 1,092.0) and the pooled placebo group (median = 80.25; IQR, 8.7 to 1,278.3). The proportion of patients reporting convulsive seizures lasting longer than 30 minutes was higher in the cannabidiol 20 mg/kg/day group (10.5%) than in the cannabidiol 10 mg/kg/day group (2.7%) and the pooled placebo group (3.9%). Across all treatment groups, the mean number of prior ASMs used by patients was approximately 7 and the mean number of current ASMs being used at baseline was almost 3.

In the CARE 4 trial, the mean age of patients enrolled in the trial was 15.3 years (SD = 9.8 years) and 15.6 years (SD = 8.7 years) in the cannabidiol 20 mg/kg/day and the volume-matched placebo groups, respectively. The median number of drop seizures at baseline was higher in the volume-matched placebo group (median = 74.67; IQR, 11.2 to 3,174.6) than in the cannabidiol 20 mg/kg/day group (median = 71.43; IQR, 10.3 to 855.9). In both treatment groups, the mean number of prior ASMs used by patients was approximately 7 and the mean number of current ASMs being used at baseline was almost 3.

Efficacy Results

Percent Change From Baseline in Drop Seizure Frequency

At the end of the 14-week treatment period in the CARE 3 trial, a reduction in the median percent change in drop seizure frequency was associated with treatment with cannabidiol 20 mg/kg/day (median difference = –21.6%; 95% confidence interval [CI], –34.8 to –6.7; P = 0.0047) and cannabidiol 10 mg/kg/day (median difference = –19.2; 95% CI, –31.2 to –7.7; P = 0.0016) compared to the pooled placebo group.

At the end of the treatment period in the CARE 4 trial, a reduction in the median percent change in drop seizure frequency was associated with treatment with cannabidiol 20 mg/kg/day compared to volume-matched placebo (median difference = –17.2%; 95% CI, –30.3 to –4.1; P = 0.0135).

Reduction in the Number of Drop Seizures From Baseline of Greater Than or Equal to 50%

In the CARE 3 trial, during the treatment period, the difference in the proportion of patients who experienced at least a 50% reduction in the number of drop seizures from baseline between the cannabidiol 20 mg/kg/day group and the pooled placebo group was 25.0% (95% CI, 11.5% to 38.5%) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 21.1% (95% CI, 7.6% to 34.7%).

In the CARE 4 trial, the difference in the proportion of patients who experienced at least a 50% reduction in drop seizure frequency from baseline between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group was 20.7% (95% CI, 6.8% to 34.5%).

Reduction in the number of Drop Seizures From Baseline of Greater Than or Equal to 25%

In the CARE 3 trial, during the treatment period, the difference in the proportion of patients who experienced at least a 25% reduction in the number of drop seizure frequency from baseline between the cannabidiol 20 mg/kg/day group and the pooled placebo group was 18.4% (95% CI, 2.8% to 34.0%) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 19.6% (95% CI, 3.9% to 35.3%).

In the CARE 4 trial, the difference in the proportion of patients who experienced at least a 25% reduction in drop seizure frequency from baseline between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group was 20.4% (95% CI, 5.8% to 35.1%).

Reduction in the Number of Drop Seizures From Baseline of Greater Than or Equal to 75%

In the CARE 3 trial, during the treatment period, the difference in the proportion of patients who experienced at least a 75% reduction in the number of drop seizures from baseline between the cannabidiol 20 mg/kg/day group and the pooled placebo group was 22.4% (95% CI, 12.0% to 55.13%) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 8.3% (95% CI, 0.3% to 16.3%).

In the CARE 4 trial, the difference in the proportion of patients who experienced at least a 75% reduction in the number of drop seizures from baseline during the treatment period between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group was 11.5% (95% CI, 1.3% to 21.8%).

Reduction in the Number of Drop Seizures From Baseline of 100%

No patient experienced a 100% reduction in the number of drop seizures during the treatment period in either the CARE 3 or the CARE 4 trial.

Number of Inpatient Hospitalizations due to Epilepsy

In the CARE 3 trial, the number of patients who experienced 1 or more inpatient hospitalizations due to epilepsy was 7 (9.2%) in the cannabidiol 20 mg/kg/day group, 6 (8.2%) in the cannabidiol 10 mg/kg/day group, and 6 (7.9%) in the pooled placebo group.

In the CARE 4 trial, the number of patients who experienced 1 or more inpatient hospitalizations due to epilepsy was 10 (11.6%) and 5 (5.9%) in the cannabidiol 20 mg/kg/day and volumed-matched placebo groups, respectively.

QOLCE for Patients Aged 2 to 18 Years

In the CARE 3 trial, overall QOLCE scores were available for 33 patients (43.4%) in the cannabidiol 20 mg/kg/day group, 36 patients (49.3%) in the cannabidiol 10 mg/kg/day group, and 38 patients (50%) in the pooled placebo group. At baseline, the overall mean QOLCE scores were comparable across the cannabidiol 20 mg/kg/day group, the cannabidiol 10 mg/kg/day group, and the pooled placebo group, at 41.6 (SD = 15.6), 40.6 (SD = 15.4), and 41.4 (SD = 16.1), respectively. The adjusted mean treatment difference in change from baseline in overall QOLCE scores between the cannabidiol 20 mg/kg/day group and the pooled placebo group was –5.1 (95% CI, –11.4 to 1.2) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 1.6 (95% CI, –4.5 to 7.8).

In the CARE 4 trial, overall QOLCE scores were available for 26 patients (30.2%) and 38 patients (44.7%) in the cannabidiol 20 mg/kg/day group and the volumed-matched placebo group, respectively. At baseline, the overall mean QOLCE scores were comparable between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group, at 39.5 (SD = 12.6) and 39.1 (SD = 15.2), respectively. The adjusted mean treatment difference in change from baseline in overall QOLCE scores between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group was 3.7 (95% CI, –3.3 to 10.7).

QOLIE-31-P for Patients Aged 19 Years and Older

In the CARE 3 trial, total QOLIE-31-P scores were available for 13 patients in the cannabidiol 20 mg/kg/day group, 14 patients in the cannabidiol 10 mg/kg/day group, and 10 patients in the pooled placebo group. At baseline, the total mean QOLIE-31-P scores were 50.2 (SD = 16.6) in the cannabidiol 20 mg/kg/day group, 56.0 (SD = 19.2) in the cannabidiol 10 mg/kg/day group, and 62.5 (SD = 13.6) in the pooled placebo group. The adjusted mean treatment difference in change from baseline in total QOLIE-31-P scores between the cannabidiol 20 mg/kg/day group and the pooled placebo group was 2.9 (95% CI, –7.3 to 13.1) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 3.6 (95% CI, –7.0 to 14.3).

In the CARE 4 trial, total QOLIE-31-P scores were available for 14 patients (16.3%) in the cannabidiol 20 mg/kg/day group and 14 patients (16.5%) in the volume-matched placebo group. At baseline, the total QOLIE-31-P scores were comparable between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group at 55.8 (SD = 13.5) and 57.3 (SD = 19.5), respectively. The adjusted mean treatment difference in total QOLIE-31-P scores was not calculated due to the low number of patients who completed the assessment.

Drop Seizure–Free Days

In the CARE 3 trial, the mean number of drop seizure–free days had increased in all treatment groups at the end of the treatment period. Compared to placebo, treatment with cannabidiol 20 mg/kg/day was associated with an increase in the mean number of drop seizure–free days of 4.6 days (95% CI, 2.5 days to 6.8 days), while treatment with cannabidiol 10 mg/kg/day was associated with an increase of 3.3 days (95% CI, 1.2 days to 5.5 days).

In the CARE 4 trial, the mean number of drop seizure–free days had increased in both treatment groups at the end of the treatment period. Compared to volume-matched placebo, treatment with cannabidiol 20 mg/kg/day was associated with a mean increase in drop seizure–free days of 2.7 days (95% CI, 0.7 days to 4.7 days).

Other Outcomes

The following outcomes were noted as being meaningful by the patient group and the clinical experts consulted by CADTH but were not assessed in either the CARE 3 or CARE 4 trial: SUDEP rate, employment, and caregiver quality of life.

Harms Results

Adverse Events

In the CARE 3 trial, at least 1 adverse event (AE) was reported by 93.9% of patients in the cannabidiol 20 mg/kg/day group, 83.6% of patients in the cannabidiol 10 mg/kg/day group, and 72.4% of patients in the pooled placebo group.

In the CARE 4 trial, at least 1 AE was reported by 86.0% and 69.4% of patients in the cannabidiol 20 mg/kg/day and the volume-matched placebo group, respectively.

Serious Adverse Events

In the CARE 3 trial, at least 1 serious AE (SAE) was reported by 15.9% of patients in the cannabidiol 20 mg/kg/day group, 19.4% of patients in the cannabidiol 10 mg/kg/day group, and 10.5% of patients in the pooled placebo group.

In the CARE 4 trial, at least 1 SAE was reported by 23.3% and 4.7% of patients in the cannabidiol 20 mg/kg/day and the volume-matched placebo groups, respectively.

Treatment Discontinuation due to AEs

In the CARE 3 trial, discontinuation of treatment due to AEs was reported by 7.3% of patients in the cannabidiol 20 mg/kg/day group, 1.5% of patients in the cannabidiol 10 mg/kg/day group, and 1.3% of patients in the pooled placebo group.

In the CARE 4 trial, discontinuation of treatment due to AEs was reported by 14.0% and 1.2% of patients in the cannabidiol 20 mg/kg/day group and the volume-matched placebo group, respectively.

Mortality

There were no reported deaths in the CARE 3 trial. In the CARE 4 trial, 1 death was recorded (1.2% of patients) due to acute respiratory distress syndrome in the cannabidiol 20 mg/kg/day group.

Notable Harms

Notable harms of interest were related to nervous system disorders (i.e., somnolence, status epilepticus, and sedation), hepatocellular injury or investigation (i.e., increased levels of alanine transaminase [ALT] and aspartate transaminase [AST], bilirubin elevation, and serum transaminase elevation), and gastrointestinal disorders (i.e., diarrhea, vomiting, and constipation).

In the CARE 3 trial, somnolence, status epilepticus, and sedation were reported in 30.5%, 4.9%, and 3.7% of patients, respectively, in the cannabidiol 20 mg/kg/day group; in 20.9%, 10.4%, and 3.0% of patients, respectively, in the cannabidiol 10 mg/kg/day group; and in 5.3%, 3.9%, and 1.3% of patients, respectively, in the pooled placebo group. Increased levels of ALT, AST, and serum transaminase were reported in 4.9%, 3.7%, and 1.2% of patients, respectively, in the cannabidiol 20 mg/kg/day group; in 4.5%, 3.7%, and 1.5% of patients, respectively, in the cannabidiol 10 mg/kg/day group; and in 1.3%, 1.3%, and 0% of patients, respectively, in the pooled placebo group. The following investigations were not reported in the CARE 3 trial: abnormal liver function test, acute hepatic failure, and hepatotoxicity. Diarrhea, vomiting, and constipation were reported in 14.6%, 12.2%, and 4.9% of patients, respectively, in the cannabidiol 20 mg/kg/day group; in 10.4%, 6.0%, and 4.5% of patients, respectively, in the cannabidiol 10 mg/kg/day group; and in 7.9%, 11.8%, and 3.9% of patients, respectively, in the pooled placebo group.

In the CARE 4 trial, somnolence, sedation, and status epilepticus were reported in 15.1%, 8.1%, and 1.2% of patients, respectively, in the cannabidiol 20 mg/kg/day group and in 9.4%, 1.2%, and 1.2% of patients, respectively, in the volume-matched placebo group. The following hepatocellular injury or investigation AEs were reported in the cannabidiol 20 mg/kg/day group: increased ALT (9.3%), increased AST (7.0%), abnormal liver function test (4.7%), acute hepatic failure (3.5%), serum transaminase elevation (2.3%), hepatic failure (1.2%), and hepatotoxicity (1.2%). In the volume-matched placebo group, increased levels of ALT and AST were reported in 2.4% and 1.2% of patients, respectively. Diarrhea, vomiting, and constipation were reported in 18.6%, 10.5%, and 7.0% of patients, respectively, in the cannabidiol 20 mg/kg/day group and in 8.2%, 16.5%, and 4.7% of patients, respectively, in the volume-matched placebo group.

Critical Appraisal

Both the CARE 3 and CARE 4 trials were multicentre, double-blind phase III RCTs, designed in accordance with the European Medicines Agency Guideline on Clinical Investigation of Medicine Products in the Treatment of Epileptic Disorders and the FDA Guidelines for the Clinical Evaluation of Antiepileptic Drugs (Adults and Children).

In both trials, patients were randomized centrally using interactive voice response system (IVRS) technology, which is typically adequate for concealing allocation until treatment assignment. IVRS technology was also used to dispense the investigational product, allowing the treatment concealment for both patients and investigators. Although the CARE 3 trial included 4 treatment types (cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, volume-matched placebo 20 mg/kg/day, and volume-matched placebo 10 mg/kg/day), the study participants were randomized using a 1:1:1 randomization ratio to the cannabidiol 20 mg/kg/day treatment group, the cannabidiol 10 mg/kg/day treatment group, and the placebo treatment group. Patients in the placebo group were split in half to receive either the 20 mg/kg/day placebo or the 10 mg/kg/day placebo, and the study results were reported based on the pooled placebo group. While this approach is acceptable, it relies on the assumption that randomization was successful in each group. Differences in baseline characteristics between patients in the cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, and pooled placebo groups were noted in the following: proportion of patients reporting convulsive seizures longer than 30 minutes and reporting concomitant use of benzodiazepine derivatives. According to the clinical experts consulted by CADTH for the purpose of this review, because of the rarity of LGS, it is unknown if these differences in baseline characteristics could influence treatment response. At CADTH’s request, the sponsor reported that the assumptions related to splitting the placebo group and pooling results for analyses were not formally tested. However, the sponsor noted, as described in Clinical Study Report for the CARE 3 trial, that posthoc sensitivity analyses were conducted on the primary end point to determine if pooling the placebo group had an effect on the results. The analyses indicated that the comparisons of the efficacy response when cannabidiol 20 mg/kg/day and 10 mg/kg/day dosages were compared to their corresponding placebo dosages (i.e., 20 mg/kg/day (10 mg/kg/day, respectively) were consistent with the results of similar comparisons in the pooled analyses. In the CARE 4 trial, the 1:1 randomization ratio and the randomization stratification factors appeared appropriate, and no notable baseline imbalances were observed between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group.

In the CARE 3 trial, a higher proportion of patients in the cannabidiol 20 mg/kg/day group discontinued the study than in the cannabidiol 10 mg/kg/day and pooled placebo groups (approximately 12% versus approximately 3%). The higher discontinuation rate in the cannabidiol 20 mg/kg/day group appeared to be driven by AEs. In the CARE 4 trial, a higher proportion of patients in the cannabidiol 20 mg/kg/day group discontinued the study than in the volume-matched placebo group (9.3% versus 1.2%). The higher discontinuation rate in the cannabidiol 20 mg/kg/day group appeared to be driven by AEs. The application of the missing not at random (MNAR) assumption and sensitivity analysis exploring missing efficacy results due to treatment discontinuation suggest that bias due to uneven discontinuation was unlikely.

In both the CARE 3 and CARE 4 trials, all efficacy outcome questionnaires and assessments were to be completed by a caregiver. To maintain consistency, the same caregiver, if the patient had multiple caregivers, was to complete and answer the questionnaire and assessment. Seizure information in both studies was ascertained using an IVRS diary, while paper diaries were used to capture use of the investigational product, use of concomitant medications, and occurrence of AEs. According to input from the clinical experts, patient and parent or caregiver reports of seizures tend to be accurate for motor seizures but are not very reliable or accurate for nonmotor seizures. However, seizure diaries are the standard method of collecting data for clinical trials, and the International League Against Epilepsy recommends the use of dairies for collecting seizure frequency data.¹⁸ Both the CARE 3 and CARE 4 trials assessed HRQoL — outcomes deemed important by the patients and clinicians — using validated and reliable instruments: the QOLCE and QOLIE-31-P questionnaires. The double-blind nature of the trials minimized the risk of bias in the measurement of subjective items on the QOLCE and QOLIE-31-P questionnaires. However, comparative efficacy conclusions based on the HRQoL outcomes are limited since the QOLCE and QOLIE-31-P questionnaires were not part of the hierarchical testing procedure and because of the low completion rates across the treatment groups. Total QOLCE scores were available for 47.6% and 36.3% of patients in the CARE 3 and CARE 4 trials, respectively, while total QOLIE-31-P scores were available for 16.5% and 16.4% of patients in the CARE 3 and CARE 4 trials, respectively. Consequently, the assessment of HRQoL in both trials is at high risk of attrition bias, although the extent and direction of the bias cannot be determined since it is not clear if the patients who completed the questionnaires were systematically different from the patients who did not. As the completion rates were similar between the treatment groups within the CARE 3 and CARE 4 trials, there is little risk that attrition bias favoured any 1 treatment group.

The analysis of efficacy results in the CARE 3 and CARE 4 trials followed a defined statistical analysis plan. The primary and key secondary end points were addressed using a hierarchical gate-keeping procedure that controlled for type I errors. The sponsor conducted additional sensitivity analyses of the primary efficacy end point using the per-protocol analysis set and testing the MNAR assumption. In all scenarios, the sensitivity analyses were consistent with the primary efficacy analysis.

The clinical experts consulted by CADTH for the purpose of this review were unable to assess if the results of the CARE 3 and CARE 4 trials were applicable to the patients seen in the Canadian clinical setting. However, the clinical experts did note trial details that were applicable to the Canadian clinical setting and others that were not representative of the Canadian clinical setting. The clinical experts noted that the treatment periods in the CARE 3 and CARE 4 trials were long enough to detect a meaningful treatment response on seizures in patients with LGS; however, the clinical experts were uncertain if the treatment response observed could be sustained in the long-term. Moreover, a longer study period would be required to detect a treatment response on cognitive functioning. In the Canadian setting, where medicinal cannabis as adjunctive treatment for seizures associated with LGS can be accessed through the medical cannabis pathway supported by Health Canada, the clinical experts would not impose any sort of washout period before initiating pharmaceutical cannabidiol.

GRADE Summary of Findings and Certainty of the Evidence

The selection of outcomes for Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) assessment was based on the sponsor’s summary of clinical evidence, consultation with clinical experts, and input received from patient groups and public drug plans. The following list of outcomes was finalized in consultation with expert committee members:

• change from baseline in number of drop seizures during the treatment period; proportion of patients who experienced a greater than or equal to 25%, greater than or equal to 50%, greater than or equal to 75%, or 100% reduction in the number of drop seizures from baseline

• number of drop seizure–free days; and number of inpatient hospitalizations due to epilepsy

• change from baseline in quality of life as measured by the QOLCE questionnaire (for patients aged 2 to 18 years) or the QOLIE-31-P questionnaire (for patients aged 19 years and older)

• notable harms, including serious adverse events and hepatocellular injury.

For the pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for the outcomes considered most relevant to inform CADTH’s expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group.^19,20

Following the GRADE approach, evidence from RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty of evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when a threshold was available) or to the null. The presence or absence of an important effect were based on thresholds informed by the clinical experts consulted for this review.

For the GRADE assessments, findings from the CARE 3 and CARE 4 trials were considered together and summarized narratively per outcome, because these studies were similar in population, interventions, design, and outcome measures.²¹

Table 2 presents the GRADE summary of findings for cannabidiol versus placebo.

Long-Term Extension Studies

To inform the longer-term safety and tolerability of cannabidiol as an adjunctive treatment in children and adults with inadequately controlled LGS, the results of 1 open-label extension (OLE) study — the CARE 5 trial — were summarized.²²

Description of Study

The CARE 5 trial was a multicentre, single-arm, OLE, phase III study with the primary objective of evaluating the longer-term safety and tolerability of cannabidiol as adjunctive treatment in patients aged 2 years and older with LGS (N = 366) who had completed the CARE 3 and CARE 4 trials (the “core studies”). The long-term efficacy of cannabidiol as adjunctive treatment in patients aged 2 years and older with LGS was evaluated as the secondary objective of the CARE 5 trial. Patients enrolled in the CARE 5 trial received adjunctive cannabidiol in addition to their usual treatment; the trial consisted of a 2-week titration period, a maintenance period, and a 10-day taper period. Patients were titrated up to 10 mg/kg/day to 20 mg/kg/day cannabidiol using the recommended titration schedule. The patients continued at their 10 mg/kg/day to 20 mg/kg/day dosage throughout the maintenance period. During the maintenance period, dosing adjustments by the investigators were permitted should a patient experience intolerance (dose decrease) or require better seizure control (dose increase) until the optimal dose was achieved. A maximum dosage of 30 mg/kg/day was permitted, if deemed necessary by the investigator. For patients whose dose had been decreased, subsequent dose increases were considered provided the treatment was adequately tolerated. Following the end-of-treatment or withdrawal visit, doses were tapered at home (10% per day for 10 days) until the end-of-taper visit. Patients could receive treatment for up to 6 years (312 weeks after visit 1), depending on the

Table 2: Summary of Findings for Cannabidiol 10 mg/kg/day and Cannabidiol 20 mg/kg/day Versus Volume-Matched Placebo for Patients With Seizures Associated With

CI = confidence interval; HRQoL = health-related quality of life; LGS = Lennox-Gastaut syndrome; MID = minimally important difference; NE = not estimated; QOLCE = Quality of Life in Childhood Epilepsy; QOLIE-31-P = Quality of Life in Epilepsy, version 2; RCT = randomized controlled trial; SD = standard deviation.

Note: Study limitations (which refer to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and publication bias were considered when assessing the certainty of the evidence. All serious concerns in these domains that led to the level of certainty being rated down are documented in the table footnotes.

^aRated down 1 level for serious imprecision. In the CARE 3 trial, the 95% CI included the potential for no clinically meaningful benefit. In the absence of an empirically derived MID, a between-group difference of 25% was used as the clinically meaningful threshold, based on clinical expert input. Any increase (improvement) not reaching this threshold indicates there is uncertainty in the clinically meaningful treatment effect of cannabidiol compared with placebo. In the CARE 4 trial, the treatment effect estimates and the lower bounds of the 95% CI for the difference between groups included the possibility of a trivial effect (little to no difference) compared with placebo. In the absence of an empirically derived MID, a between-group difference of 25% was used as the clinically meaningful threshold, based on clinical expert input. Any increase (improvement) not reaching this threshold indicates there is uncertainty in the clinically meaningful treatment effect of cannabidiol compared with placebo. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^bRated down 1 level for serious imprecision. A 20% to 30% difference in the proportion of patients reporting a reduction in drop seizure from baseline of at least 50% was considered meaningful, based on the input from the clinical experts. The observed point estimate just met the lower bounds of the MID suggested by the clinical experts. The 95% CIs of the point estimates were wide. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^cStatistical testing for this outcome was not adjusted for multiplicity. The results are considered supportive evidence.

^dRated down 1 level for serious imprecision. In the absence of an empirically derived MID, a 20% to 30% difference was used as the clinically meaningful threshold, based on clinical expert input. Any increase (improvement) not reaching this threshold indicates there is uncertainty in the clinically meaningful treatment effect of cannabidiol compared with placebo. In the CARE 3 trial, the point estimate and 95% CI are less than the 20% threshold. In the CARE 4 trial, the point estimate just meets the 20% threshold, while the lower bounds of the 95% CI fall under the threshold. Potential to rate down for serious inconsistency. The results of the CARE 3 and CARE 4 trials are different, with the effect estimate and lower 95% CI found under the threshold in the CARE 3 trial. Given that the estimate is close to the 20% threshold provided by the clinical experts, there was no rating down for inconsistency. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^eRated down 1 level for serious inconsistency. The point estimate in the CARE 4 trial is lower than that estimated in the CARE 3 trial. Rated down 1 level for serious imprecision. In the absence of an empirically derived MID, a between-group difference of 15% to 20% was used as the clinically meaningful threshold, based on clinical expert input. Any increase (improvement) not reaching this threshold indicates there is uncertainty in the clinically meaningful treatment effect of cannabidiol compared with placebo. In the CARE 4 trial, the point estimate and 95% CI are less than the 15% threshold. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^fRated down 2 levels for serious imprecision, based on there being zero events in both treatment groups. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^gRated down 1 level for serious imprecision. The 95% CI of the treatment difference included the point estimate for placebo response. In the absence of an empirically derived MID and no suggested MID from the clinical experts, the null was employed. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^hRated down 1 level for serious study limitation. Risk of bias (attrition) due to missing outcome data as results were available for less than 50% of randomized patients in the CARE 3 and CARE 4 trials. Rated down 1 level for serious imprecision. No MID estimate specific to the LGS population was identified or provided by the sponsor. Using the null, the treatment effect and the lower bound of the 95% CI included the potential for decrease (worsening) HRQoL. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. Clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^IRated down 2 levels for serious study limitation. Risk of bias due to missing outcome data as results were available for only 16% of randomized patients in the CARE 3 and CARE 4 trials. Rated down 1 level for serious imprecision. No MID estimate specific to the LGS population was identified or provided by the sponsor. Applying the MID of 5.19 established for patients with partial-onset seizure,, the treatment effect and the lower bound of the 95% CI included the potential for decrease (worsening) HRQoL. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^jRated down 2 levels for serious imprecision. In the absence of an empirically derived MID, a between-group difference of 10% was used as the clinically meaningful threshold, based on clinical expert input. Any increase (improvement) not reaching this threshold indicates there is uncertainty in the clinically meaningful treatment effect of cannabidiol compared with placebo. Difference of treatment effect could not be estimated due to the small number of events. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^kBetween-group differences in harms were not statistically tested.

^lRated down 1 level for serious imprecision. There were important concerns about the small number of events that precluded estimating a treatment effect. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

^mRated down 1 level for serious imprecision. There were important concerns about the small number of events that precluded estimating a treatment effect. Potential to rate down for serious indirectness. Although the certainty of evidence was not rated down for serious indirectness, important concerns regarding how well the evidence applies to patients with LGS in Canada were noted by CADTH. These concerns included the lack of study sites in Canada and the lack of representation of patients who had experience with medicinally sourced cannabis. The clinical experts advised that the generalizability of the CARE 3 and CARE 4 trials is difficult to determine because of the rarity of LGS and the variability in individual patient clinical characteristics.

Sources: Clinical Study Reports for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

protocols used in the country of enrolment. The end-of-treatment visit occurred after a maximum of 6 years’ treatment [312 weeks after visit 1], following early withdrawal from the study, or following an unscheduled “end-of-treatment visit” conducted no earlier than 730 days after visit 1.

The CARE 5 trial was conducted across 75 sites in 8 countries (Australia, Spain, France, Israel, Netherlands, Poland, the UK, and the US). Approximately 78% of patients were from the US. On average, patients were aged 15.9 years (SD = 9.5) and were concurrently taking 3.4 ASMs (SD = 1.38) at the time of enrollment. Of the 366 patients with LGS enrolled, 66.4% completed the treatment period, 20.5% continued to the taper phase, and 18.3% completed the taper phase.

Efficacy Results

Efficacy end points were analyzed in the safety analysis set. The retention rates for the safety analysis set at weeks 37 to 48 (12 months), weeks 85 to 96, weeks 133 to 144 (36 months), weeks 181 to 192, and weeks 241 to 252 were 82% (299 of 366 patients), 64% (236 of 366 patients), 59% (216 of 366 patients), 6% (22 of 366 patients), and 2% (8 of 366 patients), respectively. Missing data were addressed using the last observation carried forward method.

The proportion of patients experiencing drop seizure–free status at weeks 37 to 48 (12 months), weeks 133 to 144 (36 months), and weeks 253 to 264 (66 months) were 7% (24 of 364 patients), 8% (30 of 364 patients), and 9% (34 of 364 patients), respectively. The median percent change from baseline in drop seizure frequency during the same OLE periods from baseline of the core studies were –55.3% (IQR, –83.8% to –16.6%; n = 364), –59.1% (IQR, –85.7% to –15.2%; n = 364), and –59.4% (IQR, –87.1% to –16.0%; n = 364), respectively. The mean percent change in drop seizure frequency during the same OLE periods from baseline of the core studies were –34.9% (SD = 82.77; n = 364), –32.3% (SD = 106.11; n = 364), and –30.9% (SD = 127.21; n = 364), respectively. The proportion of patients who experienced a reduction in drop seizure frequency of 50% or greater during the same periods in the CARE 5 trial was 53.8% (196 of 364 patients), 56.3% (205 of 364 patients), and 58% (211 of 364 patients), respectively.

Among patients aged between 2 years and 18 years, the mean change in overall HRQoL score, as measured from the QOLCE score at baseline to last visit, was 5.5 (SD = 13.71; n = 152). Among patients aged 19 years and older, the mean change in the overall quality-of-life subscale weighted score (QOLIE-31-P) from baseline to last visit was 6.4 (SD = 28.63; n = 55).

Harms Results

A total of 353 patients with LGS (96.4%) reported experiencing 1 or more AEs during the CARE 5 trial. The most common treatment-emergent AEs were convulsion (38.5%), diarrhea (38.3%), and pyrexia (34.4%). SAEs were reported by 157 patients with LGS (42.9%). The most commonly reported SAEs were convulsion (12%), status epilepticus (11.5%), and pneumonia (7.1%). Discontinuation of treatment due to AEs was reported in 43 patients with LGS (11.7%). The most common reasons for treatment discontinuation due to AEs were convulsion (1.9%), diarrhea (1.6%), and vomiting (1.4%). Twelve patients with LGS (3.3%) died during the study. SUDEP as the cause of death was recorded in 4 patients (1.1%).

Critical Appraisal

The single-group, open-label, nonrandomized design of the CARE 5 OLE study makes interpretation of the long-term efficacy and safety of cannabidiol challenging. The lack of comparison with an active comparator and/or placebo precludes the ability to draw causal inference to assess the relative long-term therapeutic benefit or safety of cannabidiol. Although patient and caregiver self-counts of drop seizures and motor seizures were noted to be reliable by the clinical experts consulted by CADTH for the purpose of this review, self-counting of other types of seizures is not considered accurate. The results for caregiver-reported and patient-reported outcomes were inconclusive due to the open-label design of the trial and the substantial decline in the number of patients available to provide these assessments over time. Moreover, it is uncertain if the sample size (N = 366) was sufficient to detect rare AEs. As enrolment in the CARE 5 trial was contingent on the completion of a core study, thereby excluding patients who discontinued the CARE 3 or CARE 4 trials due to AEs or lack of response, it is possible that patients in the CARE 5 trial represent a select population that is more tolerant of cannabidiol. Therefore, response bias cannot be ruled out. Finally, the results may be biased due to attrition, as approximately a third of patients did not complete the study and there was wide variance in the follow-up duration for individuals. None of the CARE 5 trial sites were in Canada. Due to the rarity of LGS and a lack of robust population-based studies on LGS in Canada, the clinical experts were unable to determine if the patient population included in the CARE 5 trial was reflective of patients seen in the clinical practice setting across Canada. One clinical expert added that they would not, however, expect patients with LGS living in Canada to differ from patients with LGS living elsewhere. Adherence to the treatment regimen was not reported on and, as such, overall exposure to cannabidiol during the OLE study period is uncertain.

Indirect Comparisons

No indirect treatment comparisons were included in this submission.

Studies Addressing Gaps in the Evidence From the Systematic Review

No studies addressing gaps in the evidence from the systematic review were included in this submission.

Conclusions

LGS is a rare and complex epilepsy syndrome associated with multiple and concurrent seizure types that vary across people with the condition. Seizures associated with LGS can be difficult to treat and often become resistant to ASMs. Evidence of moderate certainty from 2 randomized, double-blind, placebo-controlled phase III trials (the CARE 3 and CARE 4 trials) in patients aged between 2 years and 55 years with LGS suggests that adjunctive treatment with cannabidiol 10 mg/kg/day or 20 mg/kg/day may result in a reduction in drop seizure frequency (primary analysis) for certain patients compared with placebo, when added to usual care, with a greater proportion of patients treated with cannabidiol reporting a reduction in drop seizure frequency of 50% or more (key secondary outcome). Although a median treatment difference in drop seizure frequency reduction of between 17% and 21% did not meet the threshold for clinical relevance, as identified by the clinical experts consulted by CADTH for the purpose of this review, the primary analysis was supported by key secondary analysis. No patients were reported at the end of the 14-week treatment period to have experienced a 100% reduction in drop seizures, and the clinical significance of the number of drop seizure–free days was uncertain. The included trials were limited to patients who reported having 2 or more drop seizures per week during the 28-day baseline period. According to input from the clinical experts, it is reasonable to expect that improvement in seizure frequency would translate to improved HRQoL. However, the ability of the CARE 3 and CARE 4 trials to assess the treatment effect of cannabidiol on HRQoL was limited due to low completion rates of the QOLCE and QOLIE-31-P questionnaires. The long-term safety, tolerability, and efficacy of cannabidiol was assessed in the CARE 5 OLE study. Confidence in the effect estimates from the OLE study was limited, primarily due to the CARE 5 trial not being designed to assess the efficacy of cannabidiol compared with an active treatment or placebo, a diminishing sample size over the study period, and the open-label design, which may have biased outcomes such as HRQoL and harms. As patients with LGS have limited treatment options for seizure control, both the patient group and the clinical expert input emphasized that there is an unmet need for new medications to achieve seizure control, and ideally seizure freedom, in patients with seizures associated with LGS.

Introduction

The objective of this report is to review and critically appraise the evidence submitted by the sponsor on the beneficial and harmful effects of cannabidiol 100 mg/mL oral solution for use as adjunctive therapy for the treatment of seizures associated with LGS in patients aged 2 years and older.

Disease Background

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following has been summarized and validated by the CADTH review team.

LGS is a complex epilepsy syndrome associated with multiple seizure types that vary across people with the condition. LGS is a rare and lifelong condition that presents in the first 4 years of life (peak age of onset: 3 to 5 years) in in children who had otherwise experienced age-appropriate development. LGS is associated with refractory and multiple treatment-resistant seizure types, cognitive and behavioural impairments, and poor outcomes into adulthood.^1-4 According to the clinical experts consulted by CADTH, LGS has various etiologies and people with LGS exhibit multiple seizure types with distinctive EEG features. The most prominent seizure types in LGS are tonic seizures (stiffening of the body, and sometimes upward eye gaze, dilated pupils, and altered breathing patterns) and atypical absence seizures (staring spells with distinctive EEG pattern), followed by myoclonic jerks (sudden muscle jerks), tonic or atonic drops (brief loss of muscle tone), generalized tonic-clonic seizures (muscle stiffness and rhythmic jerking), and focal seizures.³ Atonic and tonic seizures can be accompanied by dangerous falls or “drop seizures” that often lead to injury.⁵ High seizure frequency is an independent predictor of early death in people with chronic epilepsy.²³ Indeed, LGS is considered a life-threatening condition associated with high rates of SUDEP and a risk of death among children 14 times higher than that of children, adolescents, and young adults in the US general population.^6,7 For children with LGS, the reported mortality ratio is 13.92, compared to 3.11 for any kind of epilepsy,⁷ and the risk of death from neurologic causes (e.g., prolonged seizures, status epilepticus) is 179 greater in children with LGS compared to the general population.⁷ Clinical consensus suggests that the best prevention strategy for death related to epilepsy is to reduce the frequency of seizures through effective treatments.^3,6,8,11

Refractory epilepsy (i.e., epilepsy for which the patient, after adequate trials of 2 appropriately chosen and used ASMs, as monotherapies or in combination, does not experience sustained seizure freedom)²⁴ is associated with poor prognosis in people with LGS. Approximately 80% to 90% of people with LGS continue having seizures into adulthood, and nearly all have cognitive and behavioural impairments along with intellectual functioning and developmental delays within 5 years of diagnosis.^3,6,8 As a result, people with LGS often require lifelong round-the-clock care that demands adjustments in virtually all aspects of the lives of their caregivers and family members. The high seizure burden associated with LGS therefore negatively impacts HRQoL for both people with the condition and their caregivers, as well as siblings and other household members.^3,25-27 These impacts include high caregiver anxiety driven by the potential for injury, cognitive decline, or death of the person with LGS; financial burden; and, in applicable household environments, sibling psychological distress from witnessing seizures, not getting as much attention from parents, and experiencing concern for their parents’ stress or unhappiness.^26,27

Currently, there are no specific diagnostic tests or biological markers for the diagnosis for LGS.^6,8 Instead, LGS is diagnosed clinically, based on the presence of multiple seizure types, moderate to severe cognitive and behavioural impairment, and an abnormal EEG pattern of slow spike-wave complexes.^28,29 LGS presentation is diverse, and symptoms can evolve over time, resulting in a delay between initial symptom onset and emergence of the full triad of clinical features.²⁹ The clinical experts consulted by CADTH also noted that part of the diagnostic process and management involves the examination of underlying structural abnormalities. In some cases, enhanced access to advanced genetic testing, including epilepsy gene panels and whole exome or genome sequencing, has identified genetic etiologies for patients who would have previously had unknown causes of LGS.

LGS affects between 3% and 10% of children with epilepsy, more commonly males. The peak age for onset is 3 to 5 years, with extreme incidence occurring in the first and 10th years of life.⁹ Estimates of the incidence and prevalence rates of LGS from studies worldwide are highly variable due to variability in diagnosis criteria, the size of the study populations, the ages of the enrolled patients, and the length of the follow-up period: prevalence ranged from 4.2 to 60.8 per 100,000 people across studies for “probably” LGS and 2.9 to 28 per 100,000 for a confirmed and narrow definition of LGS.³⁰ Approximately 48,000 children and adults in the US are estimated to have LGS.³¹ The prevalence of LGS in Canada, as estimated by the sponsor, is 12 in 100,000 people.¹⁰

Standards of Therapy

Contents within this section have been informed by materials submitted by the sponsor and clinical expert input. The following has been summarized and validated by the CADTH review team.

Based on clinical expert input solicited by CADTH for the purpose of this review, the primary objective in managing LGS is the reduction of seizures through the administration of ASM. Although available pharmacological treatments for LGS primarily target seizure frequency, the clinical experts noted that there is growing emphasis on measuring the impact of therapies on nonseizure outcomes (e.g., behaviour, independence in activities of daily living, patient and caregiver quality of life) in developmental and epileptic encephalopathies such as LGS.

The current ASMs that have been evaluated in the context of randomized trials and are available for patients with LGS are sodium valproate ordivalproex sodium (expert panel recommendation),³² lamotrigine and topiramate (level A evidence for drop attacks),³³ rufinamide (effective add-on treatment for motor and nonmotor seizures as per 3 class I studies with level A evidence),³³ clobazam (probably an effective add-on treatment for motor and nonmotor seizures per 2 class II studies),³³ felbamate (adjunctive therapy as per the American Academy of Neurology and the American Epilepsy Society; FDA approved, but not available in Canada),^33,34 fenfluramine (recently approved in the US after the OLE phase III trial NCT 03355209, but not available in Canada), and cannabidiol (FDA approved).^35-38 Other ASMs commonly used, in the absence of randomized trial evaluations, include valproic acid ordivalproex sodium (first treatment option by expert panel),³² levetiracetam, ethosuximide, perampanel, brivaracetam, and zonisamide (unavailable in Canada).^39,40 In Canada, the only drugs currently indicated specifically for LGS are rufinamide¹² (to be taken in conjunction with 2 or more ASMs) and lamotrigine¹³ (to be taken with other ASMs). While lamotrigine is available through the Ontario Drug Benefit program, rufinamide must be obtained through the Exceptional Access Program.¹⁴ In addition to ASMs, other therapeutic options include dietary therapies such as ketogenic, modified Atkins, or low–glycemic index diets; neuromodulation including VNS and deep brain stimulations; and nonresective surgeries such as corpus callosotomy. In some cases, although uncommon, patients may benefit from surgical resection if the source of seizure activity can be identified.

Based on the Ontario Epilepsy Guidelines, first-line ASMs for LGS are rufinamide and valproic acid (indicated generally for epilepsy), while clobazam, lamotrigine, topiramate, and perampanel are recommended as adjunctive treatment.⁴¹ The clinical experts consulted by CADTH, however, suggested valproate as first-line therapy, followed by lamotrigine (adjunctive), rufinamide (second adjunctive), and then topiramate or clobazam or felbamate (subsequent adjunctives) in accordance with the treatment algorithm proposed by Cross et al.³² Recently, the LGS Transition of Care Advisory Board outlined a 3-tier model in which valproate, clobazam, lamotrigine, rufinamide, topiramate, and cannabidiol are considered tier 1 therapies; levetiracetam, perampanel, and zonisamide are considered tier 2 therapies; and felbamate, lacosamide, brivaracetam, and cenobamate are considered tier 3 therapies.³⁸ If the LGS becomes refractory, however, there are no specific recommended combinations of ASMs; instead, a heterogeneous combination of ASMs is used to optimize seizure control.³⁷

The clinical experts consulted by CADTH noted that purified cannabidiol from a licensed producer (“artisanal CBD” or medical cannabis) has been available in Canada for years and has been used extensively for the treatment of drug-resistant epilepsy in children and adults, including patients with LGS, albeit often at considerably lower doses than those used in clinical trials of cannabidiol.

Drug Under Review

The key characteristics of cannabidiol are summarized in Table 3, alongside other treatments available for seizures associated with LGS.

The exact mechanism of action of cannabidiol is unknown.¹⁵ Cannabidiol reduces neuronal hyperexcitability through the modulation of intracellular calcium via the GPR55 and TRPV1 receptors and the modulation of adenosine-mediated signalling through the inhibition of adenosine transport via the ENT1 transporter.¹⁵

Cannabidiol received a Notice of Compliance from Health Canada on November 15, 2023, through the standard review pathway as adjunctive therapy for the treatment of seizures associated with LGS, Dravet syndrome, or tuberous sclerosis complex in patients aged 2 years and older. The sponsor requested that CADTH review each neurologic disorder as a separate submission. CADTH accepted the sponsor’s request on June 21, 2023. Accordingly, the indication of focus for the current review of cannabidiol is for use as adjunctive therapy for the treatment of seizures associated with LGS in patients aged 2 years and older.

The product monograph–recommended starting dosage of cannabidiol is 2.5 mg/kg twice daily.¹⁵ After 1 week, the dosage should be increased to a maintenance dosage of 5 mg/kg twice daily (10 mg/kg/day). Based on individual clinical response and tolerability, each dose can be increased in weekly increments of 2.5 mg/kg, administered twice daily (5 mg/kg/day), up to a maximum recommended dosage of 10 mg/kg twice daily (20 mg/kg/day). For patients in whom a more rapid titration from 10 mg/kg/day to 20 mg/kg/day is warranted, the dose may be increased no more frequently than every other day. The product monograph also provides a dosing adjustment for patients with moderate or severe hepatic impairment, due to risk of hepatocellular injury related to cannabidiol.¹⁵

According to the product monograph, cannabidiol can cause somnolence and sedation. Other common adverse reactions associated with cannabidiol use are decreased appetite; diarrhea; transaminase elevations and liver impairment and toxicity; fatigue, malaise, and asthenia; rash; insomnia, sleep disorders and poor-quality sleep; and infections.¹⁵

Table 3: Key Characteristics of Cannabidiol and Other ASMs That Constitute Usual Care

ALT = alanine transaminase; ASM = antiseizure medication; AST = aspartate transaminase; CNS = central nervous system; DRESS = drug reaction with eosinophilia and systemic symptoms; GABA = gamma-aminobutyric acid; LGS = Lennox-Gastaut syndrome; TSC = tuberous sclerosis complex.

^aHealth Canada–approved indication.

^bAs recommended by the Ontario Epilepsy Guidelines.

Sources: Sponsor’s clinical evidence summary;⁴² Health Canada–approved product monographs for rufinamide, valproic acid, clobazam, lamotrigine, levetiracetam, topiramate, perampanel, lacosamide, and brivaracetam;^43-52 sponsor-submitted proposed product monograph for cannabidiol;¹⁵ American Academy of Neurology–American Epilepsy Society Guidelines 2004 and 2018;^33,53 Ontario Epilepsy Guidelines 2020.⁴¹

Stakeholder Perspectives

Patient Group Input

This section was prepared by the CADTH review team based on the input provided by patient groups. The full original patient input(s) received by CADTH have been included in the Stakeholder section of this report.

The CEA provided patient group input. The CEA is a network of organizations supporting and advocating for people living with epilepsy and their families. Patient input was sourced from the knowledge and experiences of patients, caregivers, clinicians, volunteers, donors, and funders, collected from 24 member associations. According to input from the CEA, current treatments may fail in 30% of patients with epilepsy. According to the input, people living with uncontrolled epilepsy are often socially isolated due to stigma and to fear of rejection in social, work, and educational settings. Patients often experience depression and anxiety upon initial diagnosis and often continuously experience these conditions if ASMs stop working. Caregivers and family members are also impacted by epilepsy as their lives revolve around the seizures experienced by patients under their care. Anxiety among caregivers is common as they worry about when the next seizure will occur, the consequences of the epilepsy, and how to navigate social gatherings (e.g., a young person with epilepsy gets invited to a birthday party). In addition, caregivers often experience compassion fatigue since they cannot leave the person with epilepsy alone, are often sleep deprived due to sleep interruptions or anxiety, and can be affected by side effects the patient may experience from medications (e.g., mood swings, sexual dysfunction, suicidal thoughts, memory loss, fatigue, exhaustion). Both patients and caregivers emphasized the importance of treatment that results in seizure freedom. However, patients and caregivers noted that they would accept a treatment that resulted in a reduction in the absolute number of seizures, as even a reduction in seizure frequency could improve overall quality of life. Because people with intractable epilepsy are very often unemployed or underemployed, not covered under employer-funded insurance plans, and/or have restricted income, most drugs to treat their epilepsy are inaccessible. Accordingly, the CEA stressed the importance of new medications being placed on the provincial formulary so that patients with intractable epilepsy have access to novel treatments. Input provided by the CEA did not include experience with cannabidiol.

Clinician Input

Input From Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise in the diagnosis and management of the condition for which the drug is indicated. Clinical experts are a critical part of the review team and are involved in all phases of the review process (e.g., providing guidance on the development of the review protocol, assisting in the critical appraisal of clinical evidence, interpreting the clinical relevance of the results, and providing guidance on the potential place in therapy). The following input was provided by 2 clinical specialists with expertise in the diagnosis and management of LGS.

Unmet Needs

The clinical experts consulted by CADTH for the purpose of this review noted that, despite the multiple treatment options available, there are currently no treatments available to reverse the course of LGS. Moreover, the overall prognosis of LGS remains unfavourable, and not all patients experience response to the available treatments. LGS often becomes refractory to current options and requires multiple medication adjustments and combinations. In addition, the side effects of taking multiple ASMs (e.g., sedation, cognitive slowing) are associated with reduced quality of life, to the point where treatment becomes intolerable and must be discontinued. This challenge is further exacerbated by drug-drug interactions, which are complex and require close clinical monitoring and periodic bloodwork for drug levels, including to guard against incipient hepatotoxicity or bone marrow suppression. The clinical experts added that there is a need for treatment that can meaningfully improve quality of life for both patients and their caregivers.

Place in Therapy

The clinical experts explained that rational polytherapy involves combining ASMs with different mechanisms of action to ensure better seizure control and fewer AEs. While the mechanisms of action of currently available therapies for LGS mostly involve modulating the sodium and calcium channel, gamma-aminobutyric acid, or glutamate receptors, it is hypothesized that cannabidiol works through the endocannabinoid system, which regulates neural excitability. The clinical experts noted that despite its novel mechanism of action compared to available therapies, cannabidiol does not address the underlying disease process any more than other available treatments. Accordingly, the clinical experts suggest that cannabidiol would complement other available treatments as a symptom management treatment. The clinical experts opined that cannabidiol could be combined with 1 or 2 first-line ASMs. The clinical experts also felt that it would be reasonable to require adequate trials of 1 or 2 other ASMs before the use of cannabidiol.

The clinical experts noted that purified cannabidiol, referred to as cannabis for medical purposes, has been available in Canada for several years from licensed producers and has been used extensively (although often at lower doses) for the treatment of drug-resistant epilepsy in children and adults, including patients with LGS. These artisanal cannabidiol treatments are generally not covered by provincial formularies or most private drug insurance plans. The clinical experts expected that the approval of cannabidiol would lead to a shift away from the use of artisanal cannabidiol in jurisdictions where cannabidiol is reimbursed by either public or private drug insurance plans.

Patient Population

According to input from the clinical experts, it is difficult to predict which patients with LGS would be most likely to benefit from cannabidiol. According to the clinical experts, there have been no studies exploring antiseizure treatment response based on LGS etiology. Patients who have not experienced response with multiple ASMs are generally less likely to experience response to the next ASM; however, these are the patients who are most in need of novel therapies. As noted by the clinical experts, idiopathic cases of LGS, with no neurologic impairment before the onset of seizures, tend to respond better to treatment, whereas LGS cases converted from West syndrome or caused by extensive brain malformations or acquired brain lesions will often have less favourable outcomes.

The clinical experts suggest that patients should be screened for treatment according to clinician judgment based on seizure characterization and frequency, etiology investigation, and previous ASMs trialled, along with EEG interpretation.

Assessing the Response to Treatment

According to the clinical experts’ input, a clinically meaningful response to treatment in epilepsy is assessed in terms of the median reduction in seizure frequency over 28-day periods, a reduction in the frequency of seizures of 50% or more, and seizure freedom rates (i.e., reduction of total seizures per day and seizure-free days per month). The clinical experts added that seizure frequency should be assessed every 4 weeks. One clinical expert also added that assessing the frequency of motor and nonmotor seizures rather than drop seizures (i.e., motor seizures that could result in a fall) would be more meaningful. The clinical experts noted that improved quality of life for both patients and caregivers is an important secondary outcome.

Discontinuing Treatment

According to the clinical experts’ input, treatment with cannabidiol should be discontinued if patients develop persistent and progressive elevation of transaminases or recurrent vomiting and diarrhea, which would compromise the absorption of ASMs. In addition, treatment with cannabidiol should be reassessed if a patient develops status epilepticus with no other reasonable explanation. The clinical experts noted that concomitant use of cannabidiol and clobazam requires careful monitoring of clobazam and its metabolite (desmethylclobazam) levels, which are common causes of toxicity and may require clobazam dosing adjustments due to the risk of persistent somnolence.^54,55

Prescribing Considerations

According to input received from the clinical experts, the prescribing and monitoring of cannabidiol for LGS should be limited to neurologists who have special expertise in the management of epilepsy in pediatric or adult patients. The clinical experts’ reasoning for this criterion is that specialized expertise is required for accurate LGS diagnosis and for the surveillance and management of drug-drug interactions.

Clinician Group Input

No clinician group input was received for this submission.

Drug Program Input

The drug programs provide input on each drug reviewed through CADTH’s reimbursement review processes by identifying issues that may impact their ability to implement a recommendation. The implementation questions and corresponding responses from the clinical experts consulted by CADTH are summarized in Table 4.

Clinical Evidence

The objective of CADTH’s Clinical Review Report is to review and critically appraise the clinical evidence submitted by the sponsor on the beneficial and harmful effects of cannabidiol 100 mg/mL oral solution for use as adjunctive therapy in the treatment of seizures in patients with LGS. The focus will be placed on comparing cannabidiol to relevant comparators and identifying gaps in the current evidence.

A summary of the clinical evidence included by the sponsor in the review of cannabidiol is presented in 2 main sections, with CADTH’s critical appraisal of the evidence included at the end of each section. The first main section, the systematic review, includes RCTs that were selected according to the sponsor’s systematic review protocol. CADTH’s assessment of the certainty of the evidence in this first section using the GRADE approach follows the critical appraisal of the evidence. The second main section details 1 sponsor-submitted long-term extension study. No indirect evidence or additional studies that were considered to address important gaps in the systematic review evidence were submitted by the sponsor.

Included Studies

Clinical evidence from the following is included in the CADTH review and appraised in this document:

• 2 RCTs identified in systematic review (the CARE 3 and CARE 4 trials)^16,17

• 1 long-term extension study (the CARE 5 trial).²²

Systematic Review

Contents within this section have been informed by materials submitted by the sponsor. The following has been summarized and validated by the CADTH review team.

Description of Studies

Characteristics of the included studies are summarized in Table 5.

Table 5: Details of Studies Included in the Systematic Review

ASM = antiseizure medication; CTR = Clinical Trial Record; EEG = electroencephalogram; LGS = Lennox-Gastaut syndrome; QOLCE = Quality of Life in Childhood Epilepsy; QOLIE-31-P = Quality of Life in Epilepsy, version 2; RCT = randomized controlled trial; S/CGIC = Subject/Caregiver Global Impression of Change; S/CGICSD = Subject/Caregiver Global Impression of Change in Seizure Duration; VNS = vagus nerve stimulation.

Note: Health Canada reviewer’s report was included.⁶⁸

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

Two sponsor-conducted studies were included in the systematic review: the CARE 3 and CARE 4 trials.^16,17

Both the CARE 3 and CARE 4 trials were multicentre, randomized, double-blind, phase III RCTs evaluating the efficacy of cannabidiol as adjunctive treatment in reducing drop seizure frequency in patients with LGS.

In the CARE 3 trial, 225 patients across 29 sites in 4 countries (US, Spain, France, and UK) were randomized to receive treatment with either cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, or placebo. A schematic of the CARE 3 study design is presented in Figure 1. Prior to randomization in the CARE 3 trial, patients entered (at visit 1) a 28-day screening period that included tests to detect the presence of delta-9-tetrahydrocannabinol metabolites to confirm that cannabis was not taken immediately before entering the trial and to ensure that medications or interventions for epilepsy were stable for 4 weeks before screening. This period was also used to establish the baseline number and type of drop and nondrop seizures. To ensure the correct trial population was randomized, an independent review by the Epilepsy Study Consortium was conducted to verify the seizure types experienced by screened patients. Confirmation of seizure types from the Epilepsy Study Consortium was required before randomization at visit 2. Eligible patients were randomly allocated with a 1:1:1 ratio to 1 of the 3 treatment arms using an IVRS. Patients in the placebo group were split into 2 equivalent cohorts: half receiving 10 mg/kg/day dosing volumes, and half receiving 20 mg/kg/day dosing volumes. The allocation of investigational medicinal product was done according to randomization schedules produced by an independent statistician. The randomization was stratified by age group (2 to 5 years, 6 to 11 years, 12 to 17 years, and 18 to 55 years). The IVRS identified the pack number to be dispensed to the patient at each visit according to the treatment assigned in the randomization schedule. Patients returned to the clinic for an end-of-treatment visit after the 14-week treatment period on day 99 (visit 8), or earlier if they withdrew prematurely. Patients were invited to continue to receive cannabidiol in an OLE trial under a separate protocol immediately following visit 8.

Figure 1: CARE 3 Study Design

GWP42003-P = cannabidiol; OLE = open-label extension.

^a For patients in France only, the visit window was fewer than 3 days rather than at least 3 days.

^b For patients in France only, visit 7 was completed at the clinic and was not conducted by telephone.

^c Visit 9 was for patients who did not enter the OLE trial or who withdrew from the study early; could be conducted by telephone.

^d Visit 10 was for patients who did not enter the OLE trial at visit 8 or for those who withdrew early and tapered the investigational product. Patients who completed treatment but opted not to enter the OLE trial, or who withdrew from the trial early, had weekly (± 3 days) safety telephone calls from visit 9 (or date of final dosing) until visit 10.

Sources: Clinical Study Report for CARE 3 trial;¹⁶ sponsor’s summary of clinical evidence.

In the CARE 4 trial, 171 patients across 24 sites in 3 countries (US, Netherlands, and Poland) were randomized to receive treatment with either cannabidiol 20 mg/kg/day or placebo. A schematic of the CARE 4 study design is presented in Figure 2. Similar to the CARE 3 trial, a 28-day screening period was included to perform tests to detect the presence of delta-9-tetrahydrocannabinol metabolites, to ensure that all patients’ medications or interventions for epilepsy were stable for 4 weeks before screening, and to establish the baseline number and type of drop and nondrop seizures. Eligible patients were randomly allocated to 20 mg/kg/day cannabidiol or an equivalent volume of placebo in a 1:1 ratio using an IVRS. The allocation of investigational medicinal product, randomization stratification factors, and dispensing of treatment were done in the same manner as described for the CARE 3 trial.

Figure 2: CARE 4 Study Design

CBD-OS = cannabidiol; OLE = open-label extension.

^a Visit 9 was for patients who did not enter the OLE trial at visit 8 or for those who withdrew early and tapered the investigation product. Patients who completed treatment but opted not to enter the OLE trial, or who withdrew from the trial early, had weekly (± 3 days) safety telephone calls from visit 9 (or date of final dosing) until visit 10.

^b Visit 10 was for patients who did not enter the OLE trial or who withdrew from the study early; could be conducted by telephone.

Sources: Clinical Study Report and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

Protocol Amendments

The CARE 3 and CARE 4 trials underwent 7 and 4 protocol amendments, respectively.

In the CARE 3 trial, the intention-to-treat analysis set was revised per FDA request; statistical analysis testing of the primary end point and additional statistical analysis of the secondary end points were modified; and the statistical methods related to the handling of missing data were changed in amendment 4 (February 17, 2015). Statistical analysis for the primary and secondary end points was amended again in amendment 5 (March 27, 2015) and amendment 6 (June 11, 2015). Other amendments were mostly administrative changes and clarifications.

In the CARE 4 trial, recommendations received from the FDA were incorporated at the time of protocol amendment 3 (June 3, 2015) and included adding stratification by age across treatment arms, follow-up procedures for potential cases of drug-induced liver injury, and updated recommended target doses and titration regimens based on the recommendations of the data safety monitoring committee of the CARE 3 trial. Other amendments were mostly administrative changes and clarifications.

Populations

Inclusion and Exclusion Criteria

Key inclusion and exclusion criteria for the CARE 3 and CARE 4 trials are summarized in Table 5. Patients eligible for inclusion in the CARE 3 and CARE 4 trials were aged 2 to 55 years and had a clinical diagnosis of LGS, which included written documentation of having a history of a slow (< 3.0 Hz) spike-and-wave pattern in an EEG before their enrolment into the baseline period and evidence of more than 1 generalized seizure, including drop seizures (atonic, tonic, or tonic-clonic), for at least 6 months. During screening, care was taken not to include patients with benign myoclonic epilepsy of infancy, atypical benign partial epilepsy (pseudo-Lennox syndrome), or continuous spike-waves of slow sleep. In addition, patients eligible for study inclusion were required to have experienced at least 2 drop seizures each week during the first 28 days of the baseline period; to have experienced documented treatment failures on more than 1 ASM; and to be taking 1 or more ASMs at a dose that had been stable for at least 4 weeks before screening. All medications or interventions for epilepsy, including ketogenic diet and VNS, must have been stable for 4 weeks before screening, and the patient had to be willing to maintain a stable regimen throughout the trial. The key exclusion criteria were the etiology of the patient’s seizures being a progressive neurologic disease, the patient taking more than 4 concurrent ASMs, or the patient having taken corticotropins in the 6 months before screening. Any patients who were currently using or had used recreational or medicinal cannabis or synthetic cannabinoid-based medications within the 3 months before trial entry and were unwilling to abstain for the duration for the study were also excluded.

Interventions

Definition of Treatment and Maintenance Period

In both the CARE 3 and CARE 4 trials, the treatment period was defined as day 1 to day 99, or the date of last dose as recorded on the “end of treatment study outcome” case report form, whichever occurred earlier. The maintenance period was defined as day 15 to day 99, or the date of the last dose as recorded on the case report form.

Investigational Product

Cannabidiol

Patients randomized to receive cannabidiol received the treatment at a dosage of either 10 mg/kg/day (CARE 3 trial) or 20 mg/kg/day (CARE 3 and CARE 4 trials), administered orally by the patient or their caregiver twice each day: half in the morning and half in the evening. Patients were titrated from a starting dosage of 2.5 mg/kg/day up to 10 mg/kg/day over 7 days, or up to 20 mg/kg/day over 11 days, and remained at this dosage for the duration of the treatment period.

The dosage titration for cannabidiol 10 mg/kg/day was as follows:

• 2.5 mg/kg/day on days 1 and 2

• 5.0 mg/kg/day on days 3 and 4

• 7.5 mg/kg/day on days 5 and 6

• 10.0 mg/kg/day from day 7 onward for maintenance.

The dose titration for cannabidiol 20 mg/kg/day was as follows:

• 2.5 mg/kg/day on days 1 and 2

• 5.0 mg/kg/day on days 3 and 4

• 7.5 mg/kg/day on days 5 and 6

• 10.0 mg/kg/day on days 7 and 8

• 15.0 mg/kg/day on days 9 and 10

• 20 mg/kg/day from day 11 onward for maintenance.

The first dose and last dose of the cannabidiol were taken at visit 2 (day 1) and visit 8, respectively. Following the end of the treatment period, treatment with cannabidiol was gradually tapered by 10% each day over a 10-day taper period for patients who did not immediately enter the OLE trial. The taper period was interrupted if the patient requested to enter the OLE trial within a 7-day time frame. For patients who withdrew from the CARE 3 and CARE 4 trials early, the taper period followed the withdrawal visit unless continued dosing was not possible due to an AE.

Placebo

Patients randomized to the control group in both studies received volume-matched placebo treatment following the same titration schedule up to 7 or 10 days, depending on whether treatment with placebo was being compared to cannabidiol 10 mg/kg/day or 20 mg/kg/day.

Withdrawal From Treatment

All patients had the right to withdraw at any point during the treatment. In addition, any patients meeting the following criteria were withdrawn from the trial:

• administrative decision by the investigator, sponsor, or a regulatory authority

• pregnancy

• protocol deviation considered to potentially compromise the safety of the patients

• withdrawal of patient consent or assent or withdrawal of parent or legal representative consent

• lost to follow-up

• ALT or AST more than 3 times the upper limit of normal and total bilirubin greater than 2 times the upper limit of normal or with an international normalized ratio greater than 1.5

• ALT or AST greater than 3 times the upper limit of normal with the appearance of fatigue, nausea, vomiting, right upper quadrant pain or tenderness, fever, rash, and/or eosinophilia (> 5%)

• ALT or AST greater than 8 times the upper limit of normal

• ALT or AST greater than 5 times the upper limit of normal for more than 2 weeks.

Patients may have also been withdrawn from the trial if they met or experienced any of the following criteria: patient noncompliance; an AE that would compromise the continued safe participation of the patient in the trial based on the opinion of the investigator; evidence of drug abuse or diversion; suicidal ideation or behaviour of type 4 or 5 during the treatment period as evaluated with the Columbia–Suicide Severity Rating Scale.

Prior, Concomitant, and Prohibited Medications

Prior Medications

Prior to screening, patients were required to be taking 1 or more ASMs at a dose that had been stable for at least 4 weeks. Although nonpharmacological interventions for epilepsy (e.g., ketogenic diet, VSN) were not counted as ASMs, their use must also have been stable for 4 weeks before screening.

Concomitant Medications

Throughout the study period, dosing of concomitant ASMs and any nonpharmacological regimens for epilepsy was to remain stable. However, due to potential pharmacological interactions between cannabidiol and other concurrently administered medication, the dosage for concomitant ASMs was permitted to be modified based on patients’ clinical symptoms or changes in plasma levels following the administration of cannabidiol. Any modifications of the dosage of concomitant ASMs were based on clinical need following discussion with the medical advisors.

The use of rescue medication was permitted when necessary.

Prohibited Therapy

The following medications were not permitted during the study period, starting from the acquisition of patient consent or assent:

• Any new medications or interventions for epilepsy, including ketogenic diets and VSN, or change in dosing.

• Recreational or medicinal cannabis or synthetic cannabinoid-based medications within 3 months before or during the trial.

• Any other investigational products taken as part of a clinical trial within 6 months before or during, the trial.

• Long-term systemic steroids, excluding inhaled medication for asthma treatment, or any other daily medications known to exacerbate epilepsy. Exceptions were made for prophylactic medication, for example for idiopathic nephrotic syndrome or asthma.

Outcomes

The primary, key secondary, other secondary, and exploratory outcomes for both the CARE 3 and CARE 4 trials were the same (Table 5).

A list of the efficacy end points assessed in this Clinical Review Report is provided in Table 6, followed by descriptions of the outcome measures. Summarized end points are based on outcomes included in the sponsor’s summary of clinical evidence as well as any outcomes identified as important to this review according to the clinical experts consulted by CADTH and stakeholder input from patient and public drug plans. Using the same considerations, the CADTH review team selected the end points that were considered most relevant to inform CADTH’s expert committee deliberations and finalized this list of end points in consultation with members of the expert committee. All summarized efficacy end points were assessed using GRADE, and they included percent change in number of drop seizures; proportion of patients who experienced at least a 25%, at least a 50%, at least a 75%, or a 100% reduction in drop seizures from baseline; number of inpatient hospitalizations due to epilepsy; and number of drop seizure–free days. Patient-reported outcomes that reflected a patient’s quality of life or burden from disease and treatment were considered important by the patient groups and the clinical experts consulted by CADTH for the purpose of this review. While impression of change, as measured by Subject/Caregiver Global Impression Change score as a key secondary outcome, was considered by the clinical experts consulted by CADTH to be a practical measure of HRQoL in the clinical practice setting, the clinical experts added that other measures of HRQoL better captured and quantified the multiple domains of quality of life. According to input from the clinical experts, the QOLCE and QOLIE-31-P scores captured the multiple domains that affect patient’s quality of life (e.g., sleep, cognition, and patient or caregiver assessment) and were therefore assessed using GRADE. Select notable harms outcomes considered important for informing CADTH’s expert committee deliberations were also assessed using GRADE. The following notable harms were recognized as important based on the product monograph and according to the clinical experts consulted by CADTH and could be associated with treatment discontinuation SAE and hepatocellular injury.

Efficacy outcomes that were not assessed using GRADE are summarized in Appendix 1 (Table 26).

Drop Seizures

A drop seizure was defined as an attack or spell (atonic, tonic, or tonic-clonic) involving the entire body, trunk, or head that led or could have led to a fall, an injury, slumping in a chair, or the patient’s head hitting a surface. Drop seizures were the basis of several study end points, including the primary end point: the percent change from baseline in the number of drop seizures during the treatment period. Patients or their caregivers were instructed to record the number and type of drop seizures (atonic, tonic, tonic-clonic) and nondrop seizures from the screening (visit 1) until the completion of dosing (visit 8 or withdrawal visit) using the IVRS diary.

Table 6: Outcomes Summarized From the Studies Included in the Systematic Review

QOLCE = Quality of Life in Childhood Epilepsy; QOLIE-31-P = Quality of Life in Epilepsy, version 2; SAE = serious adverse events.

^aStatistical testing for these end points was adjusted for multiple comparisons (e.g., hierarchal testing).

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

The frequency of drop seizures was based on a 28-day average. Percent change from baseline in drop seizure frequency, averaged per 28 days, during the treatment period was calculated as follows:

The frequency during each period was based on 28-day averages, calculated as follows:

A minimally important difference (MID) for the reduction of seizure frequency from baseline in patients with milder forms of epilepsy was considered by the sponsor to be at least 50% based on historical regulatory approval processes from the European Medicines Agency and the FDA. Currently, there is no empirically estimated MID for seizure reduction specifically for patients with LGS who experience more severe forms of seizures that are treatment refractory and who may benefit from a modest reduction in seizure frequency.⁶⁹ According to input from the clinical experts consulted by CADTH, a 25% to 30% reduction in mean or median drop seizure frequency over the placebo response would be considered clinically meaningful.

The number of drop seizure–free days was assessed as an exploratory outcome. The number of drop seizure–free days during each period was based on 28-day averages and calculated as:

The clinical experts could not estimate a threshold of how many additional drop seizure–free days over the placebo response may be considered clinically meaningful.

Treatment Response

The key secondary outcome was the number of patients who experienced at least a 50% reduction in drop seizures from baseline during the treatment periods. Other response thresholds were examined: patients who experienced at least a 25%, at least a75%, or 100% reduction in drop seizures from baseline were assessed as secondary outcomes. As previously mentioned, there are not empirically estimated MIDs for each threshold of response or for which of the thresholds defines a clinically important change or difference. According to input from the clinical experts consulted by CADTH for the purpose of this review, the following clinically meaningful differences in patients who experienced treatment response between the cannabidiol and placebo treatment groups may be considered:

• A between-group difference of 20% to 30% in patients who experienced treatment response at the greater than or equal to 25% and greater than or equal to 50% thresholds.

• A between-group difference of 15% to 20% in patients who experienced treatment response at the greater than or equal to 75% threshold.

• A between-group difference of 10% in patients who experienced treatment response at the 100% threshold.

Inpatient Hospitalization due to Epilepsy

Inpatient hospitalization due to epilepsy was a secondary outcome in both trials. Whether hospitalization was due to epilepsy was based on investigator opinion. According to input from the clinical experts consulted by CADTH for the purpose of this review, a 10% reduction in the number of hospitalizations due to epilepsy between the cannabidiol and placebo treatment groups would be considered clinically meaningful.

Health-Related Quality of Life

HRQoL as measured by the QOLCE or the QOLIE-31-P score was a secondary outcome in the included pivotal trials.

The QOLCE score is a disease-specific, parent-reported measure of HRQoL for children and adolescents with epilepsy. The QOLCE score captures the following domains of HRQoL: physical activities, well-being, cognition, social activities, behaviour, general health, and quality of life, which are grouped into 16 subscales. The subscale and overall quality-of-life scores range from 0 to 100, with a higher score indicative of better quality of life. There is no established MID for QOLCE score for patients with LGS or other seizure disorders.

The QOLIE-31-P is a self-administered, disease-specific survey of HRQoL in adults aged 18 years and older with epilepsy. The scale consists of 31 items related to a patient’s perception of their health and daily activities, which are grouped into 7 subscales: energy/fatigue, emotional well-being, social function, cognitive function, medication effects, seizure worry, and overall quality of life. The subscale and total scores range from 0 to 100, with a higher score indicative of better functioning. The last item of the QOLIE-31-P, which does not contribute to the total or subscale scores, asks patients to rank the importance of each subscale topic from 1 (the most important topic) to 7 (the least important topic). There is no established MID for QOLIE-31-P score in patients with LGS. However, an MID of 5.19 points in change in the total score for partial-onset seizures has been reported (refer to Table 7).⁷⁰

The QOLCE questionnaire was completed by the parent or caregiver of patients aged 2 to 18 years at visits 2 and 8. The QOLIE-31-P questionnaire was completed by patients aged 19 years or older at visits 2 and 8. For patients who were unable to complete the QOLIE-31-P independently, caregiver assistance was allowed.

Harms

An AE was defined as any new, unfavourable, unintended signs of symptoms (including abnormal laboratory findings) or diagnosis or worsening of a preexisting condition that was present following screening (visit 1) and at the posttreatment, safety follow-up visit (visit 10) and that may or may not have been considered to be related to the investigational product. All AEs were coded by system organ class, preferred term, and lowest level term using MedDRA, version 17.1. The number of patients who had the following AEs were summarized by treatment group as well as system organ class and preferred term: all-causality AEs; treatment-emergent AEs; serious treatment-emergent AEs; discontinuation of investigational product due to treatment-emergent AEs; and deaths. The following notable harms were of interest to the CADTH clinical review team: SAEs and hepatocellular injury.

Table 7: Summary of Outcome Measures and Their Measurement Properties

CHQ = Child Health Questionnaire; HRQoL = health-related quality of life; LGS = Lennox-Gastaut syndrome; MIC = minimally important change; MID = minimally important difference; POS = partial-onset seizure; QOLCE = Quality of Life in Childhood Epilepsy; QOLIE-31-P = Quality of Life in Epilepsy, version 2.

^aIn the CARE 3 and CARE 4 trials, the QOLCE questionnaire was administered to patients aged 2 to 18 years.

Sources: Sponsor’s clinical evidence summary;⁴² Sabaz et al. (2002); Sabaz et al. (2003); Cramer et al. (1998); Cramer et al. (2003); Borghs et al.(Borghs, 2012 #162).

Statistical Analysis

Sample Size Determination

Sample size determination in the CARE 3 trial was based on a 1:1:1 randomization ratio, stratification of age into 4 categories, and the assumption that patients in the placebo group would experience a mean reduction in drop seizure frequency of 18% (from baseline). In addition, it was assumed that although patients in the placebo group were split into 2 equivalent cohorts, 25 patients receiving 10 mg/kg/day dosing volumes and 25 patients receiving 20 mg/kg/day dosing volumes, the 2 cohorts could be pooled for the analyses of efficacy. Accordingly, a sample size of 50 patients per treatment group was determined to be sufficient to detect a difference of 32% between treatments (i.e., patients receiving cannabidiol would experience a reduction in drop seizure frequency of at least 50%) based on an SD of 56%, using a 2-tailed significance level of 0.05 and 80% power.

Sample size determination in the CARE 4 trial was based on a 1:1 randomization ratio and stratification of age into 4 categories. Using the assumptions and sample size and power calculation described in the CARE 3 trial, it was determined that a total of 100 patients would be sufficient.

Control of Type I Error

In both the CARE 3 and CARE 4 studies, type I error of the primary and key secondary efficacy end points was controlled using a hierarchical gate-keeping procedure. All statistical tests were conducted assuming a 2-sided significance level of 0.05. The null hypothesis of an end point was required to be rejected at the 2-sided significance level of 0.05 to test the hypothesis of the subsequent end point in the sequence. If a null hypothesis was not rejected, testing was stopped, and all subsequent analyses were declared not statistically significant. In both the CARE 3 and CARE 4 trials, the primary efficacy end point was tested first, and if positive, the key secondary end points were tested in sequential order. In the CARE 3 trial, each efficacy end point had 2 comparisons against placebo (cannabidiol 20 mg/kg/day and cannabidiol 10mg/kg/day). Of the outcomes reported in the clinical report, the first key secondary outcome – percent change from baseline in number of total seizures during the treatment period – was part of the testing order.

Statistical Methods

A summary of the statistical analysis employed in the CARE 3 and CARE 4 trials is presented in Table 8. In the CARE 3 trial, each end point had 2 comparisons against placebo (20 mg/kg/day cannabidiol and 10 mg/kg/day cannabidiol).

Seizure Frequency

The percent changes from baseline in drop seizure frequency, averaged per 28 days, were analyzed using a Wilcoxon rank sum test. An estimate of median difference between cannabidiol and placebo and an approximate 95% CI were calculated using the Hodges-Lehmann approach.

Patients with no seizures during the baseline period for a particular seizure type were excluded from the analysis of that seizure type.

Treatment Response

The Cochran-Mantel-Haenszel test, stratified by age group, was used to analyze the following outcomes:

• The proportion of patients who experienced a greater than or equal to 50% reduction in drop seizure frequency from baseline during the treatment period

• The proportion of patients who experienced a greater than or equal to 25% or greater than or equal to 75% reduction in drop seizure frequency from baseline

• The proportion of patients who experienced a 100% reduction in drop seizure frequency

The following statistics were presented: the proportion of patients who experienced at least a 25%, at least a 50%, at least a 75%, or a 100% reduction in drop seizure frequency; the difference in proportion with 95% CI; the estimated odds ratio (cannabidiol versus placebo) with 95% CI; and the P value from the Cochran-Mantel-Haenszel test. If no patients in a particular treatment group experienced at least a 25%, at least a 50%, at least a 75%, or a 100% reduction in drop seizure frequency, then the odds ratio and its associated 95% CI were not calculated.

HRQoL Outcomes

QOLCE and QOLIE-31-P scores were analyzed using an analysis of covariance, with baseline score and age group as covariates and treatment group as a fixed factor in the CARE 3 trial. In the CARE 4 trial, no statistical analysis of the QOLIE-31-P score was performed due to the limited number of patients who completed the questionnaire.

Inpatient Hospitalization

The number of inpatient hospitalizations due to epilepsy was analyzed using the analysis of covariance approach, with baseline number and age group as covariates and treatment group as a fixed factor.

Missing Data

Missing data were handled under the MNAR assumption. The impact of the MNAR assumption on the efficacy results — that missing data were systematic — was explored by multiple imputation on drop seizure frequency, based on time point, corresponding to each of the 14 days of the treatment period.

Sensitivity Analysis

Several nonparametric and regression-based sensitivity analyses were conducted for the primary efficacy end point to assess the various assumptions related to the distribution of the outcome measures and missing data. These analyses are summarized in Table 8.

Subgroup Analyses

In both the CARE 3 and CARE 4 trials, subgroup analyses were performed on the intention-to-treatment analysis set for the primary efficacy end point and on the secondary efficacy end points of proportion of patients who experienced at least a 25%, at least a 50%, at least a 75%, or a 100% reduction in drop seizure frequency and drop seizure freedom. All statistical significance was tested at the 0.05 level without formal adjustment for multiplicity. Subgroup analysis on just the treatment period was performed as described for the primary efficacy end point. Subgroup analyses for proportion of patients who experienced at least a 25%, at least a 50%, at least a 75%, or a 100% reduction in drop seizure frequency and for drop seizure freedom were performed using the Fisher exact test.

The following subgroups analyzed in both the CARE 3 and CARE 4 trials were of interest to CADTH:

• clobazam use (yes, no)

• baseline average drop seizure frequency per 28 days (≤ observed tertile 1, > observed tertile 1 to ≤ observed tertile 2, > observed tertile 2; the observed tertile values were rounded to the nearest 5)

• number of current ASMs (< 3 to ≥ 3)

• number of prior ASMs (< 6, ≥ 6).

Table 8: Statistical Analysis of Efficacy End Points

ANCOVA = analysis of covariance; CMH = Cochran-Mantel-Haenszel; ITT = intention to treat; LOCF = last observation carried forward; MCMC = Markov chain Monte Carlo; MI = multiple imputation; MNAR = missing not at random; NA = not applicable; NOCB = next observation carried backward; NR = not reported; PP = per protocol; QOLCE = Quality of Life in Childhood Epilepsy; QOLIE-31-P = Quality of Life in Epilepsy, version 2.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

Analysis Populations

A summary of the analysis sets used in the CARE 3 and CARE 4 trials is presented in Table 9.

Table 9: Analysis Populations of the CARE 3 and CARE 4 Trials

AE = adverse event; IMP = investigational medicinal product; ITT = intention to treat; PP = per protocol.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

Results

Patient Disposition

Patient disposition for the CARE 3 and CARE 4 trials is summarized in Table 10.

Table 10: Summary of Patient Disposition From Studies Included in the Systematic Review

ITT = intention to treat; NA = not applicable; PP = per protocol.

^aSix patients randomized to receive cannabidiol 10 mg/kg/day and 3 patients randomized to receive placebo 10 mg/kg/day were given dosing schedules for 20 mg/kg/day patients and received greater than 10 mg/kg/day dosing volumes before the mistake was corrected. These patients are analyzed according to the treatment group to which they were randomized, unless otherwise stated.

^bWithin the safety analysis set, the patients who mistakenly received greater than 10 mg/kg/day dosing volumes (refer to footnote “a”) are analyzed within the 20 mg/kg/day dose groups.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

In the CARE 3 trial, approximately 23% of the 293 patients screened were not randomized into the trial. The reasons for screen failure were not meeting the trial’s inclusion criteria (47.1%), other reasons (33.8%), meeting 1 or more exclusion criteria (8.8%), withdrawal from the study by patient or parent or guardian (5.9%), and investigator decision (4.4%). Of the 225 patients enrolled into the CARE 3 trial, 76 (33.8%) were randomized to receive cannabidiol 20 mg/day/day; 73 (32.4%) to receive cannabidiol 10 mg/kg/day; 38 (16.9%) to receive placebo 20 mg/kg/day; and 38 (16.9%) to receive placebo 10 mg/kg/day. All randomized patients received at least 1 dose of an investigational product and were thus included in the safety analysis set. Of those randomized, approximately 3% of patients in the cannabidiol 10 mg/kg/day and the 2 placebo groups discontinued the study. In the cannabidiol 20 mg/kg/day group, approximately 12% of patients discontinued the study, mainly due to AEs (5.3%) and withdrawal by the patient or their parent or guardian (2.6%).

In the CARE 4 trial, approximately 14.5% of the 200 patients screened were not randomized into the trial. The reasons for screen failure were not meeting the trial’s inclusion criteria (69.0%), other reasons (17.2%), withdrawal from study by patient or parent or guardian (10.3%), and investigator decision (3.4%). Of the 171 patients enrolled into the CARE 4 trial, 86 (50.3%) were randomized to receive cannabidiol 20 mg/kg/day and 85 (49.7%) to receive the volume-matched placebo. Approximately 16% of patients in the cannabidiol 20 mg/kg/day treatment group discontinued the study, mainly due to AEs (9.3%) and to meeting the withdrawal criteria (4.7%). In the volume-matched placebo group, 1 patient (1.2%) discontinued the study due to AEs.

Protocol Deviation

Protocol deviations in the CARE 3 and CARE 4 trials are summarized in Table 11.

In the CARE 3 trial, important deviations were recorded in 35.4% of patients in the cannabidiol 20 mg/kg/day group, 26.7% of patients in the cannabidiol 10 mg/kg/day group, and 34.2% of patients in the pooled placebo group. The most recorded protocol deviation was receipt of the wrong treatment or incorrect dose, which occurred more frequently in the cannabidiol 20 mg/kg/day group (23.2%) than in the cannabidiol 10 mg/kg/day group (14.9%) and the pooled placebo group (17.1%).

In the CARE 4 trial, important deviations were recorded in 8.1% of patients in the cannabidiol 20 mg/kg/day group and 10.6% of patients in the volume-matched placebo group. The most recorded protocol deviation was the receipt of the wrong treatment or incorrect dose, which occurred in 5.8% and 7.1% of patients in the cannabidiol 20 mg/kg/day treatment group and the volume-matched placebo group, respectively.

Table 11: Summary of Protocol Deviations From the CARE 3 and CARE 4 Trials

AED = antiepileptic drug; PK = pharmacokinetic; THC = tetrahydrocannabinol.

^aEach occurrence is counted, including multiple events of the same important protocol deviation for a single patient.

^bAmong patients younger than 12 years.

^cPatient was randomized without the laboratory results of visit 2 because the principal investigator did not receive the reports in a timely manner. The results were received after visit 3 and confirmed that hepatic enzymes were within permitted ranges at visit 1 and visit 2.

^dPatient was identified as having deviations relating to PK sampling times at visit 2 or visit 8, which may have affected cannabidiol (and major metabolite) PK summary statistics.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ additional information received from sponsor (November 28, 2023).

Baseline Characteristics

The baseline characteristics outlined in Table 12 are limited to those that are most relevant to this review or were considered by the review team to affect the outcomes or interpretation of the study results.

In the CARE 3 trial, the mean age of patients enrolled in the trial was 16.01 years (SD = 10.77 years) in the cannabidiol 20 mg/kg/day group, 15.43 years (SD = 9.48 years) in the cannabidiol 10 mg/kg/day group, and 15.29 years (SD = 9.26 years) in the pooled placebo group. Approximately 11.5% of the patients in the CARE 3 trial were aged 2 to 5 years, 32.4% were aged 6 to 11 years, 26.2% were aged 12 to 17 years, and 29.8% were aged 18 to 55 years. The mean patient weight at baseline was 41.0 kg (SD = 20.6 kg), 44.25 kg (SD = 26.2 kg), and 45.7 kg (SD = 23.2 kg) in the cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, and pooled placebo groups, respectively. In terms of patient seizure history at baseline, the median number of drop seizures varied from 80.25 (range, 8.7 to 1,278.3) to 86.9 (range, 14.0 to 7,494.0) per 28 days. Overall, 8.4% and 23.6% of patients were reported to be on a ketogenic diet or have had a VNS, respectively, at baseline. In the CARE 3 trial, the mean number of drop seizures at baseline was higher in the cannabidiol 10 mg/kg/day group (mean = 277.8; SD = 902.4) than in the cannabidiol 20 mg/kg/day group (mean = 142.4; SD = 174.3) and the pooled placebo group (mean = 192.9; SD = 313.0), and the proportion of patients reporting convulsive seizures longer than 30 minutes was higher in the cannabidiol 20 mg/kg/day group (10.5%) than in the cannabidiol 10 mg/kg/day group (2.7%) and the pooled placebo group (3.9%).

In the CARE 4 trial, the mean age of patients enrolled was 15.5 years (SD = 8.7 years) and 15.3 years (SD = 9.8 years) in the cannabidiol 20 mg/kg/day and the volume-matched placebo groups, respectively. Approximately 13.4% of patients in the CARE 4 trial were aged 2 to 5 years, 31.0% were aged 6 to 11 years, 21.6% were aged 12 to 17 years, and 33.9% were aged 18 to 55 years. The mean patient weight at baseline was 42.7 kg (SD = 22.6 kg) and 43.0 kg (SD = 23.0 kg) in the cannabidiol 20 mg/kg/day and the volume-matched placebo groups, respectively. In terms of seizure history, the median number of drop seizures was 71.4 (range, 10.3 to 855.9) and 74.7 (range, 11.2 to 3174.6) in the cannabidiol 20 mg/kg/day and the volume-matched placebo groups, respectively. Overall, 8.2% and 29.8% of patients were reported to be on a ketogenic diet or to have had a VNS, respectively, at baseline.

Table 12: Summary of Baseline Characteristics From the CARE 3 and CARE 4 Trials (ITT Population)

ASM = antiseizure medication; ITT = intention to treat; SD = standard deviation; VNS = vagus nerve stimulation.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence; exposure to Study Treatments.

Exposure and treatment adherence to the investigational products are summarized in Table 13.

Exposure

In the CARE 3 trial, the mean daily dose of the investigational products was 8 mL (SD = 4.1 mL) in the cannabidiol 20 mg/kg/day group, 4.4 mL (SD = 2.7 mL) in the cannabidiol 10 mg/kg/day group, and 6.9 mL (SD = 4.4 mL) in the pooled placebo group.

In the CARE 4 trial, the mean daily dose of the investigational products was 8.57 mL (SD = 4.50 mL) in the cannabidiol 20 mg/kg/day group and 8.59 mL (SD = 4.60 mL) in the volume-matched placebo group.

Treatment Adherence

Overall adherence to the investigational products in CARE3 was 89% in the cannabidiol 20 mg/kg/day group, 95% in the cannabidiol 10 mg/kg/day group and 95% in the pooled placebo group. In the CARE 3 trial, a missed dose (morning or evening) was reported more often in the cannabidiol 20 mg/kg/day group (17.1%) than in the cannabidiol 10 mg/kg/day group (9%) and the pooled placebo group (5.3%). Missed days (neither the morning nor evening dose taken) were reported more often in the cannabidiol 20 mg/kg/day group (6.1%) than in the cannabidiol 10 mg/kg/day group (3.0%) and the pooled placebo group (1.3%).

Overall adherence to the investigational products in CARE4 was 84% and 99% in the cannabidiol 20 mg/kg/day group and volume-matched placebo group, respectively. In the CARE 4 trial, a missed dose was reported more often in the cannabidiol 20 mg/kg/day group (15.1%) than in the volume-matched placebo group (9.4%), and missed days were also reported more often in the cannabidiol 20 mg/kg/day group (7.0%) than in the volume-matched placebo groups (2.4%).

Table 13: Summary of Patient Exposure From the CARE 3 and CARE 4 Trials (Safety Analysis Set)

SD = standard deviation.

^aDaily dose post-titration.

^bDaily doses are for the pooled placebo (20 mg/kg/day and 10 mg/kg/day) group.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

Concomitant Medications and Co-Interventions

Concomitant medications, therapies, and rescue medications used in the CARE 3 and CARE 4 trials are summarized in Table 14. Only concomitant medications, therapies, and rescue medications used by 10% of patients in any treatment group are presented. According to input from the clinical experts consulted by CADTH, zonisamide and felbamate are not available in Canada.

Concomitant Medications

In the CARE 3 trial, 224 patients (99.6%) reported taking 1 or more ASMs during the trial. The use of benzodiazepine derivatives was reported in 65% of patients, of which clobazam (48.9%) and clonazepam (12.9%) were the most commonly used (≥ 10% of patients). Carboxamide derivatives were used in approximately a third of all patients, with the most commonly used being rufinamide (28.9%). Fatty acid derivates were used in 44.1% of patients, with valproic acid reported as the most used derivative (37.8%). The use of other antiepileptics was reported in 83.1% of patients in the CARE 3 trial and included the use of levetiracetam (30.7%), lamotrigine (29.8%), zonisamide (14.2%), topiramate (15.1%), lacosamide (13.4%), and felbamate (12.4%).

In the CARE 4 trial, all 171 patients (100%) reported taking 1 or more ASMs during the trial. The use of benzodiazepine derivatives was reported in 63.2% of patients, of which clobazam (49.1%) and clonazepam (12.9%) were the most commonly used (≥ 10% of patients). Carboxamide derivatives were used in approximately a third of all patients, with the most commonly used being rufinamide (26.0%). Fatty acid derivates were used in 43.9% of patients, with valproic acid reported as the most used derivative (40.4%). The use of other antiepileptics was reported in 81.3% of patients in the CARE 4 trial and included the use of levetiracetam (33.9%), lamotrigine (37.4%), zonisamide (9.4%), topiramate (14.6%), lacosamide (12.9%), and felbamate (11.7%).

The overall use of concomitant ASMs was generally balanced between the CARE 3 and CARE 4 trials. In the CARE 3 trial, the use of benzodiazepine derivatives was higher in the cannabidiol 10 mg/kg/day group (77.6%) than in the cannabidiol 20 mg/kg/day group (59.8%) and the pooled placebo group (60.5%). In the CARE 4 trial, the use of phenobarbital was greater in the cannabidiol 20 mg/kg/day group than in the volume-matched placebo group (11.6% versus 2.4%), while the use of levetiracetam was greater in the volume-matched placebo group than in the cannabidiol 20 mg/kg/day group (40.0% versus 27.9%).

Concomitant Antiepileptic Therapies

Overall use of ketogenic diets or VNS in the CARE 3 trial was reported in 8.4% and 23.6% of patients, respectively. The proportion of patients using a ketogenic diet in the cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, and pooled placebo groups was 7.3%, 10.4%, and 7.9%, respectively. VNS use was reported in 22.0%, 20.9%, and 27.6% of patients in the cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, and pooled placebo groups, respectively.

In the CARE 4 trial, overall use of ketogenic diets and VNS was reported in 8.2% and 29.8% of patients, respectively. Use of a ketogenic diet or VNS was reported in 4.7% and 30.2% of patients in the cannabidiol 20 mg/kg/day group and in 11.8% and 29.4% of patients in the pooled placebo group, respectively.

Rescue Medication

In the CARE 3 trial, 120 patients (53.3%) used rescue medications during the trial. The most frequently used rescue medication was diazepam (35.6%). The proportion of patients requiring rescue medications in the cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, and pooled placebo groups was 56.1%, 49.3%, and 53.9%, respectively.

In the CARE 4 trial, 103 patients (60.2%) used rescue medications during the trial. The most frequently used rescue medication was diazepam (42.7%). The proportion of patients requiring rescue medications in the cannabidiol 20 mg/kg/day and volume-matched placebo groups was 61.6% and 58.8%, respectively.

Table 14: Summary of Concomitant Antiseizure Medications, Therapies, and Rescue Medications Used in the CARE 3 and CARE 4 Trials (Safety Analysis Set)

VNS = vagus nerve stimulation.

^aRescue medication used by > 5% of patients in any treatment group reported.

^bAntiseizure medications that are not available in Canada.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

Efficacy

Summary of key efficacy results from the CARE 3 and CARE 4 trials are presented in Table 15.

Percent Change From Baseline in Drop Seizure Frequency

At the end of the treatment period in the CARE 3 trial, a reduction in median percent change in drop seizure frequency was associated with treatment with cannabidiol 20 mg/kg/day (median difference = –21.6%; 95% CI, –34.8 to –6.7; P = 0.0047) and cannabidiol 10 mg/kg/day (median difference = –19.2%; 95% CI, –31.2 to –7.7; P = 0.0016) compared to placebo.

At the end of the treatment period in the CARE 4 trial, a reduction in the median percent change in drop seizure frequency was associated with treatment with cannabidiol 20 mg/kg/day compared to volume-matched placebo (median difference = –17.2%; 95% CI, –30.3 to –4.1; P = 0.0135).

Sensitivity Analysis

In both the CARE 3 and CARE 4 trials, sensitivity analyses using the per-protocol analysis set, and during the maintenance period using the intention-to-treat analysis set, were consistent with the primary analysis. The results of the sensitivity analyses that used imputed data for unreported days in the IVRS (using the worst case of the last observation carried forward, the next observation carried backward, and the mean daily number of seizures during the treatment period) and for MNAR (using a model with multiple imputations) were consistent with the primary analysis. The assumption of data MNAR did also not alter the results of the primary analysis.

Subgroup Analysis

The results of the subgroup analysis of percent change from baseline in drop seizure frequency are presented in Appendix 1. The treatment effect estimates generally favoured cannabidiol, compared with placebo, in the CARE 3 trial (Table 24) and the CARE 4 trial (Table 25).

Treatment Response

Reduction in Drop Seizures From Baseline of Greater Than or Equal to 50%

In the CARE 3 trial, during the treatment period, the difference in the proportion of patients who experienced at least a 50% reduction in drop seizure frequency from baseline between the cannabidiol 20 mg/kg/day group and the pooled placebo group was 25.0% (95% CI, 11.5% to 38.5%) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 21.1% (95% CI, 7.6% to 34.5%).

In the CARE 4 trial, the difference in the proportion of patients who experienced at least a 50% reduction in drop seizure frequency from baseline between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group was 20.7% (95% CI, 6.8% to 34.5%).

Reduction in Drop Seizures From Baseline of Greater Than or Equal to 25%

In the CARE 3 trial, during the treatment period, the difference in the proportion of patients who experienced at least a 25% reduction in drop seizure frequency from baseline between the cannabidiol 20 mg/kg/day group and the pooled placebo group was 18.4% (95% CI, 2.8% to 34.0%) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 19.6% (95% CI, 3.9% to 35.3%).

In the CARE 4 trial, the difference in the proportion of patients who experienced at least a 25% reduction in drop seizure frequency from baseline between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group was 20.4% (95% CI, 5.8% to 35.1%).

Reduction in Drop Seizures From Baseline of Greater Than or Equal to 75%

In the CARE 3 trial, during the treatment period, the difference in the proportion of patients who experienced at least a 75% reduction in drop seizure frequency from baseline between the cannabidiol 20 mg/kg/day group and the pooled placebo group was 22.4% (95% CI, 12.0% to 32.7%) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 8.3% (95% CI, 0.3% to 16.3%).

In the CARE 4 trial, the difference in the proportion of patients who experienced at least a 75% reduction in drop seizure frequency from baseline .during the treatment period between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group was 11.5% (95% CI, 1.3% to 21.8%).

Reduction in Drop Seizures From Baseline of 100%

No patient experienced a 100% reduction in drop seizure frequency during the treatment period in either the CARE 3 or CARE 4 trial.

Number of Inpatient Hospitalizations due to Epilepsy

In the CARE 3 trial, the number of patients who experienced 1 or more inpatient hospitalizations due to epilepsy was 7 (9.2%) in the cannabidiol 20 mg/kg/day group, 6 (8.2%) in the cannabidiol 10 mg/kg/day group, and 6 (7.9%) in the pooled placebo group.

In the CARE 4 trial, the number of patients who experienced 1 or more inpatient hospitalizations due to epilepsy was 10 (11.6%) and 5 (5.9%) in the cannabidiol 20 mg/kg/day and volumed-matched placebo groups, respectively.

Health-Related Quality of Life

QOLCE for Patients Aged 2 to 18 Years

In the CARE 3 trial, overall QOLCE scores were available for 33 patients (43.4%) in the cannabidiol 20 mg/kg/day group, 36 patients (49.3%) in the cannabidiol 10 mg/kg/day group, and in 38 patients (50%) in the pooled placebo group. At baseline, the overall mean QOLCE scores were comparable across the cannabidiol 20 mg/kg/day group, the cannabidiol 10 mg/kg/day group, and the pooled placebo group at 41.6 (SD = 16.5), 40.6 (SD = 15.4), and 41.4 (SD = 16.1), respectively. The adjusted mean treatment difference in change from baseline in overall QOLCE scores between the cannabidiol 20 mg/kg/day group and the pooled placebo group was –5.1 (95% CI, –11.4 to 1.2) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 1.6 (95% CI, –4.5 to 7.8).

In the CARE 4 trial, overall QOLCE scores were available for 26 (30.2%) and 38 (44.7%) patients in the cannabidiol 20 mg/kg/day group and the volumed-matched placebo group, respectively. At baseline, the overall mean QOLCE scores were comparable between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group at 39.5 (SD = 12.6) and 39.1 (SD = 15.2), respectively. The adjusted mean treatment difference in change from baseline in overall QOLCE scores between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group was 3.7 (95% CI, –3.3 to 10.7).

QOLIE-31-P for Patients Aged 19 Years and Older

In the CARE 3 trial, total QOLIE-31-P scores were available for 13 patients in the cannabidiol 20 mg/kg/day group, 14 patients in the cannabidiol 10 mg/kg/day group, and 10 patients in the pooled placebo group. At baseline, the total mean QOLIE-31-P scores were 50.2 (SD = 16.6) in the cannabidiol 20 mg/kg/day group, 56.0 (SD = 19.2) in the cannabidiol 10 mg/kg/day group, and 62.5 (SD = 13.6) in the pooled placebo group. The adjusted mean treatment difference in change from baseline in total QOLIE-31-P scores between the cannabidiol 20 mg/kg/day group and the pooled placebo group was 2.9 (95% CI, –7.3 to 13.1) and between the cannabidiol 10 mg/kg/day group and the pooled placebo group was 3.6 (95% CI, –7.0 to 14.3).

In the CARE 4 trial, total QOLIE-31-P scores were available for 14 patients (16.3%) in the cannabidiol 20 mg/kg/day group and 14 patients (16.5%) in the volume-matched placebo group. At baseline, the total QOLIE-31-P scores were comparable between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group at 55.8 (SD = 13.5) and 57.3 (SD = 19.5), respectively. The adjusted mean treatment difference in total QOLIE-31-P scores was not calculated due to the low number of patients who completed the assessment.

Drop Seizure–Free Days

In the CARE 3 trial, the mean number of drop seizure–free days at baseline was 5.3 (SD = 6.4) in the cannabidiol 20 mg/kg/day group, 5.3 (SD = 5.9) in the 10 mg/kg/day group, and 4.6 (SD = 5.5) in the pooled placebo group. The mean number of drop seizure–free days had increased in all treatment groups at the end of the treatment period. The change in the mean number of drop seizure–free days was 6.8 days (SD = 8.2) in the cannabidiol 20 mg/kg/day group, 5.5 days (SD = 6.7) in the cannabidiol 10 mg/kg/day group, and 2.3 days (SD = 5.1) in the pooled placebo group. Compared to placebo, treatment with cannabidiol 20 mg/kg/day was associated with an increase in the mean number of drop seizure–free days of 4.6 days (95% CI, 2.5 days to 6.8 days), while treatment with cannabidiol 10 mg/kg/day was associated with an increase of 3.3 days (95% CI, 1.2 days to 5.5 days).

In the CARE 4 trial, the mean number of drop seizure–free days at baseline was 6.4 (SD = 6.1) and 5.1 (SD = 5.5) in the cannabidiol 20 mg/kg/day group and the volume-matched placebo group, respectively. The mean number of drop seizure–free days had increased in both treatment groups at the end of the treatment period. The change in the mean number of drop seizure–free days was 5.8 days (SD = 7.4) in the cannabidiol 20 mg/kg/day group and 3.1 (SD, 5.4) days in the volume-matched placebo group. Compared to volume-matched placebo, treatment with cannabidiol 20 mg/kg/day was associated with a mean increase in drop seizure–free days of 2.7 days (95% CI, 0.7 days to 4.7 days).

Other Outcomes

The following outcomes were noted as being meaningful by the patient group and by the clinical experts consulted by CADTH but were not assessed in either the CARE 3 or CARE 4 trial: SUDEP rate, employment, and caregiver quality of life.

Table 15: Summary of Key Efficacy Results From the CARE 3 and CARE 4 Trials

CI = confidence interval; NA = not available; NE = not estimated; QOLCE = Quality of Life in Childhood Epilepsy; QOLIE-31-P = Quality of Life in Epilepsy, version 2; Q1 = lower quartile; Q3 = upper quartile; SD = standard deviation.

^aAdjusted for multiplicity using a hierarchical gate-keeping procedure.

^bThe P value from the Wilcoxon rank sum test is presented.

^cThe P value calculated from a Cochran-Mantel-Haenszel test is presented.

^dThis outcome was outside the adjustment for multiplicity.

^eThe 95% CI and the P value from the analysis of covariance model are presented.

^fThe change from baseline was presented for QOLIE-31-P because in the CARE 4 trial, the mean difference between groups was not assessed due to the small number of patients completing the questionnaire.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

Harms

A summary of the harms results in the CARE 3 and CARE 4 trials is presented in Table 16.

Adverse Events

In the CARE 3 trial, at least 1 AE was reported by 93.9% of patients in the cannabidiol 20 mg/kg/day group, 83.6% of patients in the cannabidiol 10 mg/kg/day group, and 72.4% of patients in the pooled placebo group. The most common AEs (reported in > 10% of patients) in the cannabidiol 20 mg/kg/day group were somnolence (30.5%), decreased appetite (25.6%), diarrhea (14.6%), upper respiratory tract infection (13.4%), pyrexia (12.2%), vomiting (12.2%), and nasopharyngitis (11.0%). In the cannabidiol 10 mg/kg/day group, the most common AEs reported were somnolence (20.9%), decreased appetite (16.4%), upper respiratory tract infection (16.4%), diarrhea (10.4%), and status epilepticus (10.4%). In the pooled placebo group, the most common AEs reported were pyrexia (15.8%), upper respiratory tract infection (14.5%), and vomiting (11.8%).

In the CARE 4 trial, at least 1 AE was reported by 86.0% and 69.4% of patients in the cannabidiol 20 mg/kg/day and the volume-matched placebo groups, respectively. Among patients randomized to receive cannabidiol 20 mg/kg/day, the most common AEs (reported in > 10% of patients) were diarrhea (18.6%), somnolence (15.1%), decreased appetite (12.8%), pyrexia (12.8%), and vomiting (10.5%). In the volume-matched placebo group, the most common AE reported was vomiting (16.5%).

Table 16: Summary of Harms Results From the CARE 3 and CARE 4 Trials

AE = adverse event; ALT = alanine aminotransferase; AST = aspartate aminotransferase; NA = not available; NR = not reported; SAE = serious adverse event.

^aOccurring in more than 10% of patients in any treatment group.

^bOccurring in at least 2% of patients in any treatment group.

Sources: Clinical Study Report for CARE 3¹⁶ and CARE 4 trials;¹⁷ sponsor’s summary of clinical evidence.

Serious AEs

In the CARE 3 trial, at least 1 SAE was reported by 15.9% of patients in the cannabidiol 20 mg/kg/day group, 19.4% of patients in the cannabidiol 10 mg/kg/day group, and 10.5% of patients in the pooled placebo group. The most common SAEs (reported in > 2% of patients) in the cannabidiol 20 mg/kg/day group were status epilepticus (4.9%), convulsion (2.4%), and pneumonia (2.4%). In the cannabidiol 10 mg/kg/day group, the most common SAEs were status epilepticus (10.4%) and pneumonia (4.5%). In the pooled placebo group, the most common SAE was status epilepticus (3.9%).

In the CARE 4 trial, at least 1 SAE was reported by 23.3% and 4.7% of patients in the cannabidiol 20 mg/kg/day and the volume-matched placebo groups, respectively. Among patients randomized to receive cannabidiol 20 mg/kg/day, the most common SAEs were increased AST (4.7%), increased ALT (4.7%), acute respiratory failure (3.5%), increased gamma-glutamyltransferase (3.5%), pneumonia (3.5%), viral infection (2.3%), convulsion (2.3%), and hypoxia (2.3%). In the volume-matched placebo group, the most common SAE was dehydration (2.4%).

Treatment Discontinuation Due to AEs

In the CARE 3 trial, discontinuation of treatment due to AEs was reported by 7.3% of patients in the cannabidiol 20 mg/kg/day group, 1.5% of patients in the cannabidiol 10 mg/kg/day group, and 1.3% of patients in the pooled placebo group. The most common reasons for treatment discontinuation due to AEs (reported in more than 2% of patients) were all reported in the cannabidiol 20 mg/kg/day group: increased levels of ALT (3.7%), AST (2.4%), and gamma-glutamyltransferase (2.4%).

In the CARE 4 trial, discontinuation of treatment due to AEs was reported in 14.0% and 1.2% of patients in the cannabidiol 20 mg/kg/day group and the volume-matched placebo group, respectively. The most common reasons for treatment discontinuation due to AEs (reported in more than 2% of patients) were all reported in the cannabidiol 20 mg/kg/day group: increased levels of ALT (3.5%), AST (3.5%), and gamma-glutamyltransferase (2.3%).

Mortality

There were no reported deaths in the CARE 3 trial. In the CARE 4 trial, 1 death was recorded by investigators (1.2% of patients) due to acute respiratory distress syndrome in the cannabidiol 20 mg/kg/day group.

Notable Harms

Notable harms of interest were related to nervous system disorders (i.e., somnolence, status epilepticus, and sedation), hepatocellular injury or investigation (i.e., increased levels of ALT and AST, bilirubin elevation, and serum transaminase elevation), and gastrointestinal disorders (i.e., diarrhea, vomiting, and constipation).

In the CARE 3 trial, somnolence, status epilepticus, and sedation were reported in 30.5%, 4.9%, and 3.7% of patients, respectively, in the cannabidiol 20 mg/kg/day group; in 20.9%, 10.4%, and 3.0% of patients, respectively, in the cannabidiol 10 mg/kg/day group; and in 5.3%, 3.9%, and 1.3% of patients, respectively, in the pooled placebo group. Increased levels of ALT, AST, and serum transaminase were reported in 4.9%, 3.7%, and 1.2% of patients, respectively, in the cannabidiol 20 mg/kg/day group; in 4.5%, 3.7%, and 1.5% of patients, respectively, in the cannabidiol 10 mg/kg/day group; and in 1.3%, 1.3%, and 0% of patients, respectively, in the pooled placebo group. The following investigations were not reported in the CARE 3 trial: abnormal liver function test, acute hepatic failure, bilirubin elevation, and hepatotoxicity. Diarrhea, vomiting, and constipation were reported in 14.6%, 12.2%, and 4.9% of patients, respectively, in the cannabidiol 20 mg/kg/day group; in 10.4%, 6.0%, and 4.5% of patients, respectively, in the cannabidiol 10 mg/kg/day group; and in 7.9%, 11.8%, and 3.9% of patients, respectively, in the pooled placebo group.

In the CARE 4 trial, somnolence, sedation, and status epilepticus were reported in 15.1%, 8.1%, and 1.2% of patients, respectively, in the cannabidiol 20 mg/kg/day group and in 9.4%, 1.2%, and 1.2% of patients, respectively, in the volume-matched placebo group. The following hepatocellular injury and investigation AEs were reported in the cannabidiol 20 mg/kg/day group: increased ALT (9.3%), increased AST (7.0%), abnormal liver function test (4.7%), acute hepatic failure (3.5%), serum transaminase elevation (2.3%), hepatic failure (1.2%), and hepatotoxicity (1.2%). In the volume-matched placebo group, increased levels of ALT and AST were reported in 2.4% and 1.2% of patients, respectively. Diarrhea, vomiting, and constipation were reported in 18.6%, 10.5%, and 7.0% of patients, respectively, in the cannabidiol 20 mg/kg/day group and in 8.2%, 16.5%, and 4.7% of patients in the volume-matched placebo group.

Critical Appraisal

Internal Validity

Both the CARE 3 and CARE 4 trials were multicentre, randomized, double-blind phase III trials, designed in accordance with the European Medicines Agency Guideline on Clinical Investigation of Medicine Products in the Treatment of Epileptic Disorders and the FDA Guidelines for the Clinical Evaluation of Antiepileptic Drugs (Adults and Children). The treatment period (including the titration period), 10-day taper period, and 28-day follow-up period employed in the trial were chosen in consultation with an external advisory board and the FDA Center for Drug Evaluation and Research.^76,77 Patients were randomized centrally using IVRS technology, which is typically adequate for concealing allocation until treatment assignment. IVRS technology was also used to dispense the investigational product, allowing the treatment concealment for both patients and the investigators.

Although the CARE 3 trial included 4 treatment types (cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, volume-matched placebo 20 mg/kg/day, and volume-matched placebo 10 mg/kg/day), the study participants were randomized using a 1:1:1 randomization ratio to the cannabidiol 20 mg/kg/day treatment group, the cannabidiol 10 mg/kg/day treatment group, and the placebo treatment group. Patients in the placebo group were split in half to receive either the 20 mg/kg/day placebo or the 10 mg/kg/day placebo, and the study results were reported based on the pooled placebo group. While this approach is acceptable, it relies on the assumption that randomization was successful in each group. Differences in baseline characteristics between patients in the cannabidiol 20 mg/kg/day, cannabidiol 10 mg/kg/day, and pooled placebo groups were noted in the following: proportion of patients reporting convulsive seizures longer than 30 minutes and reporting concomitant use of benzodiazepine derivatives at baseline. According to the clinical experts consulted by CADTH for the purpose of this review, benzodiazepine derivates, such as clobazam, could strengthen the treatment effect of cannabidiol due to synergism; they did not anticipate any significant effect modification through the use of the other concomitant ASMs. The clinical experts, however, were uncertain if the other imbalances could influence treatment response.

At CADTH’s request, the sponsor reported that the assumptions related to splitting the placebo group and pooling results for analyses were not formally tested. However, the sponsor noted, as described in the Clinical Study Report for the CARE 3 trial, that posthoc sensitivity analyses were conducted on the primary outcome to determine if pooling the placebo group had an effect on the results. The analyses indicated that the efficacy response when cannabidiol 20 mg/kg/day and 10 mg/kg/day dosages were compared to their corresponding placebo dosages (i.e., 20 mg/kg/day and 10 mg/kg/day, respectively) were consistent with the results of similar comparisons in the pooled analyses. In the CARE 4 trial, the 1:1 randomization ratio and the randomization stratification factors appeared appropriate, and no notable baseline imbalances were observed between the cannabidiol 20 mg/kg/day group and the volume-matched placebo group.

In the CARE 3 trial, 23.2% of patients screened for entry into the trial did not meet the screening criteria; the reason they did not meet the screening criteria was listed as “other reason” for approximately a third of these patients. Moreover, a higher proportion of patients in the cannabidiol 20 mg/kg/day group discontinued the study than in the cannabidiol 10 mg/kg/day and pooled placebo groups (approximately 12% versus approximately 3%). The higher discontinuation rate in the cannabidiol 20 mg/kg/day group appeared to be driven by AEs. In addition, 6 patients randomized to the cannabidiol 10 mg/kg/day group and 3 patients randomized to receive volume-matched placebo at 10 mg/kg/day were accidentally titrated to 20 mg/kg/day. This error was detected and, by day 42, those patients were returned to the 10 mg/kg/day dosage. For the purpose of the safety analysis, these patients were moved to their respective 20 mg/kg/day dosage groups. In the CARE 4 trial, a higher proportion of patients in the cannabidiol 20 mg/kg/day group discontinued the study than in the volume-matched placebo group (9.3% versus 1.2%). The higher discontinuation rate in the cannabidiol 20 mg/kg/day group appeared to be driven by AEs. The application of the MNAR assumption and sensitivity analysis exploring missing efficacy results due to treatment discontinuation suggest that bias due to uneven discontinuation was unlikely.

In both the CARE 3 and CARE 4 trials, all efficacy outcome questionnaires and assessments were to be completed by the caregiver. To maintain consistency, the same caregiver, if the patients had multiple caregivers, was to complete and answer the questionnaire and assessment. Seizure information in both studies was ascertained using an IVRS diary, while paper diaries were used to capture use of the investigational product, use of concomitant medications, and occurrence of AEs. A review of clinical outcomes used in LGS-focused research noted that while diaries enable day-to-day reporting, without a standardized questionnaire that can serve a prompt for the diary, the daily entry may be biased.⁷⁸ According to input from the clinical experts, patient and parent or caregiver reports of seizures tend to be accurate for motor seizures but are not very reliable or accurate for nonmotor seizures. Indeed, studies examining the accuracy of clinical reports of seizures found that while absence seizures were least accurately identified (14%) compared to myoclonic seizures (48%) and complex or simple partial seizures (63%), all atonic and generalized tonic seizures were correctly identified by parents.⁷⁹ Accordingly, it may be assumed that the frequencies of drop seizures reported by patients and/or caregivers in the CARE 3 and CARE 4 trials are reliable estimates of the true frequency experienced by patients. Seizure diaries are the standard method of collecting data for clinical trials, and the International League Against Epilepsy recommends the use of dairies for collecting seizure frequency data.¹⁸

Both the CARE 3 and CARE 4 trials assessed HRQoL — outcomes deemed important by the patients and clinicians — using validated and reliable instruments: the QOLCE and the QOLIE-31-P questionnaires. The double-blind nature of the trials minimized the risk of bias in the measurement of subjective items on the QOLCE and QOLIE-31-P questionnaires. However, comparative efficacy conclusions based on the HRQoL outcomes are limited since the QOLCE and the QOLIE-31-P questionnaires were not part of the hierarchical testing procedure and because of the low completion rate across the treatment groups. Total QOLCE scores were available for 47.6% and 36.3% of patients in the CARE 3 and CARE 4 trials, respectively, while total QOLIE-31-P scores were available for 16.5% and 16.4% of patients in the CARE 3 and CARE 4 trials, respectively. Consequently, the assessment of HRQoL in both trials is at high risk of attrition bias, although the extent and direction of the bias cannot be determined since it is not clear if the patients who completed the questionnaires were systematically different from those who did not. As the completion rates were similar between the treatment groups within the CARE 3 and CARE 4 trials, there is little risk that attrition bias favoured any 1 treatment group.

The analysis of efficacy results in the CARE 3 and CARE 4 trials followed a defined statistical analysis plan. The primary and key secondary end points were addressed using a hierarchical gate-keeping procedure that controlled for type I errors. The sponsor conducted additional sensitivity analysis of the primary efficacy end point using the per-protocol analysis set and testing the MNAR assumption. In all scenarios, the sensitivity analyses were consistent with the primary efficacy analysis. Subgroup analyses were performed in both the CARE 3 and CARE 4 trials for the primary efficacy end point to examine the consistency of the treatment effect observed. Although the subgroups were prespecified, strong conclusions could not be drawn due to the small sample size and the absence of statistical testing for interaction effects.

External Validity

Due to the rarity of the LGS and a lack of population-based data on LGS, the clinical experts consulted by CADTH for the purpose of this review were unable to assess if the results of the CARE 3 and CARE 4 trials were applicable to the patients seen in the Canadian clinical setting. One clinical expert added that they do not, however, expect patients with LGS living in Canada to differ from such patients elsewhere. The clinical experts did note trial details that were applicable to the Canadian clinical setting, and others that were not representative of Canadian clinical practice. The generalizability of the evidence, including the clinical expert input, is summarized in Table 17.

Table 17: Assessment of Generalizability of Evidence for Cannabidiol

ASM = antiseizure medication; HRQoL = health-related quality of life; LGS = Lennox-Gastaut syndrome; QOLCE = Quality of Life in Childhood Epilepsy; QOLIE-31-P = Quality of Life in Epilepsy, version 2; VNS = vagus nerve stimulation.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For the pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for the outcomes considered most relevant to inform CADTH’s expert committee deliberations, and a final certainty rating was determined as outlined by the GRADE Working Group:^19,20

• High certainty: We are very confident that the true effect lies close to that of the estimate of the effect.

• Moderate certainty: We are moderately confident in the effect estimate. The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. We use the word “likely” for evidence of moderate certainty (e.g., “X intervention likely results in Y outcome”).

• Low certainty: Our confidence in the effect estimate is limited. The true effect may be substantially different from the estimate of the effect. We use the word “may” for evidence of low certainty (e.g., “X intervention may result in Y outcome”).

• Very low certainty: We have very little confidence in the effect estimate. The true effect is likely to be substantially different from the estimate of the effect. We describe evidence of very low certainty as “very uncertain.”

Following the GRADE approach, evidence from RCTs started as high-certainty evidence and could be rated down for concerns related to study limitations (which refers to internal validity or risk of bias), inconsistency across studies, indirectness, imprecision of effects, and/or publication bias.

When possible, certainty was rated in the context of the presence of an important (nontrivial) treatment effect; if this was not possible, certainty was rated in the context of the presence of any treatment effect (i.e., the clinical importance is unclear). In all cases, the target of the certainty of evidence assessment was based on the point estimate and where it was located relative to the threshold for a clinically important effect (when such a threshold was available) or to the null. The presence or absence of an important effect were based on thresholds informed by the clinical experts consulted for this review.

For the GRADE assessments, findings from the CARE 3 and CARE 4 trials were considered together and summarized narratively per outcome because these studies were similar in population, interventions, design, and outcome measures.²¹

Results of GRADE Assessments

Table 2 presents the GRADE summary of findings for cannabidiol versus placebo.

Long-Term Extension Studies

Contents within this section have been informed by materials submitted by the sponsor. The following has been summarized and validated by the CADTH review team.

To inform the long-term safety and tolerability of cannabidiol as an adjunctive treatment in children and adults with inadequately controlled LGS, the results of 1 sponsor-conducted OLE study — the CARE 5 trial — were summarized.²²

Description of Studies

The CARE 5 trial was a multicentre, single-arm, OLE, phase III study with the primary objective of evaluating the long-term safety and tolerability of cannabidiol as adjunctive treatment in patients aged 2 years and older with LGS (N = 366) who had completed the CARE 3 and CARE 4 trials (the “core studies”). The long-term efficacy of cannabidiol as adjunctive treatment in patients aged 2 years and older with LGS was the secondary objective of the CARE 5 trial.

The CARE 5 trial was conducted across 75 sites in 8 countries (Australia, Spain, France, Israel, Netherlands, Poland, UK, and US); approximately 78% of patients were living in the US. A schematic of the CARE 5 study design is presented in Figure 3. The CARE 5 trial consisted of a titration period and a maintenance period, followed by a 10-day taper period. Between visits, safety telephone calls were made every 4 weeks starting at visit 2, with additional safety telephone calls made every 7 days for patients who were titrated to dosages above 20 mg/kg/day, until stable dosing was achieved. All safety telephone calls assessed AEs, epilepsy-related hospitalization, concomitant medications, and/or changes to medications. The end-of-treatment visit occurred either after a maximum duration of 6 years (312 weeks after visit 1), depending on the country of residence; following early withdrawal from the study; or following an unscheduled end-of-treatment visit conducted no earlier than 730 days after visit 1. A safety telephone call was made 2 weeks after the end-of-taper visit, or the date of final dosing, to administer the Cannabis Withdrawal Scale or the Pediatric Cannabinoid Withdrawal Scale surveys.

Populations

Patients eligible for inclusion in the CARE 5 trial were required to have completed all scheduled visits in the treatment phase of the CARE 3 and CARE 4 trials. Patients were excluded from entering the CARE 5 trial if they met any of the following exclusion criteria:

• currently using or has used recreational or medicinal cannabis or synthetic cannabinoid-based medications within the 3 months before study entry, not including cannabidiol received during the core study

• unwilling to abstain from using recreational or medicinal cannabis or synthetic cannabinoid-based medications during the study

• a history of symptoms (e.g., dizziness, light-headedness, blurred vision, palpitations, weakness, syncope) related to a drop in blood pressure due to postural changes

• any history of suicidal behaviour or any suicidal ideation of type 4 or 5 on the Columbia–Suicide Severity Rating Scale at visit 1

• significantly impaired hepatic function at the end-of-treatment visit of the core study or at visit 1 of the CARE 5 trial if reassessed

• is of childbearing potential or has a partner of childbearing potential, unless willing to ensure that they or their partner use highly effective contraception for the duration of the study and for 3 months thereafter.

Figure 3: CARE 5 Study Design

Notes: All participants were instructed to begin titration of the OLE IMP on the evening of visit 1 (day 1). Participants with no gap in IMP dosing between the core study and the OLE (including tapered dosing) took their final dose of core study IMP on the morning of visit 1. The maximum duration of the OLE study at the Poland site was 6 years (312 weeks after visit 1).

^a Between visits, safety telephone calls were made every 4 weeks (± 7 days). In addition, during titration of dosages above 20 mg/kg/day, safety telephone calls were made every 7 days (± 3 days) until stable dosing was achieved. All safety telephone calls assessed adverse events, epilepsy-related hospitalizations, concomitant medications, and/or changes to medication.

^b The end-of-treatment visit occurred after a maximum of 6 years’ treatment (312 weeks after visit 1), following early withdrawal from the study, or following an unscheduled end-of-treatment visit conducted no earlier than 730 days after visit 1.

^c A safety telephone call was made 2 weeks (± 3 days) after the end-of-taper visit (or the date of final dosing) to collect seizure information, to administer the Cannabis Withdrawal Scale or Pediatric Cannabinoid Withdrawal Scale surveys, and to assess adverse events, epilepsy-related hospitalization, concomitant medications, and/or changes to medication.

^d Could be conducted by telephone.

^e IMP dispensing appointments took place between all visits after visit 7, 14 weeks after the previous visit.

Source: Sponsor’s summary of clinical evidence.⁴²

Interventions

All patients enrolled in the CARE 5 trial received adjunctive treatment with cannabidiol, which consisted of a 2-week titration period, a maintenance period, and a 10-day taper period. Patients were instructed to begin titration of the investigational product on the evening of visit 1. Patients with no gaps in investigational dosing between the core studies and the OLE, including tapered dosing, took their final dose of core study investigational product on the morning of OLE visit 1. Patients were instructed to take cannabidiol twice daily (morning and evening) immediately after their usual ASM administration, or as otherwise specified by the investigator. Patients were titrated up to 10 mg/kg/day to 20 mg/kg/day cannabidiol using a titration schedule. In the event of intolerance during titration, the patient was maintained on a dosage below 10 mg/kg/day or 20 mg/kg/day. A titration rate faster than recommended was considered if there was an increase in seizures, following consultation with the medical monitor. The patients continued at their 10 mg/kg/day to 20 mg/kg/day dosage during the maintenance period. During the maintenance period, dosing adjustments by the investigators were permitted should a patient experience intolerance (dose decrease) or require better seizure control (dose increase) until the optimal dose was achieved. If deemed necessary by the investigator, a maximum dosage of 30 mg/kg/day was permitted. Dosage increases above 20 mg/kg/day were recommended to be done slowly, with maximum increments of 2.5 mg/kg every 5 to 7 days. Among patients whose dosage had been decreased, dosage increases were still considered, provided there was adequate tolerance. Following the end-of-treatment or withdrawal visit, doses were tapered at home (10% per day for 10 days) until the end-of-taper visit. Patients could receive treatment for up to a maximum of 6 years (312 weeks after visit 1), depending on the country of enrolment.

Outcomes

Primary End Point

The primary end point in the CARE 5 trial was the frequency of AEs.

Secondary End Points

The following seizure information was evaluated as secondary efficacy end points: drop seizure frequency, nondrop seizure frequency, number of patients who were drop seizure free, and response rate (defined in terms of a ≥ 25%, ≥ 50%, ≥ 75%, or 100% reduction of in drop seizure frequency).

The secondary efficacy end points evaluated in the CARE 5 trial that were also relevant to this review were:

• HRQoL (as measured by the QOLCE and QOLIE-31-P scores)

• need for hospitalization due to epilepsy.

Statistical Analysis

No formal sample size calculation was performed.

Safety and efficacy outcomes were summarized using descriptive statistical methods. Changes from baseline were presented where baseline data were available from the core studies. Efficacy end points were analyzed in the safety analysis set. Missing data were addressed using the last observation carried forward method.

There was no formal hypothesis testing.

Results

Patient Disposition

Patient disposition for the CARE 5 OLE study is summarized in Table 18.

Of the 681 patients screened and enrolled in the study, 366 (100%) were patients with LGS. Of these 366 patients, one-third (33.6%) discontinued the study, mainly due to withdrawal by patient or parent or guardian (13.1%), AEs (10.4%), and “other reasons” (4.9%). Most of the patients with LGS (79.5%) did not continue to the taper phase, which lasted up to 10 days. However, the majority of patients with LGS who did continue to the taper phase (20.5%) completed the taper phase (18.3%). All enrolled patients received at least 1 dose of cannabidiol and were included in the safety analysis set.

The retention rates for the safety analysis set at weeks 37 to 48 (12 months), weeks 85 to 96, weeks 133 to 144 (36 months), weeks 181 to 192, and weeks 241 to 252 were 82% (299 of 366 patients), 64% (236 of 366 patients), 59% (216 of 366 patients), 6% (22 of 366 patients), and 2% (8 of 366 patients), respectively.

Table 18: Patient Disposition in CARE 5 Open-Label Extension Study

AE = adverse event; LGS = Lennox-Gastaut syndrome.

Source: Clinical Study Report for CARE 5 trial.²²

Baseline Characteristics

The baseline demographic and clinical characteristics of the CARE 5 OLE study for patients with LGS are summarized in Table 19.

The mean age of the patients with LGS was 15.9 years (SD = 9.5): 36 patients (9.8%) were aged 2 to 5 years, 121 patients (33.1%) were aged 6 to 11 years, 89 patients (24.3%) were aged 12 to 17 years, and 120 patients (32.8%) were aged 18 years or older. The proportion of male patients was 54.1%. The majority of patients (88.8%) were white. The median number of currently used ASMs was 3 (range, 1 to 13).

Table 19: Summary of Baseline Characteristics of Patients With LGS (Safety Analysis Set) in CARE 5 OLE Study