CADTH Reimbursement Review

Cannabidiol (Epidiolex)

Sponsor: Jazz Pharmaceuticals Canada Inc.

Therapeutic area: Dravet 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

NOC = Notice of Compliance.

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

Introduction

Dravet syndrome (DS) is a very rare form of epilepsy associated with treatment-resistant, lifelong seizures and substantial comorbidities such as intellectual disability and behavioural, sleep, and gait problems. Epilepsy onset in DS usually occurs within the first year of life with febrile or afebrile clonic and tonic-clonic, generalized, and unilateral seizures in infants who have previously developed as expected.¹ Approximately 70% to 85% of cases with clinical features of DS have mutations of the SCN1A gene.^2-4 The most common causes of death in DS are sudden unexpected death in epilepsy (SUDEP) and status epilepticus. In 2011, the estimated incidence of DS was 1 in 33,000 live births worldwide, with a prevalence estimated at 1 in 45,700 children younger than 18 years of age.⁵ Epidemiological data specific to the Canadian landscape are scarce; however, Dravet Canada estimates that 1,000 individuals in Canada have DS, which accounts for 1% of the general epilepsy population.⁶ According to Orphanet, the estimated prevalence of DS in Canada is 1 in 40,000.⁷

The diagnosis of DS is based primarily on clinical observations. Confirmatory genetic testing for an SCN1A variant can be necessary when there is clinical uncertainty in the diagnosis. Treatment includes valproic acid and clobazam initially, but these are usually insufficient to control seizures. In patients with DS that is refractory to initial therapies, clinicians may add other antiseizure medications (ASMs), including stiripentol, topiramate, levetiracetam, clonazepam, and rufinamide. Cannabidiol is also recommended as an adjunctive treatment option for patients whose condition fails to respond to first-line ASMs.

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 (Epidiolex) 100 mg/mL oral solution in the treatment of patients 2 years of age and older with seizures associated with DS.

Stakeholder Perspectives

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

Patient Input

CADTH received 1 patient input, which was from the Canadian Epilepsy Alliance (CEA). CEA is dedicated to the promotion of independence and quality of life for people with epilepsy and their families by providing support services, information, and advocacy and increasing public awareness. Information for this submission was gathered by the president of the CEA through consultation with 24 member associations.

The CEA highlighted that individuals with uncontrolled epilepsy are at risk of social isolation and mental illness. The unpredictable nature of seizures and side effects of medications, such as anxiety, depression, mood swings, sexual dysfunction, suicidal thoughts, and exhaustion, have negative effects on patients and their family and caregivers. Currently available treatments do not control seizures in all patients. Lack of access to an approved treatment among patients with uncontrolled seizures can result in trying alternative practices and treatments such as cannabis and unregulated substances. The CEA input mentioned that any reduction in the frequency of seizures can improve quality of life among patients. Because of the frequent seizures, patients with epilepsy syndromes are often unemployed or underemployed with restricted income and without access to employer-funded insurance plans, which limit their access to drugs that are not on the provincial formulary.

Clinician Input

Input From the Clinical Experts Consulted by CADTH

Two clinical specialists with expertise in the diagnosis and management of DS provided input to this submission. Both agreed that the treatment goals of any therapy for patients with DS include improving seizure control with the improvement of health-related quality of life (HRQoL), and decreasing seizure burden without affecting the mood, cognition, or behaviour of patients. Other goals include increasing the number of seizure-free days and decreasing visits to health care facilities and the need for rescue medications. The clinical experts mentioned that cannabidiol has the potential for fewer adverse effects when compared with other drugs indicated for this condition. Initially, it is anticipated that cannabidiol would be used after valproic acid and clobazam. The experts mentioned that cannabidiol may be useful in the treatment paradigm in adult patients, as they do not seem to tolerate stiripentol as well as children do; in both populations, the need exists for drugs with fewer side effects and greater benefits.

According to the clinical experts, the frequency and change over time in seizure frequency, number of seizure-free days, decrease in seizure duration and severity, reduction of status epilepticus, and decreased use of rescue medication are important end points when assessing response to treatment. The experts mentioned they would consider an inadequate improvement in seizure frequency (approximately less than a 50% decrease from baseline) and the presence of intolerable adverse events (AEs) as factors to determine the discontinuation of the medication.

Most patients taking cannabidiol will be treated in outpatient epilepsy clinics. The clinical experts suggested that epileptologists and/or neurologists with expertise in the treatment of DS should be the health care providers monitoring response in these patients.

Clinician Group Input

No clinician group input was received by CADTH.

Drug Program Input

The drug program input highlights that the diagnosis of DS is largely clinical, and genetic testing for variants (i.e., of the SCN1A gene) alone is not sufficient for diagnosis. The input also mentioned that the reimbursement criteria for stiripentol, the drug currently approved for DS in Canada, only include a diagnosis of DS (without specific criteria around a diagnosis). Cannabidiol would require similar criteria, if appropriate.

The Health Canada indication and reimbursement criteria for stiripentol include combination treatment with both valproate and clobazam. The drug programs asked if it would be appropriate to require patients to be on both valproate and clobazam before being eligible for reimbursement for cannabidiol (i.e., similar to the indication for stiripentol and aligned with current clinical guidelines). The clinical experts consider that it is possible, based on the current available evidence.

The drug programs also inquired about the objective measures used to assess and monitor therapeutic response in clinical practice. The experts mentioned that reduction in convulsive seizures, use of rescue medication, number of hospital and emergency department visits, and number of AEs are appropriate measures, and added that an inadequate reduction in total seizures (i.e., a reduction of less than 50%) would be an adequate measure to assess response.

The drug programs wanted to know how many patients the clinical experts think will use the dose of 20 mg/kg/day. The experts approximated that 30% of patients would require this higher dose.

The drug programs noted there may be limited access to neurologists within some regions of Canada. The stiripentol reimbursement criteria in most jurisdictions indicate that the drug “must be prescribed by or in consultation with,” or the patient “must be under the care of” a neurologist or general pediatrician; hence, the drug programs suggested considering alignment with the prescribing criteria for stiripentol where these practitioners can provide the needed care in these situations.

Also, the drug programs pointed out that patients currently using medicinal cannabis or synthetic cannabinoid-based medications and transitioning to pharmaceutical-grade cannabidiol were excluded from the CARE1 and CARE2 trials and need to be considered in the deliberations.

Finally, the drug programs recommended that due to the risk of hepatocellular injury, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and total bilirubin levels should be obtained at baseline and then at 1, 3, and 6 months after starting treatment — and periodically thereafter, as clinically indicated — or within 1 month of change in cannabidiol dosing or upon a change in other medications that affect liver function.

Clinical Evidence

Systematic Review

Description of Studies

The body of evidence informing this submission consists of 2 individual studies assessing cannabidiol in patients with DS.

First, the pivotal CARE1 Part B study (N = 120 patients) was a phase III, double-blind, placebo-controlled, multicentre, randomized trial evaluating cannabidiol 20 mg/kg/day (n = 61) against placebo (n = 59) as an adjunctive therapy in patients aged 2 to 18 years whose DS is not completely controlled with current ASMs. The study evaluated seizure frequency per month, proportion of patients with a 50% or greater reduction in convulsive-seizure frequency, number of seizure-free days, presence of status epilepticus, HRQoL scores, amount of sleep disruption, and harms. The time of treatment and assessment was 14 weeks.

Second, the pivotal CARE2 was a 3-arm study that evaluated cannabidiol 20 mg/kg/day (n = 67) and 10 mg/kg/day (n = 67) against a placebo group (n = 65). All patients in this study were also 2 to 18 years of age and were receiving multiple therapies for controlling their seizures. The study also evaluated seizure frequency per month, proportion of patients with a 50% or greater reduction in convulsive-seizure frequency, number of seizure-free days, presence of status epilepticus, HRQoL scores, amount of sleep disruption, and harms. The time of treatment and assessment was 14 weeks.

Efficacy Results

Percentage Change From Baseline in the Frequency of Convulsive Seizures During the Treatment Period

In the CARE1 Part B study, patients in the 20 mg/kg/day cannabidiol group achieved a median percentage change from baseline in convulsive-seizure frequency during the 14-week treatment period of –38.9% (95% confidence interval [CI], –69.5% to –4.8%) versus –13.3% (95% CI, –52.5% to 20.2%) for the placebo group. The estimated median difference between treatment arms was –22.8% (95% CI, –41.1% to –5.4%; P = 0.0123).

In the CARE2 study, the median percentage change from baseline during treatment was –41.2% (95% CI, –81% to 3.0%), –47.0% (95% CI, –71.4% to –10.5%), and –24.5% (95% CI, –51.9% to 4.6%) in the cannabidiol 10 mg/kg/day, cannabidiol 20 mg/kg/day, and placebo groups, respectively. The estimated median difference for cannabidiol 10 mg/kg/day versus placebo was –15.7% (95% CI, –31.3% to 3.7%; P = 0105) and –19.9% (95% CI, –33.9% to 5.3%; P = 0.008) for cannabidiol 20 mg/kg/day versus placebo.

Proportion of Patients With a 50% or Greater Reduction From Baseline in the Frequency of Convulsive Seizures During the Treatment Period

In the CARE1 Part B study, the proportion of patients with a reduction of 50% or more in their baseline convulsive-seizure frequency was greater in the cannabidiol group, with 26 of 61 patients (42.6%) experiencing this level of reduction versus 16 of 59 patients (27.1%) in the placebo group. The difference in proportions was 0.155 (95% CI, –0.013 to 0.323) in favour of the intervention. The odds of achieving this end point in the cannabidiol group were double the odds in the placebo group (odds ratio [OR] = 2.00; 95% CI, 0.93 to 4.30; P = 0.0784).

In the CARE2 study, the proportion of patients with a reduction of 50% or more from baseline in the frequency of convulsive seizures was greater in the 10 mg/kg/day group (29 of 66 patients; 43.9%) and in the 20 mg/kg/day group (33 of 67 patients; 49.3%) compared with placebo (17 of 65 patients; 26.2%). The difference in proportion versus placebo was 0.178 (95% CI, 0.017 to 0.338) in the 10 mg/kg/day group and 0.231 (95% CI, 0.071 to 0.391) in the 20 mg/kg/day group. The odds of achieving this end point were higher in both the 10 mg/kg/day group (OR = 2.21; 95% CI, 1.06 to 4.62; P = 0.0332) and the 20 mg/kg/day group (OR = 2.74; 95% CI, 1.32 to 5.70; P = 0.0069) compared with placebo.

Proportion of Patients With a 75% or Greater Reduction From Baseline in the Frequency of Convulsive Seizures During the Treatment Period

In the CARE1 Part B study, the proportion of patients with a reduction of 75% or more in their baseline convulsive-seizure frequency was greater in the 20 mg/kg/day cannabidiol group compared with placebo, with 14 of 61 patients (23%) and 7 of 59 patients (11.9%), respectively, experiencing this level of reduction. The difference in proportions was 0.111 (95% CI, –0.023 to 0.245) in favour of the intervention. The odds of achieving a 75% or greater reduction was 2.21 (95% CI, 0.82 to 5.95; P = 0.1121) in favour of the 20 mg/kg/day group.

In the CARE2 study, 12 of 67 patients (17.9%) in the 20 mg/kg/day cannabidiol group and 20 of 66 patients (30.3%) in the 10 mg/kg/day cannabidiol group achieved a 75% or greater reduction in convulsive-seizure frequency compared with 4 of 65 patients (6.2%) in the placebo group. The difference in proportion between the 10 mg/kg/day group and placebo was 0.241 (95% CI, 0.116 to 0.367) and 0.118 (95% CI, 0.009 to 0.226) in the 20 mg/kg/day group. Compared with placebo, the odds of achieving a 75% or greater reduction was 6.63 (95% CI, 2.12 to 20.73; P = 0.0004) in the 10 mg/kg/day group and 3.33 (95% CI, 1.01 to 10.92; P = 0.0468) in the 20 mg/kg/day group.

Number of Convulsive Seizure–Free Days

In the CARE2 study, the mean number of convulsive seizure–free days increased in all 3 treatment groups, although greater increases were seen in the 10 mg/kg/day and 20 mg/kg/day cannabidiol groups compared with placebo. The treatment difference was in favour of both cannabidiol doses, with a treatment difference of 2.4 (95% CI, 1.0 to 3.9; P = 0.0009) between the 10 mg/kg/day group and placebo, and 1.3 (95% CI, –0.1 to 2.8; P = 0.0683) between the 20 mg/kg/day group and placebo.

Percentage Change From Baseline in Total Seizure Frequency During the Treatment Period

In the CARE1 Part B study, a greater median percentage change in total seizure frequency was seen in the 20 mg/kg/day cannabidiol group (median difference of –28.6%; 95% CI, –70.4 to –4.0) compared with the placebo group (median difference of –9.0%; 95% CI, –51.4 to 19.6). The median difference between 20 mg/kg/day cannabidiol and placebo was –19.2 (95% CI, –39.3 to –1.2; P = 0.0335).

In the CARE2 study, the percentage reduction was 56.4 (95% CI, 47.8 to 63.6) in the 10 mg/kg/day and 47.3 (95% CI, 36.9 to 56.0) in the 20 mg/kg/day cannabidiol groups compared with 29.7 (95% CI, 16.0 to 41.1) in the placebo group.

Patients With Status Epilepticus

In both studies, there were few incidents of status epilepticus reported overall during the baseline and treatment periods, with similar rates across all treatment groups. In the CARE1 Part B study, there was only 1 case (1.6%) in the 20 mg/kg/day group versus 0 in the placebo group at the end of the treatment period. Similarly, patients in the CARE2 study presented with status epilepticus in numbers of 3 (4.5%), 9 (13.4%), and 8 (12.3%) in the cannabidiol 10 mg/kg/day, cannabidiol 20 mg/kg/day, and placebo groups, respectively, at the end of treatment.

Health-Related Quality of Life

Patients included in the CARE1 Part B and CARE2 studies had a poor quality of life based on the low mean overall Quality of Life in Childhood Epilepsy (QOLCE) scores at baseline. Nonetheless, the adjusted mean differences for all QOLCE scores in both studies were in favour of cannabidiol 20 mg/kg/day treatment in the CARE1 Part B study, with an adjusted mean difference of 1.5 points (95% CI, –3.8 to 6.8; P = 0.576) over placebo, and 3.8 points (95% CI, –0.1 to 7.8; P = 0.058) and 1.8 points (95% CI, –2.2 to 5.8; P = 0.382) points over placebo for the 10 mg/kg/day and 20 mg/kg/day doses, respectively, in the CARE2 study.

Sleep Disruption and Function

In both the CARE1 Part B and CARE2 studies, the mean baseline scores in the sleep disturbance numerical rating scale (NRS) were similar across the treatment groups. In the CARE1 Part B study, a mean treatment difference of –0.4 (95% CI, –1.5 to 0.7) in sleep disruption score was observed, with no evidence of a significant difference between the cannabidiol 20 mg/kg/day and placebo groups. Similarly, in the CARE2 study, the mean treatment difference in sleep disruption score between the 10 mg/kg/day cannabidiol and placebo groups was 0.0 (95% CI, –0.9 to 0.8), and was –0.1 (95% CI, –0.9 to 0.8) between the 20 mg/kg/day cannabidiol and placebo groups.

CADTH evaluated the mean scores at baseline for the Epworth Sleepiness Scale (ESS) and found them to be relatively high in both trials in all treatment groups (> 7.1 in the CARE1 Part B study; > 7.2 in the CARE2 study). In the CARE1 Part B study, the mean treatment difference in ESS score between the 20 mg/kg/day cannabidiol and placebo groups was 1.51 (95% CI, –0.18 to 3.19; P = 0.078) in favour of placebo. In the CARE2 study, the mean treatment difference in ESS score between the 10 mg/kg/day and placebo groups was –0.55 (95% CI, –1.86, 0.75; P = 0.404) and 0.74 (95% CI, –0.57, 2.05; P = 0.267) between the 20 mg/kg/day and placebo groups.

Resource Use

In the CARE1 Part B study, a total of 6 patients (5%) reported 1 or more inpatient hospitalizations due to epilepsy during the treatment period: 5 patients (8.2%) in the 20 mg/kg/day cannabidiol group and 1 patient (1.7%) in the placebo group. In the CARE2 study, a total of 26 patients (13.1%) reported 1 or more inpatient hospitalizations due to epilepsy: 8 patients (11.9%) in the 20 mg/kg/day cannabidiol group, 12 patients (18.2%) in the 10 mg/kg/day cannabidiol group, and 6 patients (9.2%) in the placebo group.

The number of patients using rescue medication was similar overall in both studies. In the CARE1 Part B study, 36 patients (59.0%) and 41 patients (69.5%) in the cannabidiol 20 mg/kg/day and placebo groups, respectively, used rescue medication, while in the CARE2 study, the numbers in the cannabidiol 10 mg/kg/day, cannabidiol 20 mg/kg/day, and placebo groups were 54 (84.4%), 58 (84.1%), and 54 (80%) patients, respectively.

Harms Results

In the CARE1 Part B study, 57 of 61 patients (93.4%) in the 20 mg/kg/day cannabidiol group and 44 of 59 patients (74.6%) in the placebo group reported 1 or more AEs. In the CARE2 study, 56 of 64 patients (87.5%) in the 10 mg/kg/day cannabidiol group, 62 of 69 patients (89.9%) in the 20 mg/kg/day cannabidiol group, and 58 of 65 patients (89.2%) in the placebo group reported 1 or more AEs. The most common AEs (more than 10% of patients in any treatment group) reported in both studies were somnolence, diarrhea, and decreased appetite.

In the CARE1 Part B study, 10 of 61 patients (16.4%) in the 20 mg/kg/day cannabidiol group and 3 of 59 patients (5.1%) in the placebo group reported 1 or more serious AEs (SAEs). In the CARE2 study, 13 of 64 patients (20.3%) in the 10 mg/kg/day cannabidiol group, 17 of 69 patients (24.6%) in the 20 mg/kg/day cannabidiol group, and 10 of 65 patients (15.4%) in the placebo group reported 1 or more SAEs. The most common SAEs reported in both studies were nervous system disorders, status epilepticus, somnolence, and convulsion. Pneumonia was also a common SAE reported in the CARE2 study. All SAEs were resolved in the CARE1 Part B study, while 3 patients in the 20 mg/kg/day cannabidiol group in the CARE2 study had 3 SAEs that were not resolved at the end of the trial.

Patient discontinuation from treatment due to AEs was relatively low, although higher in the 20 mg/kg/day cannabidiol groups in both studies. In the CARE1 Part B study, AEs that led to discontinuation of the medication occurred in 9 of 61 patients (14.8%) in the 20 mg/kg/day cannabidiol group and in 1 of 59 patients (1.7%) in the placebo group while, in CARE2, 5 of 69 patients (7.2%) in the 20 mg/kg/day cannabidiol group experienced AEs leading to discontinuation from the study. No patients in the 10 mg/kg/day cannabidiol group or placebo groups withdrew from the study due to AEs. No patient deaths occurred during either study.

Critical Appraisal

Both the CARE1 Part B and CARE2 studies are randomized controlled trials (RCTs) involving an adequate randomization process, with an overall balanced distribution of participants to either the cannabidiol or placebo arms. There were some observed baseline imbalances in both studies; however, these were judged to have a low risk of introducing bias. There was good adherence to the intended interventions. There were, however, some imbalances observed in the use different cointerventions; although these possible deviations could introduce bias, the impact and direction of the bias on the outcomes of interest are uncertain. Some modifying effects from variables were observed (i.e., use of stiripentol, use of clobazam, and geographical location); however, the low number of patients across subgroups in both studies warrants caution in attributing any credible effect modification from any of these variables. There were no instances of meaningful missing outcome data. In both studies, the measurements of the outcomes were appropriate. The blinding of the participants and clinical investigators that was maintained throughout the studies mitigates potential biases in this domain. Overall, both studies demonstrated adherence to methodological consistency and minimized risks across all of the domains assessed for risk of bias for most outcomes when comparing cannabidiol with placebo. Several secondary end points depicting statistically significant results lacked multiplicity control, carrying a risk of false-positives; hence, cautious interpretation is needed due to potential random error.

Overall, the patients included in the CARE1 Part B and CARE2 trials have baseline characteristics and prognostic factors similar to those encountered in the population of patients in Canada with DS, according to the clinical experts consulted by CADTH. There were some concerns of uncertainty on the applicability of the results to adult populations older than 18 years of age since no patients older than 18 years were included in either trial. However, according to the clinical experts consulted by CADTH, it is unlikely that the response observed in the CARE1 Part B and CARE2 studies will be different in terms of beneficial effects and possible harms. There is also uncertainty as to whether the results can be generalized to patients with fewer than 4 seizures per month, since patients with such characteristics were not included in these studies. The trials excluded patients using medicinal cannabis or synthetic cannabinoid-based medications and transitioning to cannabidiol (pharmaceutical). This would be a common situation in Canada; the clinical experts suggested that this is an important consideration, but it is unlikely to affect the generalizability of the results of the studies.

The question of whether cannabidiol is more efficacious than the other treatment available in Canada for patients with DS (i.e., stiripentol) when added to standard of care is still uncertain. There is no head-to-head comparison of cannabidiol against stiripentol. Furthermore, the standard-of-care treatments commonly used in patients with DS vary and make it difficult to assess this question using an indirect treatment comparison, since such differences may include issues of inconsistency or intransitivity. Given the lack of head-to-head comparisons and with the current evidence at hand, it is difficult to draw a strong conclusion on this issue.

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, the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) tool was used to assess the certainty of the evidence for 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.

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

• percentage change from baseline in convulsive-seizure frequency during the treatment period

• patients with a 50% or greater reduction from baseline in convulsive-seizure frequency during the treatment period

• patients with a 75% or greater reduction from baseline in convulsive-seizure frequency during the treatment period

• number of convulsive seizure–free days

• percentage change from baseline in total seizure frequency during the treatment period

• patients with convulsive status epilepticus

• HRQoL

• sleep and sleep disruption (measured with ESS and 0 to 10 NRS scores)

• resource utilization (use of rescue medication and inpatient hospitalizations due to epilepsy)

• harms (AEs, SAEs, harms of special interest).

Results of GRADE Assessments

The GRADE assessments included an evaluation of the main outcomes considered important by clinicians, patient groups, and stakeholders. The comparisons evaluated in the GRADE assessments of this report compared cannabidiol 10 mg/kg/day against placebo and cannabidiol 20 mg/kg/day against placebo. In Table 2 and Table 3, we present the GRADE summary of findings for both comparisons, respectively.

Table 2: Summary of Findings for Cannabidiol 10 mg/kg/day Versus Placebo for Patients With DS

AE = adverse event; CI = confidence interval; DS = Dravet syndrome; ESS = Epworth Sleepiness Scale; HRQoL = health-related quality of life; MD = mean difference; NA = not applicable; NRS = numerical rating scale; OR = odds ratio; QOLCE = Quality of Life in Childhood Epilepsy; RCT = randomized controlled trial; SAE = serious adverse event; SD = standard deviation; vs. = versus.

Note: This comparison was obtained from the CARE2 study assessing the 10 mg/kg/day arm vs. placebo.

^aRated down 1 level for imprecision. The target of our certainty is on a nontrivial effect. The 95% CI includes the null and the threshold of a 5% meaningful difference between treatment and placebo, as informed by the clinical experts.

^bRated down for imprecision. The target of our certainty is an important benefit. The 95% CI includes the threshold of meaningful difference between treatment and placebo for 20 patients more (or fewer) per 1,000 treated as considered by the clinical experts consulted by CADTH.

^cRated down 2 levels for imprecision. No thresholds or CIs were assessed. Based on sample size, the number did not reach a plausible optimal information size.

^dRated down 2 levels for imprecision. Based on the target of the certainty of a meaningful effect of the intervention, the 95% CI was considered wide and no threshold for a minimal important difference could be obtained.

^eRated down for imprecision. No CIs could be assessed. Rated down due to small sample size that did not reach a plausible optimal information size.

Table 3: Summary of Findings for Cannabidiol 20 mg/kg/day Versus Placebo for Patients With DS

AE = adverse event; CI = confidence interval; DS = Dravet syndrome; ESS = Epworth Sleepiness Scale; HRQoL = health-related quality of life; MD = mean difference; NA = not applicable; NRS = numerical rating scale; OR = odds ratio; Q1 = lower quartile; Q3 = upper quartile; QOLCE = Quality of Life in Childhood Epilepsy; RCT = randomized controlled trial; SAE = serious adverse event; SD = standard deviation.

^aRated down for imprecision. The target of our certainty is an important benefit. The CI crosses a threshold of 20 patients more (or fewer) per 1,000 treated, as considered by the clinical experts consulted by CADTH.

^bRated down 1 for imprecision. The target of the certainty is that of any beneficial effect (based on the null). Only 1 study assesses this outcome. No thresholds or CIs were evaluated.

^cThe target of the certainty is that of an important benefit. The lower bound of the CI could include a trivial effect, the threshold for which was considered to be 5%.

^dNo thresholds or CIs were assessed. Numbers are not optimal to assess if the intervention provides a large or trivial effect; hence, it was rated down 2 levels for imprecision.

^eBased on the target of the certainty of a meaningful effect of the intervention, the 95% CI was considered wide and no threshold for a minimal important difference could be obtained. Sample size was considered low in relation to a plausible optimal information size.

Source: CARE1 Part B and CARE2 studies.

Long-Term Extension Studies

Description of Studies

CARE5 was a multicentre, open-label extension (OLE) study for patients with DS or Lennox-Gastaut syndrome (LGS) who had completed the double-blind, placebo-controlled, clinical studies with cannabidiol (CARE1, CARE2, CARE3, and CARE4 trials). The objective of this OLE study was to evaluate cannabidiol’s long-term safety and tolerability and effect on seizures as an adjunctive treatment in children and adults with inadequately controlled DS or LGS.

Efficacy Results

During weeks 37 to 48 of treatment, patients with DS experienced a median percentage change of –62.6% in total seizure frequency from their original study baseline. The proportion of patients who achieved a 50% or greater reduction in total seizure frequency during weeks 37 to 48 of treatment was 59.3%. Out of all patients with DS, 70.1% experienced a 25% or greater reduction in total seizure frequency, 39.7% experienced a 75% or greater reduction in total seizure frequency, and 6.1% experienced total seizure freedom (100% reduction).

During weeks 37 to 48 of treatment, patients with DS experienced a median percentage change of –54.2% in the frequency of convulsive seizures compared with their baseline frequency from the original study. The proportion of patients who achieved a 50% or greater reduction from baseline in the frequency of convulsive seizures during weeks 37 to 48 of treatment was 52.3%. Out of all patients with DS, 67.8% experienced a 25% or greater reduction from baseline in the frequency of convulsive seizures, 34.6% experienced a 75% or greater reduction in the frequency of convulsive seizures, and 7.9% experienced convulsive-seizure freedom (100% reduction). During the last 12 weeks of treatment, 4.5% of patients with DS reported convulsive seizures greater than 30 minutes in duration compared with a baseline of 4.8% during their original study. The proportion of patients with DS with nonconvulsive seizures greater than 30 minutes in duration during the last 12 weeks of treatment was 4.8% compared with a baseline of 7.2% during their original study.

Harms Results

A total of 306 patients (97.1%) with DS had 1 or more AEs during the study, with 71 patients (22.5%) reporting AEs of mild severity, 157 patients (49.8%) reporting AEs of moderate severity, and 78 patients (24.8%) reporting severe AEs. SAEs were reported for 133 participants (42.2%) in the DS group, with the most common SAEs being status epilepticus, convulsion, and pneumonia. There were 28 patients (8.9%) with DS who stopped treatment due to AEs, with the most common AEs leading to discontinuation being convulsion, increased AST, and increased ALT. A total of 6 patients (1.9%) with DS died during the study.

Critical Appraisal

The CARE5 study is a nonrandomized, open-label, single-arm study. The lack of comparison with an active comparator precludes the ability to assess the relative long-term therapeutic benefits or safety of cannabidiol versus other ASMs. Furthermore, the lack of blinding in the CARE5 study may affect subjective measures such as patient-reported outcomes. The direction and magnitude of this potential bias remains unclear.

Since completion of the CARE1 and CARE2 studies was an eligibility criterion for enrolment into the CARE5 study, patients who discontinued from the CARE1 or CARE2 study for any reason, such as AEs, withdrawal by patient or parent, or withdrawal by investigator, were excluded from the CARE5 study. Thus, enrolment into the CARE5 study was limited to those who tolerated and whose condition responded to cannabidiol. Moreover, only 54% of patients completed the study; as such, there is a risk of bias due to missing outcomes data. The proportion of patients who adhered to the study drug during the longer follow-up was not reported.

Indirect Comparisons

No indirect treatment comparisons were submitted by the sponsor.

Studies Addressing Gaps in the Systematic Review Evidence

No studies addressing gaps in the systematic review evidence were submitted by the sponsor.

Conclusions

The body of evidence informing the effects of cannabidiol on patients with DS consisted of 2 pivotal RCTs comparing cannabidiol (at doses of 10 mg/kg/day and 20 mg/kg/day) versus placebo, and 1 single-arm OLE study. Both pivotal studies inform the effects of cannabidiol on the outcomes of convulsive and total seizure control, HRQoL, sleep disruption, resource utilization, and AEs. All of these are considered by the clinical experts, patient groups, and stakeholders to be critical outcomes for decision-making.

The evidence shows that the use of cannabidiol, when compared with placebo, results in a greater reduction from baseline in the number of convulsive seizures. A greater proportion of patients using cannabidiol achieved reductions from baseline of 50% or more in the frequency of convulsive seizures, and 75% or more compared with placebo. Cannabidiol also likely provides beneficial albeit smaller improvement effects in the number of convulsive seizure–free days and in total seizure frequency. There is still uncertainty in the effects of cannabidiol on status epilepticus, HRQoL, sleep disruption, and use of rescue medications and hospital use.

Cannabidiol was generally well tolerated. However, the evidence suggests that cannabidiol may have more AEs and SAEs when compared with placebo, although there is uncertainty on the clinical significance of these differences to inform decision-making. The performance of cannabidiol against the other comparator used in Canada for patients with DS (i.e., stiripentol) is still uncertain, as no head-to-head comparison is available. Similarly, there is no evidence on the effects of cannabidiol as a first-line adjunctive treatment in patients older than 18 years of age, nor for its use in patients with DS who experience fewer seizures (i.e., fewer than 4 seizures per month).

Overall, the use of cannabidiol yields better estimates of seizure control with an adequate safety profile when compared with placebo. The effects on long-term outcomes of behaviour and HRQoL are still uncertain.

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 in the treatment of patients with seizures associated with DS.

Disease Background

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

DS is a very rare form of epilepsy associated with treatment-resistant, lifelong seizures and substantial comorbidities, such as intellectual disability and behavioural, sleep, and gait problems.¹ Epilepsy onset in DS usually occurs within the first year of life at between 5 and 8 months of age. It is characterized by febrile or afebrile clonic and tonic-clonic, generalized, and unilateral seizures that tend to be prolonged in infants who have previously developed as expected.⁸

Later on, multiple seizure types (mainly myoclonic, atypical absences, and focal seizures) and behavioural disorders may appear, and developmental and cognitive skills may slow.⁸ Convulsive seizures, consisting of generalized clonic, generalized tonic-clonic, or unilateral clonic seizures are present throughout the lifespan of patients with DS.⁸ The frequency of convulsions often gradually decreases over time, and status epilepticus rarely occurs after the age of 10.³

It has been reported that 70% to 85% of patients with clinical features of DS have mutations of the SCN1A gene that result in the dysfunction of voltage-gated sodium channels in neurons.^2-4

The most common causes of mortality in DS are SUDEP (accounting for nearly half of the overall mortality) and status epilepticus (accounting for one-third of all deaths). Other causes include accidents, drowning, or infection.³

For patients with epilepsy of any cause, the main SUDEP risk factors include a high frequency of convulsions and antiepileptic polytherapy, which are related to the severity of epilepsy.³ The most important mechanisms of SUDEP are cardiac arrhythmias, respiratory dysfunction, and dysfunction of the brainstem arousal system. The only effective way to prevent SUDEP is through the control of seizures.⁹

DS is associated with several comorbidities that, cumulatively, have substantial and lifelong impacts on the HRQoL of both patients and caregivers.¹⁰ Having fewer seizures and additional seizure-free days have been linked to improved HRQoL in patients with DS and their caregivers.¹⁰ According to a survey of parents of children younger than 6 years of age living with DS (n = 36), reduction in severe seizures and communication issues were the most important aspects of the disease that need to be addressed with a potential new drug, identified by more than 90% of respondents.¹¹

According to a population-based study in Sweden conducted between 2007 and 2011, the estimated incidence of DS was 1 in 33,000 live births, and the prevalence in 2011 was 1 in 45,700 children younger than 18 years of age.⁵ According to a retrospective cohort study of 355 patients with DS in the UK published in 2012, the estimated incidence of gene mutation–positive DS was at least 1 in 40,900 births.¹²

The diagnosis of DS is based primarily on clinical observations of tonic-clonic seizures during the first year of life, the occurrence of myoclonic seizures and ataxia later, impaired psychomotor development following the onset of seizures, and poor response to antiepileptic drugs. Confirmatory genetic testing for an SCN1A mutation in patients with suspected DS, especially those younger than 2 years of age (in whom a clinical diagnosis can be difficult), has been shown to decrease unnecessary testing and improve access to therapies and supportive-care services for families.¹³

The diagnosis of DS is largely clinical, as genetic testing for variants of the SCN1A gene alone is not sufficient for a diagnosis, since the variants can also be seen in other conditions.¹⁴

Standards of Therapy

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

Treatment goals for patients with DS are currently focused on balancing optimal seizure control with treatment side effects and patient quality of life.¹ Control of convulsive seizures should also be prioritized over nonconvulsive seizures due to their stronger association with SUDEP and impact on HRQoL.¹ An additional therapeutic goal is to minimize other nonseizure manifestations of DS such as cognitive disabilities, behavioural difficulties, and psychiatric issues.¹⁵ For patients with DS that is refractory to ASMs, it is challenging to optimize treatment regimens that will reduce the frequency of seizures. However, as every seizure increases the future risk of poor outcomes and reduced HRQoL, even a moderate improvement in seizure control can have cumulative benefits over a patient’s lifetime.^1,16

Based on the Ontario Epilepsy Guidelines,¹⁷ DS-specific international guidelines,^1,13,18,19 and Canadian clinical expert opinion,²⁰ valproic acid and clobazam are often used initially, but these are usually insufficient to control seizures.¹⁸ Stiripentol in conjunction with clobazam and valproate is currently the only ASM specifically indicated for DS in Canada,²¹ with the other ASMs indicated for general epilepsy or prescribed off-label.²² Due to the intractability of the seizures in patients with DS, there is a tendency to place patients on multiple ASMs, with the typical patient on a median of 3 ASMs.^1,20 However, some ASMs (e.g., carbamazepine, oxcarbazepine, lamotrigine, phenytoin) may exacerbate seizures and should be avoided.¹ Clinicians treating patients with DS that is refractory to initial therapies can attempt to add other ASMs, including stiripentol, topiramate, levetiracetam, clonazepam, and rufinamide. Cannabidiol is also recommended as an adjunctive treatment option for patients whose condition fails to respond to first-line ASMs.^1,19,23 Adjunctive nonpharmacological interventions, including a ketogenic diet and vagus nerve stimulation, may also be considered as another treatment option.¹ In Canada, there is no standardized approach to treatment for patients with DS.²⁰

Drug Under Review

The key characteristics of cannabidiol are summarized in Table 4 along with those of other therapies available for the treatment of seizures associated with DS.

The requested indication for cannabidiol under review by Health Canada is as adjunctive therapy for the treatment of seizures associated with LGS, DS, or tuberous sclerosis complex in patients 2 years of age and older.

Jazz Pharmaceuticals Canada requested that CADTH review each neurologic disorder (i.e., LGS, DS, and tuberous sclerosis complex) as a separate submission and this request was accepted on June 21, 2023. The specific requested listing for Epidiolex (cannabidiol) for this review is as adjunctive therapy for the treatment of seizures associated with DS in patients 2 years of age and older. A deviation request was accepted by CADTH on September 15, 2023, permitting the reimbursement request (i.e., for patients 2 years of age or older) to deviate from the anticipated Health Canada indication (i.e., for patients 1 year of age or older).

Dosing and Administration

The recommended starting dosage is 2.5 mg/kg by mouth twice daily (5 mg/kg/day).²⁴ After 1 week, the dosage can be increased to a recommended maintenance dosage of 5 mg/kg twice daily (10 mg/kg/day). Based on individual clinical response and tolerability, the dosage 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 dosage may be increased no more frequently than every other day.

Mechanism of Action

Cannabidiol (Epidiolex) is a highly purified, plant-derived pharmaceutical formulation of cannabidiol administered as an oral solution. Although not fully elucidated, cannabidiol is thought to have novel mechanisms of action that are different from those of other ASMs.²⁴

Table 4: Key Characteristics of Cannabidiol and Antiepileptic Medications Considered Usual Care Interventions

ASM = antiseizure medication; DS = Dravet syndrome; LGS = Lennox-Gastaut syndrome; NA = not applicable; SMEI = severe myoclonic epilepsy in infancy; TSC = tuberous sclerosis complex.

^aHealth Canada–approved indication.

^bIncludes valproate.

Sources: Product monographs for cannabidiol,²⁴ valproic acid,²⁵ clobazam,²⁶ stiripentol,²¹ levetiracetam,²⁷ topiramate,²⁸ clonazepam,²⁹ and rufinamide.³⁰

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.

CADTH received 1 patient input, which was from the CEA. The CEA is dedicated to the promotion of independence and quality of life for people with epilepsy and their families by providing support services, information, advocacy, and increasing public awareness. Information for this submission was gathered by the president of the CEA. The input drew on the knowledge and experiences of the CEA’s 24 member associations and the lived experiences and first-hand knowledge of patients, caregivers, clinicians, volunteers, and supporters (donors and funders).

The CEA highlighted that individuals with uncontrolled epilepsy are at risk of social isolation and mental illness. The CEA noted that the unpredictable nature of seizures and side effects of medications, such as anxiety, depression, mood swings, sexual dysfunction, suicidal thoughts, and exhaustion, have negative effects on patients and their family and caregivers.

The CEA stated that currently available treatments do not control seizures in all patients and there is a percentage of patients with uncontrolled seizures. Lack of access to an approved treatment among patients with uncontrolled seizures will result in trying alternative practices and treatments such as cannabis and unregulated substances. The CEA noted that even a small reduction in the number of seizures can improve quality of life among patients.

According to the CEA input, because of the frequent seizures, patients with epilepsy are often unemployed or underemployed, with restricted income and without access to employer-funded insurance plans, which limit their access to the drugs that are not on the provincial formulary. Although the CEA could not comment on patient or caregiver experience with cannabidiol, it noted that each new drug that is brought to market offers hope to patients whose seizures are uncontrolled by 1 or a combination of existing therapies.

Clinician Input

Input From the Clinical Experts Consulted by CADTH

All CADTH review teams include at least 1 clinical specialist with expertise regarding 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 DS.

Unmet Needs

Treatment goals, according to the clinical experts consulted by CADTH, include improving seizure control with the improvement of HRQoL and decreasing seizure burden without affecting the mood, cognition, or behaviour of patients. Other goals include increasing the number of seizure-free days and decreasing visits to health care facilities and the need for rescue medications.

Also, the clinical experts added that reducing the risk of sudden death and preventing status epilepticus with treatments that do not adversely affect behaviour or mood or cause excessive sedation, are important needs in patients with DS.

The experts mention that patients with DS, tuberous sclerosis complex, and LGS rarely become seizure free. Patients with DS would benefit from more ASM options that improve seizure control without adverse behavioural and cognitive effects. Currently, most ASMs are either effective with many adverse effects or largely ineffective with few side effects.

Place in Therapy

Cannabidiol has a unique mechanism of action not shared by any other ASM. According to the clinical experts, cannabidiol will likely complement other ASMs and could become an important inclusion in the therapeutic options for DS. The possible advantage of cannabidiol is that it may have fewer adverse effects when compared with other drugs indicated for this condition. The clinical experts anticipated that, initially, cannabidiol would be used after valproic acid and clobazam.

Patient Population

Both experts agreed that the decision to prescribe cannabidiol hinges on the judgment of the treating neurologist. There is no need for an electroencephalogram (EEG) or MRI biomarkers. The diagnosis is usually made by an epileptologist or neurologist with expertise in epilepsy, since the diagnosis can be missed by general neurologists.

The clinical experts commented that, in their experience, using cannabidiol would be considered when the patient’s condition has failed to respond to 2 or more appropriate ASMs, including valproic acid, clobazam, stiripentol, or topiramate. They stated that, in their clinical practice, they would opt for these interventions as early as feasible. Evaluating treatment response is part of a clinician’s routine practice, incurring no additional time cost.

Assessing the Response Treatment

In clinical practice, the clinical experts suggested that the frequency of seizures and change over time, the number of seizure-free days, a decrease in seizure duration and severity, a reduction of status epilepticus, and a decrease in the use of rescue medication are important when assessing response to treatment.

Discontinuing Treatment

In general, little or no improvement in seizure frequency (approximately less than a 50% in change from baseline), or AEs from the intervention that become intolerable for patients and caregivers, are factors the clinical experts would take into consideration to determine whether to discontinue the use of the medication.

Prescribing Considerations

An outpatient epilepsy clinic is an ideal setting for initiating cannabidiol treatment and monitoring patients with DS; a hospital admission would rarely be needed. It would also be ideal to have an epileptologist or neurologist with expertise in the treatment of DS monitor response to treatment in these patients.

Additional Considerations

The clinical experts considered that the use of cannabidiol would be appropriate in adults, even though the evidence from the clinical trials generally excludes the adult population.

Clinician Group Input

No clinician group input was received by CADTH.

Drug Program Input

The drug programs provide input on each drug being 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 5.

Table 5: Summary of Drug Plan Input and Clinical Expert Response

ALT = alanine aminotransferase; ASM = antiseizure medication; AST = aspartate aminotransferase; BIA = budget impact analysis; CDEC = Canadian Drug Expert Committee; DS = Dravet syndrome.

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 (Epidiolex) 100 mg/mL oral solution as an adjunctive treatment of seizures associated with DS in patients 2 years of age and older. The focus of this report is on comparing cannabidiol with 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 4 sections, with CADTH’s critical appraisal of the evidence included at the end of each section. The first section, the systematic review, includes pivotal studies and 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 section includes sponsor-submitted long-term extension studies. The third and fourth sections would include indirect evidence and additional studies that were considered to address important gaps in the systematic review evidence, respectively; however, neither was submitted by the sponsor.

Included Studies

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

• 2 pivotal RCTs identified in the systematic review

• 1 long-term extension study.

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

A total of 2 double-blind, placebo-controlled phase III RCTs were included in the systematic literature review. The characteristics of the included studies are summarized in Table 6.

Table 6: Details of Studies Included in the Systematic Review

ASM = antiseizure medication; CaGIC = Caregiver Global Impression of Change; CaGICSD = Caregiver Global Impression of Change in Seizure Duration; DS = Dravet syndrome; ECG = electrocardiogram; ESS = Epworth Sleepiness Scale, NRS = numerical rating scale; OLE = open-label extension; QOLCE = Quality of Life in Childhood Epilepsy; RCT = randomized controlled trial; Vineland-2 = Vineland Adaptive Behavior Scales, 2nd Edition; VNS = vagus nerve stimulation.

^aOut of 23 trial sites, 22 sites randomized patients into the CARE1 Part B study.

^bOut of 43 trial sites, 38 sites randomized patients into the CARE2 study.

^cFollowing advice from the European Medicines Agency on the protocol, the response analysis of the ≥ 50% reduction in convulsive-seizure frequency was considered a key secondary end point for European regulatory submissions only.

Sources: CARE1³⁸ and CARE2 Clinical Study Reports.³⁹ Details included in the table are from the sponsor’s summary of clinical evidence.

CARE1 Part B Study

The objective of the CARE1 Part B study was to assess the efficacy and safety of cannabidiol as an adjunctive antiepileptic treatment compared with placebo with respect to the percentage change from baseline in convulsive-seizure frequency during the treatment period of the trial. This study was a multisite, randomized, double-blind, placebo-controlled trial. A total of 120 patients were randomized to double-blind treatment between March 30 and November 26, 2015. The trial was conducted at 23 trial sites in 4 countries: France, Poland, the UK, and the US. A 28-day screening period was included to perform tests to detect the presence of tetrahydrocannabinol metabolites (i.e., to confirm that the patient had not taken cannabis immediately before entering the trial) and ensure that all 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 convulsive and nonconvulsive seizures. Eligible patients were randomized to 20 mg/kg/day of cannabidiol or an equivalent volume of placebo, titrating from a 2.5 mg/kg/day dose up to a 20 mg/kg/day maintenance dose over 11 days. Patients were randomly allocated to intervention or placebo groups using an interactive voice response system or interactive web response system. The allocation of investigational medicinal product (IMP) to patient identifier was done according to randomization schedules produced by an independent statistician. The randomization was stratified by age group (2 to 5 years, 6 to 12 years, and 13 to 18 years of age). The trial design of CARE1 Part B is summarized in Figure 1.

Figure 1: CARE1 Part B Trial Design

OLE = open-label extension.

^a For patients who did not enter the OLE trial at visit B8 or for those who withdrew early and tapered the investigational medicinal 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 B9 (or date of final dosing) until visit B10.

^b For patients who did not enter the OLE trial or who withdrew from the trial early; could be conducted by telephone.

Source: CARE1 Clinical Study Report.³⁸

CARE2 Study

The CARE2 study shared the same objective and study design as the CARE1 Part B trial. A total of 199 patients were randomized to double-blind treatment between April 13, 2016, and April 9, 2018. The trial was conducted at 43 sites in 6 countries: Australia, Israel, Netherlands, Poland, Spain, and the US. Of these, 38 sites randomized patients into the trial. A 28-day screening period was included that was similar to the CARE1 Part B study. Eligible patients were randomized to receive cannabidiol 10 mg/kg/day, cannabidiol 20 mg/kg/day, placebo 10 mg/kg/day dose-volume equivalent, or placebo 20 mg/kg/day dose-volume equivalent at a 2:2:1:1 ratio. Patients assigned to either dose level titrated from a 2.5 mg/kg dose up to the assigned 10 mg/kg or 20 mg/kg maintenance dose over 11 days. Patients were randomly allocated using an interactive web response system. The allocation of IMP to patient number identifier was done according to a randomization schedule produced by an independent statistician. The randomization was stratified by age group (2 to 5 years, 6 to 12 years, and 13 to 18 years of age). The trial design of CARE2 is summarized in Figure 2.

Figure 2: CARE2 Trial Design

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

^a For patients who did not enter the OLE trial at visit B8 or who withdrew early and tapered the investigational medicinal 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 B9 (or date of final dosing) until visit B10.

^b For patients who did not enter the OLE trial or who withdrew from the trial early; could be conducted by telephone.

^c The placebo groups were pooled for the analyses of efficacy.

Source: CARE2 Clinical Study Report.³⁹

Populations

Inclusion and Exclusion Criteria

Key inclusion criteria for the CARE1 Part B and CARE2 studies were patients who were aged 2 to 18 years (inclusive), had 4 or more convulsive seizures per 28 days, had a confirmed diagnosis of DS, had been taking 1 or more ASMs at a dose that had been stable for at least 4 weeks, and had seizures that were not completely controlled by their current ASMs. Key exclusion criteria were having unstable medical conditions in the 4 weeks before screening, a history of substance abuse (including alcohol), use of recreational or medical cannabis in the previous 3 months, and initiated use of felbamate within the past year. Overall, patient inclusion and exclusion criteria were the same for both studies and are reported in more detail in Table 6.

Interventions

Both the CARE1 Part B and CARE2 trials compared 20 mg/kg/day cannabidiol versus placebo as an adjunctive antiepileptic treatment, while CARE2 also included the 10 mg/kg/day cannabidiol arm. In both trials, the same titration period was followed. Patients in the 20 mg/kg/day cannabidiol and dose volume–equivalent placebo treatment arms followed a titration schedule starting with a 2.5 mg/kg dose on days 1 and 2 during the treatment period, a 5 mg/kg dose on days 3 and 4, a 7.5 mg/kg dose on days 5 and 6, a 10 mg/kg dose on days 7 and 8, a 15 mg/kg dose on days 9 and 10, and a maintenance dose of 20 mg/kg from day 11 onward. In the CARE2 study, the 10 mg/kg/day cannabidiol and dose volume–equivalent placebo treatment arms followed the same titration schedule up to day 7, when the maintenance 10 mg/kg/day dose was reached and was to remain constant throughout the rest of the treatment period. Throughout the duration of both trials, doses of concomitant ASMs and any nonpharmacological regimens for epilepsy were to remain stable. The doses of these drugs could be modified by the treating physician. The use of rescue medications was allowed and recorded by the investigators. Any new medications or interventions for epilepsy (including a ketogenic diet) or changes in dosage within 4 weeks before or during the trial were not allowed. Also, recreational medical cannabis and synthetic cannabinoid-based medications were not allowed during the trial.

Outcomes

A list of efficacy end points assessed in this report is provided in Table 7 followed in Table 8 by descriptions of the outcome measures and properties. The 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 the stakeholder input from the patient group and the public drug plans. Using the same considerations, the CADTH review team selected end points that were considered to be 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 selected and summarized efficacy end points were assessed using GRADE. AEs and select notable harms outcomes considered important for informing CADTH’s expert committee deliberations were also assessed using GRADE. The outcomes selected included: frequency of total and convulsive seizures (change from baseline and as rates of total seizure reduction, e.g., ≥ 50% reduction), convulsive seizure–free days, and status epilepticus rates. Also, the HRQoL end point measured by the QOLCE scale, sleep disruption (ESS score and sleep disturbance NRS 0 to 10 score), and the use of rescue medications and hospitalizations due to epilepsy. Other outcomes not included in the GRADE framework are described in Appendix 1 (Table 23).

Outcomes in the Pivotal Trials

Both the CARE1 Part B and CARE2 studies had the same primary and secondary efficacy outcomes, with the exception of Cognitive Assessment Battery scores and number of convulsive seizure–free days, which were secondary end points assessed only in the CARE2 study. The outcomes assessed in the CARE1 Part B and CARE2 studies that were considered important for this review are summarized in Table 7.

The primary end point in both studies was the percentage change in convulsive seizures during the treatment period compared with baseline in patients taking cannabidiol compared with placebo. The clinical experts consulted by CADTH agreed this is a critical outcome for decision-making. In the pivotal trials, patients or their caregivers recorded the number and type of convulsive (tonic, clonic, tonic-clonic, or atonic) and nonconvulsive seizures (myoclonic, partial, or absence) each day from screening until completion of dosing using an interactive voice response diary. The seizure frequency during each period was based on 28-day averages. The proportion of patients with a reduction from baseline in convulsive-seizure frequency of 25% or greater, 50% or greater, 75% or greater, and 100% were reported as secondary end points in both studies. The percentage change for other individual seizure types, total seizure frequency, and status epilepticus were also reported as secondary end points. A minimal important difference (MID) for the reduction in seizure frequency is sometimes considered to be a relative reduction from baseline of 50% or greater,⁴⁰ although this threshold is largely arbitrary based on historical European Medicines Agency and FDA regulatory approval processes in the mainly “general” (i.e., milder) forms of epilepsy.⁴¹ Patients with DS may have a more severe form of epilepsy that is treatment-refractory, and there may be more uncertainty around a threshold of important benefit.⁴²

Sleep disruption (measured using the sleep disturbance NRS or the ESS) and HRQoL (assessed using the QOLCE questionnaire) were secondary end points assessed in both studies. For sleep disruption, caregivers were asked to indicate the number that best described their child’s sleep disruption in the past week from a scale of 0 to 10. Scores ranged from 0 (slept extremely well) to 10 (unable to sleep at all). The ESS is a self-reported questionnaire that was completed by each patient’s caregiver. The total score was based on the total of the 8 item scores in the assessment and could range from 0 to 24, with higher total scores representing greater levels of daytime sleepiness. The QOLCE questionnaire was completed by the parent or caregiver of patients aged 4 years and older. QOLCE scores ranged from 0 to 100 for each subscale, where 0 represents the lowest level of functioning and 100 represents the highest level of functioning. The overall quality-of-life score was calculated by taking the mean of the subscale scores. None of these measurements had a published MID value applicable to patients with DS or other measures of the validity, reliability, or responsiveness of the scales.

All AEs and inpatient hospitalizations due to epilepsy observed by the investigator or reported by the patient or caregiver were recorded on the patient’s case report form at all trial visits. An AE was defined as any new unfavourable or unintended signs or symptoms, or diagnosis or worsening of a pre-existing condition, that occurred following screening or at any point up to the posttreatment safety follow-up visit and that may or may not be considered related to the IMP. Any event that was the result of a trial procedure was to be recorded as an AE. Patients’ expected seizure types were not documented as AEs, but any worsening, including change in the pattern or severity of seizures, was to be documented as an AE. An AE was considered serious if it was: fatal, life-threatening, or required inpatient hospitalization or prolonged existing hospitalization; was persistently or significantly disabling or incapacitating; was a congenital anomaly or birth defect; or was a medically significant event that, based upon appropriate medical judgment, may have jeopardized the patient and may have required medical or surgical intervention to prevent 1 of the outcomes listed.

Table 7: Outcomes Summarized From the Pivotal Studies, CARE1 Part B and CARE2

ESS = Epworth Sleepiness Scale; NA = not applicable; NRS = numerical rating scale; QOLCE = Quality of Life in Childhood Epilepsy.

^aOutcomes from both cannabidiol arms (10 mg/kg/day and 20 mg/kg/day) were compared with the placebo group.

Sources: CARE1³⁸ and CARE2 Clinical Study Reports.³⁹ Details included in the table are from the sponsor’s summary of clinical evidence.

Table 8: Summary of Outcome Measures and Their Measurement Properties

EMA = European Medicines Agency; ESS = Epworth Sleepiness Scale; MID = minimal important difference; NRS = numerical rating scale.

Sources: CARE1³⁸ and CARE2³⁹ Clinical Study Reports.

Statistical Analysis

Clinical Trial End Points

In the CARE1 Part B study, the primary end point and all other secondary end points related to seizure frequency were analyzed using a Wilcoxon rank sum test. The estimated median differences for all seizure frequency–related end points between the cannabidiol and placebo treatment arms, together with a 95% CI, were calculated using the Hodges-Lehmann approach. Wilcoxon rank sum tests were also conducted in seizure frequency–related end points in the CARE2 study but were treated as part of the sensitivity analyses. In the CARE2 study, the differences between the cannabidiol and placebo treatment arms were analyzed using a negative binomial regression analysis, with the estimated ratio of least squares means of each cannabidiol treatment group to placebo (with 95% CI) presented as the primary outcome. Primary analyses were performed using the intention-to-treat (ITT) analysis set. Analyses using the per-protocol analysis sets were performed for the primary and key secondary end points only.

In both the CARE1 Part B and CARE2 studies, the proportion of patients whose condition was considered to have responded to treatment (defined as those with a ≥ 25%, ≥ 50%, ≥ 75%, or 100% reduction from baseline in the frequency of convulsive seizures during the treatment period) was analyzed using a Cochran-Mantel-Haenszel test stratified by age group.

In both studies, an analysis of covariance (ANCOVA) with baseline and age group as covariates and treatment group as fixed factor was used to analyze sleep disruption in the sleep disturbance NRS (0 to 10 scale), ESS, and QOLCE scores. A summary of the statistical analyses is presented in Table 9 and Table 24.

Table 9: Statistical Analysis of Efficacy End Points in the Pivotal Trials

ANCOVA = analysis of covariance; CaGIC = Caregiver Global Impression of Change; CaGICSD = Caregiver Global Impression of Change in Seizure Duration; CI = confidence interval; CMH = Cochran-Mantel-Haenszel; ESS = Epworth Sleepiness Scale; ITT = intention to treat; IVRS = interactive voice response system; LOCF = last observation carried forward; MI = multiple imputation; MMRM = mixed-model for repeated measures; MNAR = missing not at random; NOCB = next observation carried backward; NR = not reported; NRS = numerical rating scale; PP = per protocol; QOLCE = Quality of Life in Childhood Epilepsy; Vineland-2 = Vineland Adaptive Behavior Scales, 2nd Edition.

Sources: CARE1³⁸ and CARE2³⁹ Clinical Study Reports.

Sample Size and Power Calculation

The CARE1 Part B study planned to enrol 100 patients randomly allocated on a 1:1 basis (cannabidiol versus placebo) and stratified by age group (2 to 5 years, 6 to 12 years, and 13 to 18 years of age). The investigators assumed that patients in the placebo group would experience a mean reduction in convulsive-seizure frequency of 18% from baseline, for which a sample size of 50 patients per group would be sufficient to detect a difference of 32% in treatment effect between treatment groups. This was based on a standard deviation of 56% using a 2-sided 5% significance level and 80% power.

The CARE2 study planned to randomize a total of 186 patients across 4 treatment groups (cannabidiol 10 mg/kg/day, cannabidiol 20 mg/kg/day, placebo 10 mg/kg/day dose-volume equivalent, or placebo 20 mg/kg/day dose-volume equivalent) in a 2:2:1:1 ratio, stratified by age group (2 to 5 years, 6 to 12 years, and 13 to 18 years of age). The investigators determined that a sample size of 62 per group (after pooling the placebo groups) was required to obtain a power of at least 80% for a Wilcoxon-Mann–Whitney test comparing 2 distributions with a 2-sided significance level of 0.05. This was based on a gamma distribution for the cannabidiol groups with a scale parameter of 65.614 and a shape parameter of 1.0886, and a gamma distribution for the placebo group with a scale parameter of 40.887 and a shape parameter of 2.3059. Maximum likelihood estimates using the Newton-Raphson approximation were computed for the scale and shape parameters using data from the CARE1 Part B study.

Statistical Testing

In both the CARE1 Part B and CARE2 studies, all statistical tests were 2-sided with a 5% significance level. No formal adjustment of statistical significance for multiple testing was performed for the CARE1 Part B study.

In the CARE2 study, each end point had 2 comparisons against placebo (cannabidiol 20 mg/kg/day and 10 mg/kg/day versus placebo). The primary and key secondary end points were tested using a hierarchical gate-keeping procedure to control for type I error.

The key secondary end points were defined as follows:

• First key secondary end point: percentage change from baseline in total seizure frequency during the treatment period

• Second key secondary end point: number of patients with a 50% or greater reduction in convulsive-seizure frequency from baseline during the treatment period

• Third key secondary end point: Caregiver Global Impression of Change (CaGIC) scores

The null hypothesis of an end point at the level of 0.05 (2-sided) must have been rejected to test the hypothesis of the subsequent end point in the sequence at the level of 0.05 (2-sided). If a null hypothesis was not rejected, then testing would stop and all subsequent analyses would be declared not statistically significant.

Subgroup Analyses

In both trials, subgroup analyses were performed for the primary efficacy end point and for patients with a 25% or greater, 50% or greater, 75% or greater, or 100% reduction in the frequency of convulsive seizures from baseline using the ITT analysis set. All statistical significance was to be tested at the 0.05 level without formal adjustment for multiplicity. The same analyses were performed as for the primary efficacy variable using data from the treatment period. In the CARE1 study, for patients with a 25% or greater, 50% or greater, 75% or greater, or 100% reduction in the frequency of convulsive seizures, analyses were performed using a Fisher exact test. In the CARE2 study, for the key secondary end point of a 50% or greater reduction in convulsive-seizure frequency, patients with a 50% or greater reduction in seizure frequency were modelled using logistic regression, including stratified age group and treatment arm as covariates. The model also included covariates for each level of the effect being tested (excluding a reference level), both individually and with interactions with the treatment arm. A separate model was used for testing each effect.

The following subgroups were used In both the CARE1 Part B and CARE2 studies:

• age group (2 to 5 years, 6 to 12 years, and 13 to 18 years of age)

• sex (male, female)

• region (US, rest of the world)

• clobazam use (yes, no)

• valproic acid use (yes, no)

• stiripentol use (yes, no)

• baseline average convulsive-seizure frequency per 28 days (≤ observed tertile 1, > observed tertile 1 to ≤ observed tertile 2, > observed tertile 2)

• number of current ASMs (< 3, ≥ 3)

• number of prior ASMs (< 4, ≥ 4).

The CARE2 study also included the following subgroups:

• clobazam use and stiripentol use (yes/yes, yes/no, no/yes, no/no)

• levetiracetam use (yes, no)

• topiramate use (yes, no)

• number of prior and concurrent ASMs (< 8, ≥ 8).

Analysis Populations

A summary of study populations in the CARE1 Part B and CARE2 studies is presented in Table 10.

Table 10: Analysis Populations of the CARE1 Part B and CARE2 Studies

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

Sources: CARE1³⁸ and CARE2³⁹ Clinical Study Reports. Details included in the table are from the sponsor’s summary of clinical evidence.

Results

Patient Disposition

In the CARE1 Part B and CARE2 studies, 32% and 30% of patients failed screening, respectively, most commonly due to not meeting the eligibility criteria or reported as “other reason” (not defined). After randomization, the discontinuation rates were higher in the cannabidiol 10 mg/kg/day and 20 mg/kg/day arms versus the placebo arms, the majority due to AEs (Table 11) when considering both studies.

Table 11: Summary of Patient Disposition From Studies Included in the Systematic Review

AE = adverse event; ITT = intention to treat; PP = per protocol.

Note: Details included in the table are from the sponsor’s summary of clinical evidence.

^aTwo patients randomized to receive 10 mg/kg/day cannabidiol were mistakenly given dosing schedules for patients receiving 20 mg/kg/day at visit 2 (day 1) and thus received > 10 mg/kg/day dosing volumes before the error was corrected. In 1 patient, the error was not identified and rectified until visit 6 (day 57), while the other continued at the higher dose throughout the trial. These patients were analyzed according to the treatment group to which they were randomized, unless otherwise stated.

^bPooled results from the 10 mg/kg/day and 20 mg/kg/day dose volume–equivalent placebo groups from the CARE2 study are presented.

^cPatients could have more than 1 reason for screen failure.

Source: CARE1³⁸ and CARE2³⁹ Clinical Study Reports.

Baseline Characteristics

The baseline characteristics outlined in Table 12 are limited to those that are most relevant to this review or were felt to affect the outcomes or interpretation of the study results. In general, the initial characteristics were evenly distributed among the groups in both studies. However, slight discrepancies were observed in the variables of sex and age. These disparities are likely attributable to random variation and the limited sample size evaluated.

Exposure to Study Treatments

Exposure and adherence information for the CARE1 Part B and CARE2 studies (safety analysis set) are summarized in Table 13. Lower mean patient-days of exposure in the CARE2 study in the 20 mg/kg/day group (93.9 patient-days versus 99.5 and 98.3 patient-days in the placebo and 10 mg/kg/day groups, respectively) were aligned with higher discontinuation rates due to AEs in this treatment group.

Table 12: Summary of Baseline Characteristics From Studies Included in the Systematic Review (Safety Analysis Set)

ASM = antiseizure medication; NA = not applicable, NR = not reported, SD = standard deviation.

^aThe pooled results from the 10 mg/kg/day and 20 mg/kg/day dose volume–equivalent placebo groups from the CARE2 study are presented.

^bThese patients were all from France, where identification of race is subject to a data-protection law.

Sources: CARE1³⁸ and CARE2³⁹ Clinical Study Reports. Details included in the table are from the sponsor’s summary of clinical evidence.

Table 13: Summary of Patient Exposure From Studies Included in the Systematic Review (Safety Analysis Set)

NR = not reported; SD = standard deviation.

Source: Details included in the table are from the sponsor’s summary of clinical evidence.

Concomitant Medications and Cointerventions

The use of concomitant medications (excluding rescue medications) in the safety populations of the CARE1 Part B and CARE2 studies is summarized in Table 14.

In the CARE1 Part B study, overall, the utilization of concurrent ASMs was evenly distributed among the groups, with stiripentol being the only exception, with 30 patients (49.2%) and 21 patients (35.6%) in the cannabidiol and placebo arms, respectively, using stiripentol. All 120 patients (100%) took 1 or more concomitant ASMs during the trial. The most common classes of ASMs used concomitantly (i.e., reported in more than 50% of patients overall) were other antiepileptics, which were used by 101 patients (84.2%); benzodiazepine derivatives, which were used by 98 patients (81.7%); and fatty acid derivatives, which were used by 71 patients (59.2%). The most commonly used concomitant ASM was clobazam, which was used by 78 patients (65.0%); followed by valproic acid, which was used by 68 patients (56.6%); and stiripentol, which was used by 51 patients (42.5%).

Similarly, in the CARE2 study, the utilization of concurrent ASMs was evenly distributed among the groups, with clonazepam and topiramate being the exception. All 198 patients (100.0%) took 1 or more concomitant ASMs during the trial. The most common classes of ASMs used concomitantly (i.e., reported in more than 50% of patients overall) were other antiepileptics, which were used by 161 patients (81.3%); benzodiazepine derivatives, which were used by 148 patients (74.7%); and fatty acid derivatives, which were used by 139 patients (70.2%). The most commonly used ASM was valproic acid, which was used by 139 patients (70.2%); clobazam, which was used by 126 patients (63.6%); stiripentol, which was used by 71 patients (35.9%); levetiracetam, which was used by 54 patients (27.3%); and topiramate, which was used by 46 patients (23.2%).

Table 14: Concomitant Antiseizure Medications Used in the CARE1 Part B and CARE2 Studies (Safety Analysis Set)

NA = not applicable; NR = not reported.

^aFor the CARE2 study, the preferred terms of valproate semisodium and valproate sodium were combined with the preferred term of valproic acid.

Sources: CARE1³⁸ and CARE2³⁹ Clinical Study Reports.

Efficacy

The results of the main efficacy outcomes (from the ITT populations in the pivotal trials, CARE1 Part B and CARE2) are summarized in Table 15. These end points represent those considered important for decision-making by the clinical experts consulted by CADTH and other relevant stakeholders, such as committee members and patient representatives. These outcomes are also assessed in the GRADE framework and summary of findings. The rest of the end points submitted by the sponsor are reported in Table 23 in Appendix 1.

Table 15: Summary of Key Efficacy Results From Studies Included in the Systematic Review (ITT Analysis Set)

CI = confidence interval; ESS = Epworth Sleepiness Scale; HRQoL = health-related quality of life; ITT = intention to treat; NA = not applicable; NR = not reported; NRS = numerical rating scale; Q1 = lower quartile; Q3 = upper quartile; QOLCE = Quality of Life in Childhood Epilepsy; SD = standard deviation.

^aAnalyzed using Wilcoxon rank sum test. This statistical test was used as a sensitivity analysis in the CARE2 study.

^bNo multiplicity adjustment was made for this end point.

^cAnalyzed using negative binomial regression. This statistical test was used in the primary analysis of the CARE2 study.

Sources: CARE1³⁸ and CARE2³⁹ Clinical Study Reports.

Seizure Control

Percentage Change From Baseline in Convulsive-Seizure Frequency During the Treatment Period

In the CARE1 Part B study, patients in the 20 mg/kg/day cannabidiol group achieved a median percentage change from baseline in convulsive-seizure frequency of –38.9% (95% CI, –69.5% to –4.8%) during the 14-week treatment period compared with –13.3% (95% CI, –52.5% to 20.2%) for the placebo group. The estimated median difference between treatment arms was –22.8% (95% CI, –41.1% to –5.4%; P = 0.0123).

In the CARE2 study, the percentage reduction in convulsive-seizure frequency from baseline was 48.7% (95% CI, 37.9% to 57.6%) for the 10 mg/kg/day cannabidiol group and 45.7% (95% CI, 34.2% to 55.2%) for the 20 mg/kg/day cannabidiol group versus 26.9% (95% CI, 11.9% to 39.4%) for the placebo group. The treatment ratio between the 10 mg/kg/day cannabidiol group and the placebo group was 0.702 (95% CI, 0.538 to 0.916; P = 0.0095), and between the 20 mg/kg/day cannabidiol group and the placebo group was 0.743 (95% CI, 0.568 to 0.971; P = 0.0299).

Proportion of Patients With a 50% or Greater Reduction in Convulsive-Seizure Frequency From Baseline During the Treatment Period

In the CARE1 Part B study, the proportion of patients with a reduction of 50% or greater in their baseline frequency of convulsive seizures was 42.6% in the 20 mg/kg/day cannabidiol group and 27.1% in the placebo group. The OR for achieving this end point in the 20 mg/kg/day cannabidiol group compared with placebo was 2.00 (95% CI, 0.93 to 4.30; P = 0.0784).

In the CARE2 study, the proportion of patients with a reduction of 50% or greater in their baseline frequency of convulsive seizures was 43.9% in the 10 mg/kg/day cannabidiol group and 49.3% in the 20 mg/kg/day cannabidiol group compared with 26.3% in the placebo group. The odds of achieving this end point was an OR of 2.21 (95% CI, 1.06 to 4.62; P = 0.0332) in the 10 mg/kg/day group and an OR of 2.74 (95% CI, 1.32 to 5.70; P = 0.0069) in the 20 mg/kg/day group compared with placebo.

Proportion of Patients With a 25% or Greater, 75% or Greater, or 100% Reduction From Baseline in Convulsive-Seizure Frequency During the Treatment Period

In the CARE1 Part B study, the proportion of patients with a reduction of 75% or greater in their baseline convulsive-seizure frequency was 23.0% in the 20 mg/kg/day cannabidiol group compared with 11.9% in the placebo group. The OR for achieving a 75% or greater reduction in the frequency of convulsive seizures was 2.21 (95% CI, 0.82 to 5.95; P = 0.1121) in the 20 mg/kg/day cannabidiol group compared with the placebo group. Three patients (4.9%) in the 20 mg/kg/day cannabidiol group experienced convulsive-seizure freedom (i.e., a 100% reduction in convulsive-seizure frequency) compared with no patients in the placebo group (treatment difference of 4.9%; 95% CI, –0.5 to 10.3; P = 0.0827).

In the CARE2 study, the proportion of patients with a 75% or greater reduction in the frequency of convulsive seizures was 17.9% in the 20 mg/kg/day cannabidiol group, 30.3% in the 10 mg/kg/day cannabidiol group, and 6.2% in the placebo group. The OR for a 75% or greater reduction in the frequency of convulsive seizures was 6.63 (95% CI, 2.12 to 20.73; P = 0.0004) in the 10 mg/kg/day cannabidiol group and 3.33 (95% CI, 1.01, 10.92; P = 0.0468) in the 20 mg/kg/day cannabidiol group compared with placebo. Five patients in the cannabidiol groups reached convulsive-seizure freedom (2 patients taking 10 mg/kg/day and 3 patients taking 20 mg/kg/day cannabidiol) compared with 1 patient in the placebo group, with an OR of 2.00 (95% CI, 0.18 to 22.61; P = 0.5600) for 10 mg/kg/day cannabidiol versus placebo and an OR of 3.00 (95% CI, 0.30 to 29.61; P = 0.3640) for 20 mg/kg/day cannabidiol versus placebo.

In the CARE2 study, the proportion of patients who experienced any improvement in the frequency of convulsive seizures was 75.8% in the 10 mg/kg/day and 82.1% in the 20 mg/kg/day cannabidiol groups, while it was 73.8% in the placebo group.

Number of Convulsive Seizure–Free Days

In the CARE2 study, the mean change from baseline in the number of convulsive seizure–free days during the treatment period was 3.9 days (standard deviation [SD] = 4.8) and 2.7 days (SD = 4.0) in the 10 mg/kg/day and 20 mg/kg/day cannabidiol groups, respectively, compared with 1.7 (SD = 4.1) in the placebo group. The mean difference in the change from baseline was 2.4 days (95% CI, 1.0 to 3.9; P = 0.0009) between the 10 mg/kg/day group and placebo, and 1.3 days (95% CI, –0.1 to 2.8; P = 0.0683) between the 20 mg/kg/day group and placebo. This outcome was not reported in the CARE1 Part B study.

Percentage Change From Baseline in Total Seizure Frequency During the Treatment Period

In the CARE1 Part B study, the median percentage change in total seizure frequency in the 20 mg/kg/day cannabidiol group was –28.6% (95% CI, –70.4% to –4.0%) compared with –9.0% (95% CI, –51.4% to 19.6%) in the placebo group. The median difference between 20 mg/kg/day cannabidiol and placebo was –19.2% (95% CI, –39.3% to –1.2%; P = 0.0335).

In the CARE2 study, the median percentage reduction in the 10 mg/kg/day and 20 mg/kg/day cannabidiol groups was –56.4% (95% CI, 47.8% to 63.6%) and –47.3% (95% CI, 36.9% to 56.0%), respectively, compared with –29.7% (95% CI, 16.0% to 41.1%) in the placebo group. The treatment ratios were 0.620 (95% CI, 0.481 to 0.799; P = 0.0003) for 10 mg/kg/day cannabidiol versus placebo and 0.749 (95% CI, 0.581 to 0.965; P = 0.0255) for 20 mg/kg/day cannabidiol versus placebo.

Percentage Change From Baseline in Seizure Frequency by Individual Seizure Type During the Treatment Period

In the CARE1 Part B study, of the individual convulsive-seizure types, tonic-clonic seizures were the most common type reported by patients during the baseline period (89.2% of patients) followed by tonic seizures (24.2%), clonic seizures (20.8%), and atonic seizures (13.3%). Reductions in the frequency of all 4 individual convulsive-seizure types were observed in every treatment group. Greater percentage reductions were seen in the cannabidiol treatment group for all convulsive-seizure types except for clonic seizures, where the placebo group experienced a greater percentage change than the cannabidiol treatment group. Treatment differences were in favour of cannabidiol over placebo for all seizure types except clonic seizures. For tonic-clonic seizures, the treatment difference was –22.0 (95% CI, –42.5 to –2.1; P = 0.0254), which was statistically significant.

In the CARE2 study, of the individual types of convulsive seizures reported by patients during the baseline period, tonic-clonic seizures were the most common (91.4% of all patients), followed by tonic seizures (40.9%), clonic seizures (24.7%), and atonic seizures (15.2%); the distribution of the individual convulsive-seizure types across the 3 treatment groups was similar during the baseline period. Reductions in the frequency of all 4 types of convulsive seizures were observed in every treatment group. In all cases, greater percentage reductions were seen in the 10 mg/kg/day and 20 mg/kg/day cannabidiol groups compared with the placebo groups. In all cases, the treatment ratios for individual convulsive-seizure analyses were numerically in favour of both doses of cannabidiol over placebo but were not statistically significant. Negative binomial regression analyses showed nominal statistical significance in favour of 10 mg/kg/day cannabidiol over placebo for the reduction of atonic seizures, although caution is advised with interpreting the data due to low patient numbers.

In the CARE1 Part B study, of the individual types of nonconvulsive seizures reported by patients during the baseline period, myoclonic and absence seizures were the most common (each type was experienced by 35.8% of patients), followed by countable partial seizures (24.2%) and other seizures (5.8%). Reductions in the frequency of all 4 types of nonconvulsive seizures were seen for both treatment groups, except in the case of countable partial seizures in the placebo group, where there was an increase in seizure frequency. Greater percentage reductions were seen in the cannabidiol treatment group for all types of nonconvulsive seizures, except for absence seizures. None of the treatment differences were statistically significant.

In the CARE2 study, of the individual types of nonconvulsive seizures reported by patients during the baseline period, myoclonic and absence seizures were the most common (42.4% of patients for each), followed by countable partial seizures (34.3%), and other partial seizures (9.1%). Reductions in the frequency of all 4 types of nonconvulsive seizures during the treatment period were observed across the treatment groups. Greater percentage reductions in the 20 mg/kg/day and 10 mg/kg/day cannabidiol groups were seen for myoclonic and absence seizures compared with the placebo group. For countable partial and other partial seizures, the treatment ratios were in favour of placebo over both cannabidiol doses; none of the treatment ratios were nominally statistically significant.

Patients With Status Epilepticus

In the CARE1 Part B study, there were 0 patients and 1 (1.7%) patient with status epilepticus during the baseline period in the 20 mg/kg/day cannabidiol and placebo groups, respectively. During the treatment period, there was 1 patient (1.6%) and 0 patients with status epilepticus in the 20 mg/kg/day cannabidiol and placebo groups, respectively. In the CARE2 study, there were 4 (6.1%) and 6 (90%) patients with status epilepticus during the baseline period in the 10 mg/kg/day and 20 mg/kg/day cannabidiol groups, respectively, compared with 4 patients (6.2%) in the placebo group. During the treatment period, there were 3 (4.5%) and 9 (13.4%) patients with status epilepticus during the baseline period in the 10 mg/kg/day and 20 mg/kg/day cannabidiol groups, respectively, compared with 8 patients (12.3%) in the placebo group. The changes from baseline and between-group differences were not reported in either study.

Health-Related Quality of Life

Change in QOLCE Scores From Baseline

In the CARE1 Part B study, the mean change from baseline in QOLCE scores among patients aged 4 years or older was 5.4 (SD = 14.6) in the 20 mg/kg/day cannabidiol group and 3.8 (SD = 9.9) in the placebo group. The adjusted mean difference in the change from baseline was 1.5 (95% CI, –3.8 to 6.8; P = 0.5766).

In the CARE2 study, the mean changes from baseline in QOLCE score among patients aged 4 years or older were 6.4 (SD = 10.9) and 4.0 (SD = 11.8) in the 10 mg/kg/day and 20 mg/kg/day cannabidiol groups, respectively, compared with 2.6 (SD = 9.7) in the placebo group. The adjusted mean differences in the change from baseline were 3.8 (95% CI, –0.1 to 7.8; P = 0.0581) for the 10 mg/kg/day cannabidiol group and 1.8 (95% CI, –2.2 to 5.8; P = 0.3827) for the 20 mg/kg/day cannabidiol group compared with placebo.

Sleep Disruption and Function

Change in the Sleep Disturbance NRS 0 to 10 Score at Last Visit

In the CARE1 Part B study, the mean difference in the change from baseline was –0.4 (95% CI, –1.5 to 0.7; P = 0.4543).

In the CARE2 study, the mean treatment difference in scores for the NRS 0 to 10 scale between the 10 mg/kg/day cannabidiol group and placebo group was 0.0 (95% CI, –0.9 to 0.8; P = 0.9510); between the 20 mg/kg/day cannabidiol and placebo groups, the between-treatment difference was –0.1 (95% CI, –0.9 to 0.8; P = 0.8407).

Change in ESS Score at Last Visit

In the CARE1 Part B study, the mean treatment difference in ESS score between the 20 mg/kg/day cannabidiol and placebo groups was 1.51 (95% CI, –0.18 to 3.19; P = 0.078).

In the CARE2 study, the mean treatment difference in ESS score between the 10 mg/kg/day cannabidiol and placebo groups was –0.55 (95% CI, –1.86 to 0.75; P = 0.4041) and 0.74 (95% CI, –0.57 to 2.05; P = 0.2676) between the 20 mg/kg/day group and placebo.

Resource Use

Hospitalizations Due to Epilepsy

In the CARE1 Part B study, a total of 5 patients in the 20 mg/kg/day cannabidiol group (8.2%) and 1 patient in the placebo group (1.7%) reported 1 or more inpatients hospitalizations due to epilepsy during the treatment period.

In the CARE2 study, 8 patients (11.9%) in the 20 mg/kg/day cannabidiol group, 12 patients (18.2%) in the 10 mg/kg/day cannabidiol group, and 6 patients (9.2%) in the placebo group reported 1 or more inpatient hospitalizations due to epilepsy during the treatment period. Although reported as epilepsy-related admissions, follow-up determined that a number of admissions were unlikely to be related to epilepsy. Overall, the numbers of patients for whom follow-up confirmed epilepsy-related hospitalizations were similar across treatment groups: 3 patients (4.5%) randomized to 20 mg/kg/day cannabidiol, 7 patients (10.4%) randomized to 10 mg/kg/day cannabidiol, and 4 patients (6.2%) randomized to placebo.

Use of Rescue Medications

In the CARE1 Part B study, 36 (59.0%) and 41 (69.5%) patients in the 20 mg/kg/day cannabidiol and placebo groups, respectively, used rescue medication. In the CARE2 study, the number of patients using rescue medications in the cannabidiol 10 mg/kg/day, 20 mg/kg/day, and placebo groups were 54 (84.4%), 58 (84.1%), and 54 (80.0%), respectively.

Subgroup Effects

When evaluating the subgroup effects of the primary end points (reduction from baseline in the frequency of convulsive seizures and reduction of more than 50% from baseline in the frequency of convulsive seizures) in both studies, the results were consistent with the overall effect in favour of cannabidiol for all variables with a few exceptions:

• In the CARE1 Part B study, no use of clobazam favoured placebo.

• In the CARE2 study, for most subgroups, the difference between treatments was in favour of cannabidiol (both 10 mg/kg/day and 20 mg/kg/day) over placebo. The modifiers with the most credible effects in the 20 mg/kg/day group were region of study and baseline frequency of convulsive seizures. In the 10 mg/kg/day group, only stiripentol use had a modifying effect on the interaction tests (P = 0.046).

Harms

A summary of AEs reported in patients during the CARE1 Part B and CARE2 trials can be found in Table 16. In the CARE1 Part B and CARE2 studies, safety analyses were conducted on the safety analysis set. In the CARE2 study, 2 patients randomized to receive 10 mg/kg/day cannabidiol were titrated above the target dose in error and were included in the 20 mg/kg/day treatment groups in the safety analysis set. One patient in the 10 mg/kg/day cannabidiol group was excluded from the safety analysis set because they were randomized in error and never received IMP.

The first occurrence of an AE was most commonly reported during the 2-week titration (dose escalation) period in each treatment group in both the CARE1 Part B and CARE2 studies. In the CARE1 Part B study, AEs were reported in 39 patients (63.9%) taking cannabidiol and 27 patients (45.8%) receiving placebo. In the CARE2 study, AEs were reported in 28 patients (50.0%) in the 10 mg/kg/day cannabidiol group, 38 patients (61.3%) in the 20 mg/kg/day cannabidiol group, and 33 patients (56.9%) in the pooled placebo group. Somnolence was the most commonly reported AE followed by diarrhea and decreased appetite across both studies. In the CARE1 Part B study, the proportion of patients who received cannabidiol who experienced AEs was 93.4% compared with 74.6% of patients in the placebo group; most AEs (89%) were mild or moderate in severity. In the CARE2 study, 87.5% of patients receiving 10 mg/kg/day cannabidiol, 89.9% of patients receiving 20 mg/kg/day cannabidiol, and 89.2% of patients in the pooled placebo group had 1 or more AEs during the trial; most AEs (92.0%) were of mild or moderate severity.

Adverse Events

In the CARE1 Part B study, 57 of 61 patients (93.4%) in the 20 mg/kg/day cannabidiol group and 44 of 59 patients (74.6%) in the placebo group reported 1 or more AEs. In the CARE2 study, 56 of 64 patients (87.5%) in the 10 mg/kg/day cannabidiol group, 62 of 69 patients (89.9%) in the 20 mg/kg/day cannabidiol group, and 58 of 65 patients (89.2%) in the placebo group reported 1 or more AEs. The most common AEs reported in both studies (> 10% of patients in any treatment group) were somnolence, diarrhea, and decreased appetite.

Serious Adverse Events

The most commonly reported serious AEs (SAEs) in both trials were nervous system disorders. All SAEs were resolved in the CARE1 Part B study, while 3 patients in the 20 mg/kg/day cannabidiol group in the CARE2 study collectively had 3 SAEs that were not resolved at the end of the trial.

In the CARE1 Part B study, 10 of 61 patients (16.4%) in the 20 mg/kg/day cannabidiol group and 3 of 59 patients (5.1%) in the placebo group reported 1 or more SAEs. In the CARE2 study, 13 of 64 patients (20.3%) in the 10 mg/kg/day cannabidiol group, 17 of 69 patients (24.6%) in the 20 mg/kg/day cannabidiol group, and 10 of 65 patients (15.4%) in the placebo group reported 1 or more SAEs. The most common SAEs reported in more than 2% of patients in both studies were status epilepticus, somnolence, and convulsion. Pneumonia was also a common SAE reported in the CARE2 study.

Withdrawals Due to Adverse Events

In the CARE1 Part B study, AEs that led to treatment discontinuation occurred in 9 of 61 patients (14.8%) in the 20 mg/kg/day cannabidiol group and in 1 of 59 patients (1.7%) in the placebo group. In the CARE2 study, 5 of 69 patients (7.2%) in the 20 mg/kg/day cannabidiol group experienced AEs leading to discontinuation from the study. No patients in the 10 mg/kg/day cannabidiol group or placebo groups discontinued treatment due to AEs.

Mortality

No patient deaths occurred during either study.

Notable Harms

Notable harms (AEs of special interest) were not specified in either the CARE1 Part B or CARE2 studies. However, the clinical experts consulted by CADTH suggested the occurrence of somnolence, hypotonia, lethargy, mood disorder, and increased liver enzymes as harms to explore more closely. Of these, only the total number of investigations related to liver function were increased in the 20 mg/kg/day arm of both studies (Table 16).

Table 16: Summary of Harms Results From the CARE1 Part B and CARE2 Studies (Safety Analysis Set)

AE = adverse event, IVRS = interactive voice response system, NR = not reported, SAE = serious adverse event.

^aThe most common AEs included those reported in ≥ 10% of patients from the CARE1 Part B or CARE2 studies.

^bIncidence of status epilepticus reported as an AE by the investigator may differ from the number of events of status epilepticus reported by patient caregivers using the IVRS.

^cThose reported in ≥ 2% of patients in any group in any trial.

Sources: CARE1³⁸ and CARE2³⁹ Clinical Study Reports.

Critical Appraisal

Internal Validity

Both the CARE1 Part B and CARE2 studies are RCTs investigating the efficacy of cannabidiol compared with placebo. They both involved a randomization process that was properly implemented, ensuring an overall balanced distribution of participants to either the cannabidiol or placebo arms, reassured by the balanced distribution of the baseline characteristics among participants in the studies. There were some observed baseline imbalances in both studies. However, these were judged to have a low risk of introducing bias or to have suggested problems in the randomization process. The randomization lists from both studies were also properly generated and the allocation was successfully concealed from investigators and patients. The HRQoL assessment applied only to patients aged 4 years or older; however, since the randomization was stratified by age group, prognostic balance was likely to have been maintained for this subpopulation.

Both studies provided ground for patients to maintain good adherence to the intended interventions, confirmed by the flow of participants throughout the studies showing no important imbalances in the adherence to cannabidiol. There were, however, some imbalances observed in the use of different cointerventions. Although these possible deviations could introduce bias, the impact and direction of the bias on the outcomes of interest are uncertain. Although some modifying effects were observed (i.e., use of stiripentol and clobazam and geographical location), the low number of patients across subgroups in both studies warrants caution for stating any credible effect modification resulting from any of these variables. Further, the subgroup analyses were not adjusted for multiple comparisons, so there is an increased risk of type I error (false-positive results) for statistically significant findings.

There were no instances of meaningful missing outcome data. The low number of patients who were lost to follow-up as well as the low number of changes in doses or withdrawals confirm the completeness of the analysis, thus reducing the risk of bias in this domain. This is in addition to an ITT analysis of those patients with complete information.

In both studies, measurements of the outcomes were appropriate. Convulsive seizures are objective end points to measure by both clinicians and patients. Some nonconvulsive seizures may be challenging to detect and may lead to measurement error; however, the magnitude of this error is difficult to determine. The blinding of participants and clinical investigators that was maintained throughout the studies mitigates potential biases in this domain. Only a small difference in the mean number of patient-days of treatment was noticed in the 20 mg/kg/day cannabidiol group of the CARE2 study due to AEs, which could have opened the possibility of patients and clinicians knowing they were receiving the intervention. Similarly, other AEs such as somnolence and appetite changes may have signalled to patients and caregivers which treatment they were receiving.

The results were reported in accordance with predefined protocols, reducing the likelihood of selective reporting bias. Overall, both studies demonstrated adherence to methodological consistency and minimized risks across all of the domains that were assessed for risk of bias for most outcomes when comparing cannabidiol with placebo.

Several secondary end points depicting statistically significant results lacked multiplicity control (refer to Table 15), carrying a higher risk of false-positives; hence, cautious interpretation is needed due to potential random errors.

External Validity

Overall, patients included in the CARE1 Part B and CARE2 trials had baseline characteristics and prognostic factors that were similar to those encountered in the population of people in Canada with DS, according to the clinical experts consulted by CADTH. There were some concerns of uncertainty on the applicability of the results to adult populations older than 18 years of age, since no patients older than 18 years were included in either trial. According to the clinical experts consulted by CADTH, it is unlikely that the response observed in the CARE1 Part B and CARE2 studies would be different in terms of beneficial effects and possible harms; however, there are no trial data to confirm whether the efficacy and safety of cannabidiol would be equivalent in adults. As such, the generalizability of the data from the CARE1 Part B and CARE2 trials to adults older than 18 years of age is uncertain. Given that the severity and frequency of seizures differs between children and adults, the comparative efficacy of cannabidiol compared with usual care could differ between these groups; hence, any effect remains uncertain.

It is unknown whether the results can be generalized to patients with fewer than 4 seizures per month, since patients with such characteristics were not included in these studies.

The CARE1 Part B and CARE2 trials excluded patients using medicinal cannabis or synthetic cannabinoid-based medications and transitioning to cannabidiol (pharmaceutical). This would be a common situation in Canada, as noted by the drug programs. The clinical experts suggested that while this is an important consideration, it would not affect the generalizability of the results in such patients once they stop using other forms of medicinal cannabis.

The question as to whether cannabidiol is more efficacious than the other treatment available in Canada for patients with DS (i.e., stiripentol) when added to standard of care is still uncertain. There is no head-to-head comparison of cannabidiol against stiripentol. Furthermore, the standard-of-care treatments commonly used in patients with DS vary and make it difficult to assess this question using indirect comparison, since such differences may include issues of inconsistency or intransitivity. With the lack of head-to-head comparisons and the current evidence at hand, it is difficult to draw a strong conclusion on this issue.

GRADE Summary of Findings and Certainty of the Evidence

Methods for Assessing the Certainty of the Evidence

For pivotal studies and RCTs identified in the sponsor’s systematic review, GRADE was used to assess the certainty of the evidence for 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.^43,44

• 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 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 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 the 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.

Results of GRADE Assessments

The GRADE assessments included an evaluation of the main outcomes considered important by clinicians, patient groups, and stakeholders. The comparisons evaluated in the GRADE assessments of this report were that of cannabidiol 10 mg/kg/day against placebo and cannabidiol 20 mg/kg/day versus placebo. In Table 2 and Table 3, we present the GRADE summary of findings, respectively, for each comparison.

Long-Term Extension Studies

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

Description of Studies

To inform the long-term safety and tolerability of cannabidiol as an adjunctive treatment in children and adults with inadequately controlled DS, a summary of the results of 1 sponsor-conducted OLE study,⁴⁵ GWPCARE5 (hereafter referred to as the CARE5 study), is provided subsequently.

The CARE5 trial was a multicentre, OLE study for patients with DS or LGS who had completed the double-blind, placebo-controlled, clinical studies of cannabidiol (the CARE1, CARE2, CARE3, and CARE4 trials). The objective of the CARE5 OLE study was to evaluate the long-term safety and tolerability and the effect on seizures of cannabidiol as adjunctive treatment in children and adults with inadequately controlled DS. The secondary objectives of the CARE5 OLE were to evaluate the effect of cannabidiol as adjunctive treatment on quality of life, adaptive behaviour, need for hospitalization due to epilepsy, usage of rescue medication, maintenance of seizure frequency reduction and freedom from seizures during the OLE study, frequency of total seizures, change in duration of subtypes of seizures, number of episodes of status epilepticus, growth and development, menstruation cycles (in females), and signals indicating drug abuse liability of cannabidiol. For patients with DS only, the effect on convulsive-seizure frequency, nonconvulsive-seizure frequency, number of participants who were convulsive seizure–free, and responder rate (defined in terms of percentage reduction in convulsive-seizure frequency) were also evaluated. The following information focuses on the patients with DS enrolled in the CARE5 study who had completed the CARE1 and CARE2 studies.

The study consisted of a titration period and a maintenance period followed by a 10-day taper period. During the 10 days of the titration period, all participants titrated up to 10 mg/kg/day to 20 mg/kg/day of cannabidiol. Thereafter, patients entered the maintenance period where the optimal dosage was maintained (although increases or decreases could occur if there was intolerance or seizures were not optimally controlled). Following the end-of treatment visit or withdrawal, the doses of the IMP were down-titrated at home (10% per day for 10 days) until the end-of-taper period visit.

The baseline period was defined as the 28-day period of the original study from screening to randomization, while the treatment period was defined as day 1 to the day of the last dose up to and including the end-of-treatment visit.

Populations

Patients who completed all scheduled visits in the treatment phase of the CARE1 and CARE2, CARE3, or CARE4 studies were included. Those who used cannabis or cannabinoid-based medication, had a history of postural hypotension symptoms, had a history of suicidal behaviour, had been part of a clinical trial involving an IMP, were of child-bearing potential as a patient or partner of the patient, were pregnant, had any disease or abnormalities that put them at risk, were unwilling to abstain from blood donation during the trial, or had impaired hepatic function were excluded.

Interventions

The CARE5 study consisted of a titration period and a maintenance period followed by a 10-day taper period. Participants could receive treatment for up to a maximum of 6 years (312 weeks after visit 1), depending on the country. Cannabidiol was to be taken twice daily (morning and evening) immediately after the patient’s usual ASM administration or as otherwise specified by the investigator. All patients were titrated up to 10 mg/kg/day to 20 mg/kg/day of cannabidiol. If there was intolerance during titration, the patient was maintained on a dose below 10 mg/kg/day to 20 mg/kg/day.

Patients were titrated up to 10 mg/kg/day to 20 mg/kg/day and then remained at this dose until the end-of-treatment visit, with the option for doses to be increased (up to a maximum of 30 mg/kg/day) or decreased, if deemed necessary by the investigator. Following the end-of-treatment visit, the doses of the IMP were down-titrated at home (10% per day for 10 days) until the end-of-taper period visit.

Outcomes

A summary of outcomes reported in the CARE5 study that were relevant to this review is presented in Table 17.

The safety of cannabidiol was assessed by the AE profile and by evaluating changes in the following, relative to the prerandomization baseline of the original study: vital signs, physical examination (including height and body weight), electrocardiograms (ECGs), Columbia-Suicide Severity Rating Scale (C-SSRS) score, Cannabis Withdrawal Scale score or Pediatric Cannabis Withdrawal Scale score, and clinical laboratory parameters. Participants aged 18 years and older were administered the Cannabis Withdrawal Scale, while participants aged 4 to 17 years (inclusive) were administered the Pediatric Cannabis Withdrawal Scale. The children’s C-SSRS was used for participants aged 6 to 18 years (inclusive), while the C-SSRS was used for participants aged 19 years and older.

Baseline Characteristics

The baseline demographic and clinical characteristics of the CARE5 OLE study for patients with DS (safety analysis set) are summarized in Table 18.

Table 17: Summary of Outcomes in the CARE5 Study

AE = adverse event; CWS = Cannabis Withdrawal Scale; C-SSRS = Columbia-Suicide Severity Rating Scale; ECG = electrocardiogram; IGF-1 = insulin-like growth factor; PCWS = Pediatric Cannabis Withdrawal Scale; QOLCE = Quality of Life in Childhood Epilepsy; QOLIE = Quality of Life in Epilepsy.

^aSince the CARE1 and CARE2 studies only enrolled patients aged 2 to 18 years, no patients with Dravet syndrome were eligible to complete the QOLIE questionnaire.

Source: CARE5 Clinical Study Report.⁴⁵

Table 18: Summary of Baseline Characteristics in the CARE5 Study (Safety Analysis Set)

ASM = antiseizure medication; BMI = body mass index; SD = standard deviation.

Source: CARE5 Clinical Study Report.⁴⁵

Statistical Analysis

All data collected during this study were summarized across time. Changes from baseline were also presented where data from the original studies were available. This included seizure information, C-SSRS assessments, quality-of-life assessments, and other measures of safety (e.g., vital signs, clinical laboratory samples). Descriptive statistical methods were used throughout. There was no formal hypothesis testing.

The safety analysis set is the only analysis set in this study and included all patients who received at least 1 dose of IMP in the study.

For the seizure outcomes, when a patient withdrew after they had taken their first dose in the OLE, seizure frequencies were calculated from the data available before the patient withdrawing. The last observation carried forward methodology was used to impute missing data.

Results

Patient Disposition

Patient disposition for the CARE5 OLE study (safety analysis set) is summarized in Table 19.

Exposure to Study Treatments

Exposure information for the CARE5 OLE study (safety analysis set) is described in Table 20.

Table 19: Patient Disposition (Safety Analysis Set)

AE = adverse event; DS = Dravet syndrome.

^aOnly patients with DS were included in these counts.

Source: CARE5 Clinical Study Report.⁴⁵

Table 20: Patient Exposure in the CARE5 Study (Safety Analysis Set)

NR = not reported; SD = standard deviation.

^aThe study tracked adherence to study intervention rather than adherence. The CARE5 Clinical Study Report contains individual patient-level data on compliance and/or drug concentration data.

Source: CARE5 Clinical Study Report.⁴⁵

All 315 patients with DS (100%) took 1 or more concomitant ASMs during the study. The most common classes of ASMs (i.e., reported in > 50% of patients overall) used concomitantly by patients with DS were other antiepileptics, which were used by 266 participants (84.4%); benzodiazepine derivatives, which were used by 254 participants (80.6%); and fatty acid derivatives, which were used by 218 participants (69.2%). The most commonly used ASMs by patients with DS were valproic acid, used by 218 participants (69.2%); clobazam, used by 215 participants (68.3%); stiripentol, used by 120 participants (38.1%); and levetiracetam, used by 92 participants (29.2%).

Efficacy

The efficacy values of the CARE5 study are presented in Table 21.

Table 21: Summary of Key Efficacy Outcomes in the CARE5 Study (Safety Analysis Set)

Q1 = lower quartile; Q3 = upper quartile; QOLCE = Quality of Life in Childhood Epilepsy; SD = standard deviation.

^aEfficacy results for time points up to weeks 253 to 264 are available in the CARE5 Clinical Study Report.

^bCombined caregiver and patient scores are reported.

^cQuestionnaire data reported for patients aged 2 to 18 years.

Source: CARE5 Clinical Study Report.⁴⁵

Change in the Frequency of Total Seizures

During weeks 37 to 48 of treatment, patients with DS experienced a median percentage change from their original study baseline of –62.6% in total seizure frequency (interquartile range, –87.7 to –6.4).

Responder Rates for Total Seizure Frequency

The proportion of patients who achieved a 50% or greater reduction in total seizure frequency during weeks 37 to 48 of treatment was 59.3%. Out of all patients with DS, 70.1% experienced a 25% or greater reduction in total seizure frequency, 39.7% experienced a 75% or greater reduction in total seizure frequency, and 6.1% experienced total seizure freedom.

Change in the Frequency of Convulsive Seizures

During weeks 37 to 48 of treatment, patients with DS experienced a median percentage change from their original study baseline of –54.2% (interquartile range, –83.1 to –10.1) in convulsive-seizure frequency.

Responder Rates for Convulsive-Seizure Frequency

The proportion of patients who achieved a 50% or greater reduction in convulsive-seizure frequency during weeks 37 to 48 of treatment was 52.3%. Out of all patients with DS, 67.8% experienced a 25% or greater reduction in the frequency of convulsive seizures, 34.6% experienced a 75% or greater reduction in the frequency of convulsive seizures, and 7.9% experienced convulsive-seizure freedom (100% reduction).

Change in the Number of Episodes of Status Epilepticus

During the last 12 weeks of treatment, 4.5% of patients with DS reported convulsive seizures greater than 30 minutes in duration compared with 4.8% during the original study baseline. The proportion of patients with DS with nonconvulsive seizures greater than 30 minutes in duration during the last 12 weeks of treatment was 4.8% compared with 7.2% during the original study baseline.

Inpatient Hospitalizations Due to Epilepsy at Last Visit

During the study, 16 patients (6.4%) had 1 or more inpatient hospitalizations due to epilepsy.

Use of Rescue Medications

During the study, 258 patients (70.5%) were recorded as taking rescue medications.

Harms

A total of 306 patients with DS (97.1%) had 1 or more AEs during the study, with 71 patients (22.5%) reporting AEs of mild severity, 157 patients (49.8%) reporting AEs of moderate severity, and 78 patients (24.8%) reporting severe AEs. SAEs were reported for 133 patients (42.2%) with DS, with the most common SAEs experienced being status epilepticus, convulsion, and pneumonia. Twenty-eight patients (8.9%) with DS stopped treatment due to AEs. The most common AEs leading to discontinuation were convulsion, increased AST, and increased ALT. A total of 6 patients (1.9%) with DS died during the study, 4 (1.3%) due to SUDEP, and 1 (0.3) each due to drowning and convulsion (Table 22).

Table 22: Summary of Harms Results From CARE5 (Safety Analysis Set)

SAE = serious adverse event.

^aMost common adverse events included those reported in > 10% of patients in any treatment group from the CARE5 study.

^bMost common SAEs included those reported in ≥ 2% of patients in any treatment group from the CARE5 study.

^cMost common adverse events leading to discontinuation of the investigational medicinal product included those reported in > 1% of patients in any treatment group from the CARE5 study.

Sources: Details included in the table are from the sponsor’s summary of clinical evidence and the CARE5 Clinical Study Report.⁴⁵

Critical Appraisal

Internal Validity

The CARE5 study is a nonrandomized, open-label, single-arm extension study. The lack of a randomized comparator precludes the ability to draw strong conclusions regarding the long-term therapeutic benefits or safety of cannabidiol as an adjunctive therapy among patients with DS compared with other relevant therapies. It is also not possible to ascertain a causal relationship between cannabidiol and the observed effects.

Furthermore, the lack of blinding in the CARE5 study may affect the reporting of subjective measures such as patient-reported outcomes; the direction and magnitude of this potential bias cannot be ascertained. Moreover, only 54% of patients completed the study, so there is a risk of bias due to missing outcomes data. The median number of dosing days in the treatment phase for all participants was 444 days (range, 18 days to 1,822 days), which provided a longer follow-up time for assessment of AEs compared with the double-blind trials; however, the proportion of patients who adhered to the study drug during the longer follow-up was not reported. As such, it is not possible to determine whether there is a risk of bias due to deviations from the intended interventions.

External Validity

Completion of the CARE1 and CARE2 studies was an eligibility criterion for enrolment into the CARE5 study. As such, patients who discontinued from either the CARE1 or CARE2 study for any reason, such as AEs, withdrawal by patient or parent, or withdrawal by investigator, were excluded from the CARE5 study. Therefore, it is likely that the patients who were enrolled in the CARE5 study were those who tolerated cannabidiol better and/or had a more favourable response. The results may therefore not be generalized to all patients with DS who might be eligible for treatment with cannabidiol. Similar to the original trials, there were few patients older than 18 years of age; as such, it is not certain whether the results could be generalized to adults with DS. The outcomes reported included those that were identified as important by patients and clinicians.

Indirect Evidence

An indirect treatment comparison was not included in this submission. The reasoning provided by the sponsor was that usual care was considered the only appropriate comparator in the DS population. Patients with DS often cycle through numerous ASMs in attempts to control their seizures, but many have disease that is refractory to currently available ASMs and display heterogeneity in their response to therapies. The sponsor consulted clinical experts in Canada who confirmed that the usual care followed in the CARE trials is representative of Canadian clinical practice. Additionally, the sponsor formulated that in the real-world setting, cannabidiol will be administered as an add-on treatment. Therefore, the only comparator considered by the sponsor in this population was the established standard of care or clinical management. CADTH recognized that stiripentol is currently indicated for use in combination with clobazam and valproate to treat refractory generalized tonic-clonic seizures in patients with DS whose seizures are not adequately controlled with clobazam and valproate alone. The sponsor considered stiripentol to be a part of usual care; however, the efficacy and cost-effectiveness of cannabidiol compared with stiripentol is unknown due to a lack of published head-to-head comparisons. Indirect comparative evidence from the literature suggests that stiripentol may be associated with a higher seizure control rate than cannabidiol, but the certainty in the effect estimates remains to be determined due to the high heterogeneity in the comparisons assessed, increasing the possibility of intransitive assessments from a network meta-analysis.⁴⁶

Studies Addressing Gaps in the Systematic Review Evidence

No studies addressing gaps in the systematic review evidence were submitted by the sponsor.

Discussion

Summary of Available Evidence

The body of evidence informing this submission consists of 2 individual studies assessing cannabidiol as an adjunctive treatment in patients with DS. First, the pivotal CARE1 Part B study (N = 120 patients) evaluated cannabidiol 20 mg/kg/day against placebo in patients with DS. Second, the pivotal CARE2 study evaluated cannabidiol 20 mg/kg/day (n = 67) and cannabidiol 10 mg/kg/day (n = 67) therapy against a placebo group (n = 65). All patients in these studies were younger than 18 years of age and were receiving multiple therapies for controlling their seizures.

A supplemental body of evidence was submitted in the form of a single-arm, long-term, OLE study (CARE5) to inform the long-term safety and tolerability of cannabidiol as an adjunctive treatment in children and adults with inadequately controlled DS. No evidence from indirect treatment comparisons or studies addressing gaps in the systematic review evidence were submitted by the sponsor.

Interpretation of Results

Efficacy

For patients with DS, the end points of the studies — reduction in convulsive and total seizure frequency — are highly valued outcomes, as are those related to HRQoL, sleep disruption, medication and hospital use, and harms. The clinical experts agreed with these considerations and deemed these end points to be key factors that would inform their decisions in clinical practice.

Comparison of Cannabidiol 10 mg/kg/day Against Placebo

The reduction in the frequency of convulsive seizures (measured as number of seizures per month and then as either the percentage change from baseline or as the proportion of patients achieving a 50% or greater or 75% or greater reduction from baseline) was 1 of the main outcomes deemed relevant for decision-making. The CARE2 study (N = 131 patients) provided information on the 10 mg/kg/day dosage of cannabidiol. In this study, patients in the placebo arm had a 24.5% reduction in the frequency of seizures per month; in contrast, patients taking cannabidiol at 10 mg/kg/day had a reduction from baseline of 41.2%. This means that cannabidiol 10 mg/kg/day provides a 15.7% greater reduction than placebo. However, due to chance, this difference in reduction can go from a reduction of 31.3% to an increase of 3.7% (the 95% CI); this implies that the possible values still include a trivial effect.

Similarly, when evaluating the proportion of patients reaching a 50% or greater reduction from baseline in convulsive seizures, an improvement favouring cannabidiol was observed, although still with some uncertainty as to whether this effect can be clinically important due to imprecision in the effect estimate. For instance, in the placebo arm, 26.2% of patients reached the 50% or greater reduction threshold while, in the cannabidiol 10 mg/kg/day group, 43.9% of the patients reached this threshold (OR = 2.21; 95% CI, 1.06 to 4.62). Expressed as absolute effects, this means that 178 more patients per 1,000 treated with cannabidiol 10 mg/kg/day reached this threshold when compared with placebo; however, due to chance (the 95% CI), this number can go from 17 more patients per 1,000 (likely a trivial effect) to 338 more patients per 1,000 (a large effect). Since no MID for between-group difference could be obtained, the clinical experts consulted by CADTH considered that a difference between cannabidiol and placebo of at least 20 more patients per 1,000 achieving a 50% or greater reduction from baseline in convulsive seizures was considered a conservative threshold of meaningful clinical importance. Cannabidiol 10 mg/kg/day also had effects in increasing the control of convulsive seizures by 75% or more from baseline (deemed of higher certainty), and in increasing the number of convulsive seizure–free days by 2.4 days when compared with placebo. However, little to no difference was observed in the outcome of change in number of episodes of convulsive status epilepticus.

HRQoL was evaluated in 110 patients in the 10 mg/kg/day group of the CARE2 study and showed that cannabidiol provided improvement in the HRQoL measure, resulting in a QOLCE score that was 3.8 points higher compared with placebo, but the 95% CI was imprecise (0.1 lower to 7.8 higher), including the null effect. The clinical meaningfulness of this effect is uncertain also because no threshold of clinical benefit could be obtained.

Similarly, the ESS and the NRS 0 to 10 scores indicated the study did not provide evidence of differences between the cannabidiol and placebo arms in the sleep disruption outcome, an outcome considered relevant by patients and clinicians. Like HRQoL, thresholds of clinical importance were not possible to obtain; hence, the overall certainty for these end points was assessed based on any effect that would exclude the null. As a result, these end points were considered of low certainty due to imprecision.

The utilization of resources in the form of hospital and medication use was not meaningfully different between treatment with cannabidiol and placebo. The body of evidence also had uncertainty due to low sample size and no description of a specific effect estimate from the pivotal studies in both outcomes of resource utilization.

Comparison of Cannabidiol 20 mg/kg/day Against Placebo

Both the CARE1 Part B and CARE2 pivotal studies informed this comparison. Overall, the results were in agreement with the 10 mg/kg/day dose comparison against placebo, but with higher certainty on the effect estimates.

Reduction in convulsive-seizure frequency (measured as either the percentage change from baseline or as the proportion of patients achieving a 50% or greater or 75% or greater reduction from baseline) was the main outcome deemed relevant by experts for decision-making. Evidence from the 2 studies (N = 252 patients in total) showed that, on average, cannabidiol produces greater reductions from baseline in the frequency of seizures when compared with placebo, with cannabidiol 20 mg/kg/day providing 22.8% and 19.9% greater reductions from baseline than placebo in the CARE1 Part B and CARE2 studies, respectively.

Similarly, the proportion of patients reaching a 50% or greater reduction in seizures from baseline was higher in the intervention groups compared with the placebo group (CARE1 Part B study: OR = 2.0; 95% CI, 0.93 to 4.30; CARE2 study: OR = 2.74; 95% CI, 1.32 to 5.70). Compared with placebo, the number of patients who reached this threshold was 155 higher per 1,000 patients (from 13 fewer to 323 more patients) in the CARE1 Part B study and 231 more patients per 1,000 (from 71 more to 391 more patients) in the CARE2 study.

Similar effects were observed when assessing the outcome of a 75% or greater reduction in convulsive seizures from baseline, with an OR of 2.21 (95% CI, 0.82 to 5.95) in the CARE1 Part B study and an OR of 3.33 (75% CI, 1.01 to 10.92) in the CARE2 study. For the number of convulsive seizure–free days, total seizure frequency, and instances of convulsive status epilepticus, effect estimates in favour of cannabidiol were observed; however, the evidence was considered less certain due to imprecision in the effect estimates.

HRQoL was evaluated in 193 patients from both studies using the QOLCE score, where higher values imply better HRQoL. This was deemed a relevant outcome for clinicians and showed that cannabidiol provides a greater improvement in QOLCE score. However, the effect was considered imprecise due to a wide 95% CI ranging from a reduction to an increase in the scale points. The clinical meaningfulness of this effect is uncertain also because no MID or threshold of clinical benefit could be obtained.

The CARE1 Part B and CARE2 studies did not provide evidence of differences between the cannabidiol and placebo arms in sleep disruption outcomes in either the ESS or the NRS 0 to 10 scores, although the overall certainty was deemed low due to imprecision. When evaluating the utilization of resources in both pivotal studies, the rate of hospitalization and medication use was not meaningfully different between treatment and placebo. The body of evidence also had uncertainty due to imprecision (low sample size) from the pivotal studies.

CADTH could not establish enough evidence to support a dose-response effect of cannabidiol for any of the measured outcomes or harms due to the small number of observations and wide CIs. Although different effects between 20 mg/kg/day and 10 mg/kg/day were observed, the information is not sufficient to declare a dose response.

Other Considerations

CADTH recognizes that stiripentol is currently approved in Canada and indicated for use in combination with clobazam and valproate to treat refractory generalized tonic-clonic seizures in patients with DS whose seizures are not adequately controlled with these drugs alone. The sponsor considered stiripentol to be a part of usual care; however, clear evidence on the efficacy and cost-effectiveness of cannabidiol compared with stiripentol is lacking due to the absence of published head-to-head comparisons. Indirect comparative evidence from the literature suggests that stiripentol may be associated with a higher seizure control rate than cannabidiol,⁴⁶ but the certainty in the effect estimates remains to be determined due to the high heterogeneity in the comparisons assessed in the published literature, increasing the possibility of intransitive assessments from a network meta-analysis.

Whether the effect estimates obtained from the body of evidence in this review can be generalized to adults older than 18 years of age remains uncertain, even though data from long-term observational studies and input from the clinical experts consulted by CADTH suggest that adults with DS might achieve comparable, adequate seizure control through the use of cannabidiol.

Harms

Cannabidiol, either as 10 mg/kg/day or 20 mg/kg/day, was generally well tolerated, with some of the AEs being of higher frequency in the 20 mg/kg/day group than in the 10 mg/kg/day and placebo groups. Some exceptions observed in certain events and investigations are noted subsequently.

In the cannabidiol 10 mg/kg/day versus placebo comparison, the proportion of patients with at least 1 AE was similar for the intervention when compared with the placebo group: 56 patients (87.5%) in the intervention arms versus 58 patients (89.2%) in the placebo group. Similarly, little difference was observed in the number of patients with at least 1 SAE, with 13 patients (20.3%) versus 10 patients (15.4%) in the intervention and placebo arms, respectively. A slight increase in the investigations of liver enzymes (ALT and/or AST) was observed in the CARE2 study, with 13 (20.3%) versus 8 (12.3) patients in the intervention and control groups undergoing these evaluations, respectively. According to the experts consulted by CADTH, these measurements are not a critical concern and typically do not necessitate additional assessment beyond routine checks. Also, the number of SAEs was slightly higher in the 10 mg/kg/day group (13 patients; 20.3%) versus placebo (10 patients; 15.4%), but the numbers were considered too small to draw definite conclusions.

In the 20 mg/kg/day versus placebo comparison, the proportion of patients with AEs increased in the CARE1 Part B study, with 57 patients (93.4%) and 44 patients (74.6%) having at least 1 AE in the intervention and placebo arms, respectively; this difference was not present in the CARE2 study. Looking at the most common AEs, more gastrointestinal disorders (diarrhea and vomiting), fatigue, pyrexia, somnolence, and investigations of liver enzymes were observed with the 20 mg/kg/day group versus placebo in both studies. Meanwhile, in total, SAEs were more common in the 20 mg/kg/day group, occurring in 10 patients (16.4%) and 3 patients (5.1%) in the CARE1 Part B study’s intervention and placebo arms, respectively, and in 17 patients (24.6%) and 10 patients (15.4%) in the CARE2 study. The clinical meaningfulness of these differences in AEs and SAEs is uncertain due to the relatively small number of patients assessed.

More harms of special interest occurred in the cannabidiol 20 mg/kg/day group when compared with the 10 mg/kg/day or placebo groups. Somnolence occurred in a total of 7 patients in the cannabidiol 20 mg/kg/day arms of both studies (5 in the CARE1 Part B study and 2 in the CARE2 study) while no events were reported in the placebo arms or in the 10 mg/kg/day arm (CARE2 study). Hypotonia, lethargy, and investigations of liver enzymes also followed this pattern. However, for all events of special interest, the numbers were considered too low to draw definite conclusions.

When comparing the cannabidiol 20 mg/kg/day dose with the 10 mg/kg/day dose and placebo, more discontinuations of the intervention due to AEs were observed with the 20 mg/kg/day dose than with the 10 mg/kg/day dose or placebo. This was also observed for some AEs, such as gastrointestinal disorders and somnolence.

The clinical experts and patient groups identified sudden death as an outcome of special interest to monitor. However, no deaths were reported in any of the pivotal studies. In the long term, open-label, single-arm CARE5 study, with a follow-up of up to 260 weeks, 6 deaths were reported among 315 patients (1.9%), 4 due to SUDEP.

Conclusion

The body of evidence informing the effects of cannabidiol on patients with DS consisted of 2 pivotal RCTs comparing cannabidiol (at doses of 10 mg/kg/day and 20 mg/kg/day) versus placebo, and 1 single-arm OLE study. Both pivotal studies inform the effects of cannabidiol on the outcomes of convulsive and total seizure control, HRQoL, sleep disruption, resource utilization, and AEs. All of these are considered by the clinical experts, patient groups, and stakeholders to be critical outcomes for decision-making.

The evidence shows that the use of cannabidiol, when compared with placebo, results in a greater reduction from baseline in the number of convulsive seizures. A greater proportion of patients using cannabidiol achieved reductions from baseline of 50% or more in the frequency of convulsive seizures, and 75% or more compared with placebo. Cannabidiol also likely provides beneficial albeit smaller improvement effects in the number of convulsive seizure–free days and in total seizure frequency. There is still uncertainty in the effects of cannabidiol on status epilepticus, HRQoL, sleep disruption, and use of rescue medications and hospital use.

Cannabidiol was generally well tolerated. However, the evidence suggests that cannabidiol may have more AEs and SAEs when compared with placebo, although there is uncertainty on the clinical significance of these differences to inform decision-making. The performance of cannabidiol against the other comparator used in Canada for patients with DS (i.e., stiripentol) is still uncertain, as no head-to-head comparison is available. Similarly, there is no evidence on the effects of cannabidiol as a first-line adjunctive treatment in patients older than 18 years of age, nor for its use in patients with DS who experience fewer seizures (i.e., fewer than 4 seizures per month).

Overall, the use of cannabidiol yields better estimates of seizure control with an adequate safety profile when compared with placebo. The effects on long-term outcomes of behaviour and HRQoL are still uncertain.

References

1.Wirrell EC, Hood V, Knupp KG, et al. International consensus on diagnosis and management of Dravet syndrome. Epilepsia. 2022;63(7):1761-1777. PubMed

2.Skluzacek JV, Watts KP, Parsy O, Wical B, Camfield P. Dravet syndrome and parent associations: the IDEA League experience with comorbid conditions, mortality, management, adaptation, and grief. Epilepsia. 2011;52 Suppl 2:95-101. PubMed

3.Shmuely S, Sisodiya SM, Gunning WB, Sander JW, Thijs RD. Mortality in Dravet syndrome: a review. Epilepsy Behav. 2016;64(Pt A):69-74.

4.Wirrell EC, Laux L, Donner E, et al. Optimizing the diagnosis and management of Dravet syndrome: recommendations from a North American consensus panel. Pediatr Neurol. 2017;68:18-34.e13. PubMed

5.Rosander C, Hallböök T. Dravet syndrome in Sweden: a population-based study. Dev Med Child Neurol. 2015;57(7):628-633. PubMed

6.Dravet Canada. What is Dravet? https://dravet.ca/what-is-dravet/. Accessed 2023 Jul 19.

7.Orpha.net. Dravet syndrome. 2021; https://www.orpha.net/en/disease/detail/33069#:~:text=Disease%20definition,with%20cognitive%20and%20motor%20impairment. Accessed 2023 Jul 11.

8.Dravet Canada. The core Dravet syndrome phenotype. Epilepsia. 2011;52 Suppl 2:3-9. PubMed

9.Ruthirago D, Julayanont P, Karukote A, Shehabeldin M, Nugent K. Sudden unexpected death in epilepsy: ongoing challenges in finding mechanisms and prevention. Int J Neurosci. 2018;128(11):1052-1060. PubMed

10.Auvin S, Damera V, Martin M, Holland R, Simontacchi K, Saich A. The impact of seizure frequency on quality of life in patients with Lennox-Gastaut syndrome or Dravet syndrome. Epilepsy Behav. 2021;123:108239. PubMed

11.Juandó-Prats C, James E, Bilder DA, et al. DRAVET ENGAGE. Parent caregivers of children with Dravet syndrome: perspectives, needs, and opportunities for clinical research. Epilepsy Behav. 2021;122:108198. PubMed

12.Brunklaus A, Ellis R, Reavey E, Forbes GH, Zuberi SM. Prognostic, clinical and demographic features in SCN1A mutation-positive Dravet syndrome. Brain. 2012;135(Pt 8):2329-2336. PubMed

13.Cross JH, Caraballo RH, Nabbout R, Vigevano F, Guerrini R, Lagae L. Dravet syndrome: treatment options and management of prolonged seizures. Epilepsia. 2019;60 Suppl 3:S39-s48. PubMed

14.Andrade DM, Berg AT, Hood V, et al. Dravet syndrome: a quick transition guide for the adult neurologist. Epilepsy Res. 2021;177:106743. PubMed

15.Schubert-Bast S, Kay L, Simon A, et al. Epidemiology, healthcare resource use, and mortality in patients with probable Dravet syndrome: a population-based study on German health insurance data. Epilepsy Behav. 2022;126:108442. PubMed

16.Strzelczyk A, Schubert-Bast S. Therapeutic advances in Dravet syndrome: a targeted literature review. Expert Rev Neurother. 2020;20(10):1065-1079. PubMed

17.Ontario Epilepsy Guidelines. Clinical guidelines for the management of epilepsy in adults and children, version 2.0. Toronto (ON): The Epilepsy Research Program of the Ontario Brain Institute; 2020: https://clinictocommunity.ca/wp-content/uploads/2021/01/ManagementGuidelines_Nov2020.pdf. Accessed 2023 May 08.

18.Strzelczyk A, Schubert-Bast S. A practical guide to the treatment of Dravet syndrome with anti-seizure medication. CNS Drugs. 2022;36(3):217-237. PubMed

19.Cardenal-Muñoz E, Auvin S, Villanueva V, et al. Guidance on Dravet syndrome from infant to adult care: road map for treatment planning in Europe. Epilepsia Open. 2022;7(1):11-26. PubMed

20.Jazz Pharmaceuticals Inc. Data on File. Canadian Clinical Expert Survey. 2023 [sponsor provided reference].

21.Diacomit (stiripentol): 250 mg and 500 mg oral capsules, or 250 mg and 500 mg oral powder for suspension [product monograph]. Newmarket (ON): Biocodex Canada Inc.; 2021 Jul 14: https://pdf.hres.ca/dpd_pm/00062850.PDF. Accessed 2023 Nov 20.

22.Health Canada. Summary basis of decision (SBD) for Diacomit. 2021; https://hpr-rps.hres.ca/reg-content/summary-basis-decision-detailTwo.php?linkID=SBD00181. Accessed 2023 Jul 19.

23.National Institute for Health and Care Excellence. Epilepsies in children, young people and adults. (NICE guideline NG217) 2022; https://www.nice.org.uk/guidance/ng217. Accessed 2023 Dec 18.

24.Epidiolex (cannabidiol): 100 mg/mL, oral solution [product monograph]. Cambridge (UK): GW Research Ltd.; 2023 Nov 15: https://pdf.hres.ca/dpd_pm/00073344.PDF. Accessed 2023 Nov 20.

25.Apo-Valproic Acid (valpric acid capsules): 250 mg oral capsules; Valproic Acid Oral Solution (valproic acid oral solution): 250 mg/5 mL valproic acid (as sodium valproate), oral solution [product monograph]. Toronto (ON): Apotex Inc.; 2023 Oct 05: https://pdf.hres.ca/dpd_pm/00072687.PDF. Accessed 2023 Oct 25.

26.Apo-Clonazepam (clonazepam): 10 mg, oral tablets [product monograph]. Toronto (ON): Apotex Inc.; 2021 May 06: https://pdf.hres.ca/dpd_pm/00060946.PDF. Accessed 2023 Oct 25.

27.Ach-Levetiracetam (levetiracetam): 250 mg, 500 mg, and 750 mg oral tablets [product monograph]. Kirkland (QC): Accord Healthcare Inc.; 2023 Aug 10: https://pdf.hres.ca/dpd_pm/00072037.PDF. Accessed 2023 Oct 25.

28.Ach-Topiramate (topiramate): 25 mg, 100 mg, and 200 mg oral tablets [product monograph]. Kirkland (QC): Accord Healthcare Inc.; 2022 Dec 12: https://pdf.hres.ca/dpd_pm/00068700.PDF. Accessed 2023 Oct 25.

29.Apo-Clonazepam (clonazepam tablets): 0.5 mg and 2 mg, oral tablets [product monograph]. Toronto (ON): Apotex Inc.; 2023 Sep 15: https://pdf.hres.ca/dpd_pm/00072497.PDF. Accessed 2023 Oct 25.

30.Banzel (rufinamide): 100 mg, 200 mg and 400 mg, oral tablets [product monograph]. Mississauga (ON): Eisai Limited; 2020 Oct 14: https://pdf.hres.ca/dpd_pm/00058487.PDF. Accessed 2023 Oct 25.

31.Cross H, Vyas K, Sahebkar F, Schreiber A, Mazurkiewicz-Beldzinska A. Abstract 137: Cannabidiol (CBD) in patients with Lennox-Gastaut syndrome (LGS) and Dravet syndrome (DS) on a ketogenic diet therapy (KDT) in four phase 3 trials. Virtual Congress: 34th International Epilepsy Congress; 2021 Aug 28 - Sep 01.

32.Devinsky O, Cross JH, Laux L, et al. Trial of cannabidiol for drug-resistant seizures in the Dravet syndrome. N Engl J Med. 2017;376(21):2011-2020. PubMed

33.Madan Cohen J, Checketts D, Dunayevich E, et al. Time to onset of cannabidiol treatment effects in Dravet syndrome: analysis from two randomized controlled trials. Epilepsia. 2021;62(9):2218-2227. PubMed

34.Mazurkiewicz-Beldzinska M, Privitera M, Marsh E, et al. Time to onset of efficacy of cannabidiol (CBD) during titration in patients with Lennox-Gastaut syndrome and Dravet syndrome enrolled in three randomized controlled trials (S48.009). Neurology. 2019;92:S48.009.

35.Nabbout R, Bhathal H, Zolnowska M, et al. Abstract 245: Diarrhoea in patients with Lennox-Gastaut syndrome (LGS) or Dravet syndrome (DS) enrolled in four phase 3 clinical trials of cannabidiol (CBD). Virtual Congress: 34th International Epilepsy Congress; 2021 Aug 28 - Sep 01.