CADTH Reimbursement Review

Pitolisant Hydrochloride (Wakix)

Sponsor: Paladin Labs Inc.

Therapeutic area: Narcolepsy

This multi-part report includes:

Clinical Review

Pharmacoeconomic Review

Stakeholder Input

Clinical Review

Executive Summary

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

Introduction

Narcolepsy is a chronic neurologic condition that is caused by an imbalanced sleep-wake cycle or sleep-wake instability.¹ It is characterized by chronic, excessive episodes of drowsiness during the day, also known as excessive daytime sleepiness (EDS).² Type 1 narcolepsy is classified as EDS with cataplexy, whereas type 2 narcolepsy consists of EDS alone.¹ Cataplexy is defined as a sudden episode of partial or complete paralysis of voluntary muscles, triggered by strong emotion.³ Approximately 60% to 70% of patients with narcolepsy have cataplexy (type 1 disease).⁴ Approximately 1 in 2,000 individuals in Canada are affected by narcolepsy.² This prevalence is considered an underestimate, given the possibility of misdiagnosis and the limited availability of health care providers with experience in narcolepsy.^5-7

Narcolepsy can affect all aspects of life in work and social settings, and a patient’s day-to-day functioning, health-related quality of life (HRQoL), and productivity.⁸ Patients can experience EDS during common situations in the day, such as work or driving, and during sedentary periods.³ Narcolepsy is associated with an increased risk in comorbid conditions, including depression, anxiety, obesity, cardiovascular disease, and overall mortality.⁸ In Canada, the current treatment standard for EDS in narcolepsy is modafinil, which is thought to improve wakefulness by reducing dopamine reuptake.

Pitolisant hydrochloride is an inverse agonist/antagonist of the histamine 3 (H₃) receptor. The human H₃ receptor functions as a presynaptic autoreceptor on histamine-containing neurons.⁹ H₃ antagonists promote wakefulness by increasing histamine synthesis and release. By binding competitively to H₃ autoreceptors on presynaptic histaminergic neurons, pitolisant hydrochloride blocks the normal negative-feedback mechanisms for histamine release, increasing histaminergic transmission and resulting in enhanced histamine synthesis and release.^10-12 Pitolisant hydrochloride is administered orally, up to 40 mg daily, with 5 mg and 20 mg tablets. It is indicated for the treatment of EDS or cataplexy in adults with narcolepsy. It received Notice of Compliance on May 25, 2021, after undergoing standard review. The reimbursement request is per indication.

The objective of this clinical review is to perform a systematic review of the beneficial and harmful effects of pitolisant hydrochloride oral tablets (5 mg and 20 mg), with a daily dose up to 40 mg, for the treatment of EDS or cataplexy in adults with narcolepsy.

Stakeholder Perspectives

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

Patient Input

One patient group, Wake Up Narcolepsy (WUN), submitted patient input for this review. WUN is a patient advocacy nonprofit organization established in 2008 that aims to accelerate research, increase awareness of narcolepsy, and provide supportive services to patients. The input was based on a survey of 19 patients in Canada who have a narcolepsy diagnosis or are undiagnosed but living with narcolepsy symptoms. Most patients were aged 18 to 34 years (66%) and female (72%), and none had experience with the treatment under review.

Respondents reported EDS to be the most troubling symptom of narcolepsy, with 39% of respondents giving it a rating of 6 on a scale of 1 (not at all bothersome) to 7 (completely bothersome). The second-most troublesome symptom reported was disturbed nocturnal sleep (DNS), followed by hallucinations when falling asleep or waking up, cataplexy, and sleep paralysis. The negative impacts of narcolepsy on respondents’ lives include mental health and emotional symptoms (mood swings, anger, depression, and anxiety), missing out on social activities, difficulty managing career and job tasks, depending on others for support for daily activities, and difficulty maintaining physical health and wellness (weight gain). Treatments that respondents reported currently using for their narcolepsy include stimulants (56%), antidepressants (33%), sodium oxybate (13%), and modafinil or armodafinil (13%). Some respondents reported that the physical side effects (28%) and mental side effects (39%) of their current treatment options were moderately or extremely challenging.

Respondents would like a new drug or treatment to be more effective for symptoms of sleepiness, cataplexy, and DNS. Respondents indicated a desire to have a treatment that is easy to swallow and does not cause nausea, weight gain, or affect mood or personality. Respondents want a treatment with an extended release that allows them to stay awake longer during the day without having to take additional doses. A copy of the patient input is presented in Appendix 1.

Clinician Input

Input From Clinical Experts Consulted by CADTH

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

A number of factors make the diagnosis extremely challenging. Patients often first visit family doctors or pediatricians who may not immediately recognize this condition. Patients are misdiagnosed frequently, and more than 70% of patients with narcolepsy are undiagnosed. Although existing medications treat the underlying symptoms of narcolepsy, including primarily daytime sleepiness and cataplexy, it is believed that none of these treatment options address the fundamental underlying neurochemical abnormality of loss of hypocretin cells and secondary absence or reduction of available central nervous system (CNS) hypocretin associated with narcolepsy.

Several problems persist with existing treatment options. Not all patients respond to treatment with selective serotonin reuptake inhibitors (SSRIs), and selective norepinephrine reuptake inhibitors (SNRIs), or tricyclic antidepressants (TCAs), and some become tolerant to treatment. Tolerance to the rapid eye movement (REM)-suppressing effects of SSRI, SNRI, and TCA medications occurs frequently, leading to persistent cataplexy. Side effects of these drugs, such as stomach upset, night sweats, sexual side effects, and headaches, can be problematic, and can include sedation during the day, which can be a problem despite the anticataplectic effects of drugs. With stimulants, daytime sleepiness may not be fully resolved and/or drugs may or may not wear off at inopportune times, leading to EDS in the evenings and/or insomnia at night. The side effects of stimulants, such as appetite suppression, anxiety, increased blood pressure, cardiac effects, allergic reaction, reduced seizure threshold, fetal defects, inactivation of birth control, and hair loss, can be problematic. There can be potential for misuse or drug diversion; most patients with narcolepsy have a low likelihood of misusing existing treatment options even though they may require high doses, but for some, there could be temptation to obtain the drug for purposes not intended by the prescriber.

The consistent use of anticataplectic treatments with pitolisant hydrochloride may mask and/or minimize the potential benefits pitolisant hydrochloride might have for cataplexy, and if the benefits of pitolisant hydrochloride are minimal for cataplexy, it would be difficult to assess. In short, it is difficult to properly assess the potential benefits of pitolisant hydrochloride for cataplexy with use of ongoing anticataplectic treatments.

Based on the efficacy of pitolisant hydrochloride shown in early studies, its novel mechanism of action as a H₃ antagonist/inverse agonist, and its relatively favourable side-effect profile, it is likely to become an early treatment option. It received a strong recommendation from the American Academy of Sleep Medicine (AASM) in their most recent (2021) guidelines for the treatment of hypersomnolence disorders.¹² It will be an early drug to consider for the treatment of narcolepsy. It may find a niche as an adjunct treatment to be used in combination with other therapies to boost efficacy, and may also become a drug of choice for patients in whom stimulant and/or other therapies are contraindicated because it has no effect on the efficacy of birth control (unlike modafinil) and no significant known cardiovascular effects (unlike other stimulants). Patients most in need of intervention include those who cannot tolerate stimulant therapies, those concerned about getting pregnant, and those with a history of drug abuse. Jurisdictions should continue to provide coverage for therapies currently considered standard of care (SOC) when used in combination with pitolisant hydrochloride because the mechanism of action of pitolisant hydrochloride is quite different than any currently available drug, which is an exciting prospect for patients living with this debilitating condition.

Primary outcomes in clinical practice will likely be a reduction in EDS, a reduction in this report, treatment goals are primarily to improve quality of life. Although narcolepsy is not lethal, symptoms of EDS and cataplexy can be debilitating if left uncontrolled. In severe circumstances, sleep attacks can occur while a patient is eating or even talking to someone. Uncontrolled, these symptoms limit a patient’s ability to perform basic daily activities, such as driving, working, and interacting with people. Cataplexy (which occurs in 60% to 70% of patients with narcolepsy) is equally, if not more, debilitating when left uncontrolled. Patients cannot drive or walk outside safely because surprises can trigger a cataplectic attack. Basic daily activities, such as showering and bathing, dressing, and eating, can be dangerous and/or challenging for an untreated patient. Without treatment, most patients have very limited, if any, work options, and may not be able to attend school. The symptoms can lead to isolation, anxiety, and depression. Treatment is aimed at reducing EDS and cataplexy so that patients are not dependent on caregivers for support and can interact and be functional members of society. Treatment can significantly improve alertness and daytime abilities, allowing patients to be functional members of society. Diagnosis is often delayed, often occurring 10 years or longer after symptom onset, potentially leading to significant suffering. However, if appropriate treatment is initiated, tolerated, and maintained, up to 80% of functional capacity could be retained. the frequency, intensity, and duration cataplexy episodes, and the ability to predict episodes. Clinically meaningful responses to treatment include a reduction in the frequency, severity, and intensity of cataplexy episodes. Although frequency is easier to assess systematically, the intensity and severity of spells, as well as the perceived predictability of episodes, are more of a clinical assessment. For example, patients may describe certain emotions that no longer trigger episodes in the way they had before. Other parameters for assessment could include a reduction in other REM intrusion phenomena, if present, and the degree to which patients can resume normal functioning and daily activities.

Outcomes typically assessed in most clinical trials include the degree of reduction in EDS and in the frequency of cataplexy spells. The use of Epworth Sleepiness Scale (ESS) scores in clinical practice to determine coverage of pitolisant hydrochloride may not be ideal. The ESS is very subjective and could easily manipulate scores. In addition, there can be significant differences in the way male and female patients score their results, further skewing potential for the determination of coverage. In research trials, it is ideal if patients are blinded to what they are being offered and there is no incentive to report better or worse scores. An ESS score of 10 or lower would indicate that sleepiness is no different than in the general population. As a comparison, an ESS score for patients with narcolepsy would typically be higher than 18 (on the 24-point scale), indicating severe sleepiness; a score of 15 to 17 indicates moderate sleepiness, and a score of 11 to 14 indicates mild sleepiness.

There is very little data available to define what represents an effective reduction of cataplexy. Trials of sodium oxybate demonstrated a more than 90% reduction in cataplexy episodes. Driving is not recommended if a patient has experienced a cataplexy episode in the previous year. A minimum reduction in cataplexy episodes of 50% would be meaningful. Depending on severity and frequency, fewer than 1 episode per week would be a reasonable standard.

At this time, pitolisant hydrochloride may not be suitable for patients who wish to get pregnant or who are breastfeeding. In addition, because of a lack of data on its use in children and in older adults, it should be used with caution in those populations. Patients who are on multiple medications (particularly drugs affecting the QTc interval or those that are significant CYP2D6 inhibitors, which have more potential for drug interactions) and patients who have a history of significant kidney or liver failure also may not be ideal candidates for pitolisant hydrochloride (because of a difficult-to-predict metabolism). Patients who have had adverse reactions to opioids (including hives), or who have a history of urticaria or another skin condition, might be predisposed to allergic reactions to pitolisant hydrochloride. Ongoing treatment will be determined by lack of response and/or excess adverse side effects, like most medications. Whether it will continue to be used as an adjunct if abandoned as monotherapy is unclear. Excess adverse side effects or drug-drug interactions may necessitate withdrawal. Similarly, if a patient wishes to become pregnant, withdrawal may also be necessary.

As with other drugs for narcolepsy, there should be close follow-up of the patient in the first months of therapy. The first follow-up should occur 1 month after the initiation of pitolisant hydrochloride, then every 1 to 2 months for the next several months, and then intermittent follow-up after that, with at least yearly follow-up, at a minimum, in the long-term. Medical supervision in an outpatient setting with a physician trained in sleep medicine would be appropriate for patients with narcolepsy being treated with pitolisant hydrochloride. In the future, psychiatrists will likely become interested in using this medication for conditions and symptoms outside of narcolepsy. At this time, because the indication for pitolisant hydrochloride is only for narcolepsy, with a conditional recommendation for idiopathic hypersomnia, prescribing should be limited to those with specialty training or certification in sleep medicine.

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.

Clinical Evidence

Pivotal Studies and Protocol-Selected Studies

Description of Studies

Three double-blind, phase III, placebo-controlled, randomized controlled trials (RCTs) met the inclusion criteria for the Systematic Review.^13-15 In all 3 trials, patients were included if they had narcolepsy with cataplexy. The HARMONY 1 (NCT01067222) and HARMONY 1bis (NCT01638403) trials also included patients without cataplexy (type 1 narcolepsy). The HARMONY 1 and HARMONY 1bis trials required patients to have an ESS score of at least 14 during the baseline period, whereas the HARMONY CTP trial required an ESS score of at least 12. The HARMONY CTP (NCT01800045) trial included patients with at least 3 cataplexy attacks weekly. In all trials, patients with severe cataplexy were permitted stable doses of anticataplectic medications (except TCAs) that were administrated for at least 1 month before the start of the trial.

The HARMONY 1 and HARMONY 1bis trials were 8-week trials that assessed the superiority of pitolisant hydrochloride to placebo with regard to EDS in patients with narcolepsy. An additional efficacy objective was a noninferiority comparison between pitolisant hydrochloride and modafinil. The HARMONY CTP trial was a 7-week randomized, double-blind placebo-controlled study comparing pitolisant hydrochloride to placebo. It focused on the safety and efficacy of pitolisant hydrochloride in decreasing the frequency of cataplexy attacks in patients who had narcolepsy with cataplexy. The maximum daily dosages of pitolisant hydrochloride were 20 mg in the HARMONY 1bis trial and 40 mg in the HARMONY 1 and HARMONY CTP trials. Titration of the study drug was at the discretion of study investigators, which could have affected efficacy and potentially threatened blinding to treatment arms. Patients on anticataplectic medications represented 35% of all patients in the HARMONY 1 trial, |||||| of all patients in the HARMONY 1bis trial, and 10% of all patients in the HARMONY CTP trial.

Efficacy Results

Excessive Daytime Sleepiness

In the HARMONY 1 trial, the adjusted mean difference in the final ESS score between pitolisant hydrochloride and placebo was –3.10 (95% confidence interval [CI], –5.73 to –0.46; P = 0.022), as shown in Table 2. Sensitivity analyses of the per-protocol (PP) population, without accounting for the centre effect, showed similar results. Because the superiority of pitolisant hydrochloride over placebo for EDS was demonstrated at the a priori level of significance of alpha = 0.025, the noninferiority of pitolisant hydrochloride to modafinil was tested. The adjusted mean difference in the final ESS score between pitolisant hydrochloride and modafinil was 0.09 (95% CI, –2.31 to 2.30); thus, pitolisant hydrochloride was judged to not be noninferior to modafinil at the prespecified noninferiority margin (NIM) of 2. A patient was considered a responder when the final ESS score was less than 10. Based on this threshold, the responder rates were 13.3% in the placebo group, 45.2% in the pitolisant hydrochloride group, and 45.3% in the modafinil group. The adjusted odds ratio (OR) of response for pitolisant hydrochloride compared with placebo was 7.86 (95% CI, 1.59 to 38.86). The adjusted OR of response for pitolisant hydrochloride compared with modafinil was 1.09 (95% CI, 0.31 to 3.81).

In the HARMONY 1bis trial, the mean ESS score reductions from baseline (standard deviation [SD]) were |||||||||| in the placebo group, |||||||||| in the pitolisant hydrochloride group, and |||||||||| in the modafinil group (Table 2). The adjusted mean difference in the final ESS score between pitolisant hydrochloride and placebo was –2.19 (95% CI, –4.17 to –0.22; P = 0.030). Sensitivity analyses without reallocation by centre, and without adjustment for baseline ESS score, or after adjustment for baseline following the mean change, and the mean change over baseline methods showed similar results. Because the superiority of pitolisant hydrochloride over placebo for EDS was demonstrated at the a priori alpha level of 0.05, the noninferiority of pitolisant hydrochloride to modafinil was tested. The adjusted mean difference in the final ESS score between pitolisant hydrochloride and modafinil was 2.75 (95% CI, 1.02 to 4.48); thus, pitolisant hydrochloride was judged to not be noninferior to modafinil at the prespecified NIM of 2. A patient was considered a responder when the final ESS score was 10 or lower or the change from baseline was at least 3 points. The response proportions were |||||||||||||||||||| for the placebo, pitolisant hydrochloride, and modafinil groups, respectively. The adjusted relative risk (RR) for the difference between pitolisant hydrochloride and placebo was ||||||||||||||||||||. The adjusted RR for the difference between pitolisant hydrochloride and placebo was ||||||||||||||||||||||||||||||.

In the HARMONY CTP trial, the observed mean changes in ESS score over baseline were –1.9 (SD = 4.3) and –5.4 (SD = 4.3) in the placebo and pitolisant hydrochloride arms, respectively (Table 3). The adjusted mean difference in the change from baseline between pitolisant hydrochloride and placebo was –3.42 (95% CI, –4.96 to –1.87). Sensitivity analyses using the last observation carried forward (LOCF), the baseline observation carried forward (BOCF), and the PP population were consistent with the main analysis. A patient was considered a responder when the final ESS score was 10 or lower or the change from baseline was at least 3 points. The response proportions were 34.0% and 68.6% for the placebo and pitolisant hydrochloride groups, respectively. The adjusted OR for the difference between pitolisant hydrochloride and placebo was 4.26 (95% CI, 1.72 to 10.52).

Maintenance of Wakefulness Test

In the HARMONY 1 trial, the adjusted mean difference in final Maintenance of Wakefulness Test (MWT) score between placebo and pitolisant hydrochloride was 1.47 (95% CI, 1.01 to 2.14), and the adjusted mean difference in final MWT score between pitolisant hydrochloride and modafinil was 0.77 (95% CI, 0.52 to 1.13). This was consistent with the findings of the HARMONY 1bis trial, in which the adjusted mean difference between placebo and pitolisant hydrochloride was 1.46 (95% CI, 1.06 to 2.01). The adjusted mean difference in final MWT score between pitolisant hydrochloride and modafinil was ||||||||||||||||||||||||||||||. In the HARMONY CTP trial, the geometric mean of ratios (final divided by baseline) was 1.78 (95% CI, 1.22 to 2.60). Sensitivity analyses for all trials using the PP population were consistent with the main analysis.

Table 3: Sleepiness and Cataplexy — HARMONY CTP (ITT Population) Trial

CI = confidence interval; ESS = Epworth Sleepiness Scale; ITT = intention-to-treat; OR = odds ratio; SD = standard deviation.

^aMean of values at visits 1 and 2.

^bMean of values at visits 5 and 6. For missing values, the LOCF approach was used.

^cLinear mixed model (analysis of covariance) adjusted for baseline ESS and for centre heterogeneity (i.e., including centre as a random factor).

^dP values have not been adjusted for multiple testing (i.e., the type I error rate has not been controlled).

^eMean of week 1 and week 2 values.

^fMean of values during the stable-dose period (i.e., week 6 through week 9). For patients terminating the trial before completion, the final value was calculated as the mean of the 2 last known periods (LOCF).

^gANCOVA via mixed nonlinear model featuring a possibly overdispersed Poisson distribution and taking into account centre heterogeneity by using centre as a random factor.

^hMeasured at visit 2.

ⁱMeasured at visit 6.

^jiNonlinear mixed model taking into account centre heterogeneity.

Source: Clinical Study Report for HARMONY CTP.¹⁵

Sustained Attention to Response Task

In the HARMONY 1 trial, the adjusted mean difference in the Sustained Attention to Response Task (SART) between the pitolisant hydrochloride and placebo treatment arms was 0.82 (95% CI, 0.67 to 0.99) for NOGO scores, 0.80 (95% CI, 0.57 to 1.13) for GO scores, and 0.79 (95% CI, 0.64 to 0.99) for TOTAL scores. The adjusted mean difference between the pitolisant hydrochloride and modafinil treatment arms was 1.03 (95% CI, 0.83 to 1.28) for NOGO scores, 1.03 (95% CI, 0.56 to 1.15) for GO scores, and 0.90 (95% CI, 0.70 to 1.14) for TOTAL scores. Sensitivity analyses using the PP population was consistent with the main analysis. In the HARMONY 1bis trial, the ratio of mean change between pitolisant hydrochloride and placebo was significant (0.83; 95% CI, 0.69 to 0.99), whereas the ratio of mean change between pitolisant hydrochloride and modafinil was ||||||||||||||||||||||||||||||||||||||||||||||||||.

Clinical Global Impression of Change on EDS

In the HARMONY 1 and HARMONY 1bis trials, the Clinical Global Impression of Change (CGI-C) score for EDS was improved in a higher proportion of patients in the pitolisant hydrochloride and modafinil groups than in the placebo group. However, the change in CGI-C score was similar in the pitolisant hydrochloride and modafinil arms. In the HARMONY 1 trial, CGI-C scores for EDS improved in the subgroup of patients with a history of cataplexy, but a greater proportion of patients reported an improvement in the modafinil arm. In the HARMONY CTP trial, the mean reduction in CGI-C score for pitolisant hydrochloride compared with placebo was –0.95 (95% CI, –1.36 to –0.54). Mean CGI-C score was 3.5 (SD = 1.1) in the placebo group and 2.6 (SD = 1.1) in the pitolisant hydrochloride group. Similar results were observed for the PP population, with a mean reduction of –0.86 (95% CI, –1.29 to –0.43).

Frequency and Severity of Cataplexy Attacks

In the HARMONY 1 trial, the final mean of complete and partial cataplexy episodes (per day) was 0.68 (SD = 1.66), 0.28 (SD = 1.11), 0.65 (SD = 1.62) in the placebo, pitolisant hydrochloride, and modafinil groups, respectively. In the exposed population, the RR of daily rates of complete and partial cataplexy episodes at the end of treatment for pitolisant hydrochloride compared to placebo was 0.38 (95% CI, 0.15 to 0.93). The RR of daily rates of complete and partial cataplexy episodes at the end of treatment for pitolisant hydrochloride compared to modafinil was 0.54 (95% CI, 0.24 to 1.23). In the HARMONY 1bis trial, the mean least square of daily cataplexy rate for those with cataplexy between the final 7 days of treatment and baseline was |||||||||||||||||||||||||||||| for pitolisant hydrochloride compared to placebo.

The primary end point in the HARMONY CTP trial was the measure of anticataplectic efficacy. During the stable-dose period, the geometric means of the weekly cataplexy rate (WCR) at the end of treatment decreased to 4.51 (95% CI, 2.90 to 7.02) in the placebo group and 2.27 (95% CI, 1.51 to 3.41) in the pitolisant hydrochloride group. The ratio of geometric means during the stable-dose period was 0.51 (95% CI, 0.43 to 0.60; P < 0.0001) for pitolisant hydrochloride compared to placebo. Similar results were observed for the PP population, with a ratio of 0.50 (95% CI, 0.34 to 0.74; P < 0.0001) for pitolisant hydrochloride compared to placebo. The effect of pitolisant hydrochloride on the WCR remained consistent at 20 mg and 40 mg doses. The proportion of patients with a high frequency of weekly cataplexy episodes (> 15) during the stable-dose period was 5.6% in the pitolisant hydrochloride group and 17.6% in the placebo group (OR, 0.035; 95% CI, 0.0035 to 0.352). The effect remained consistent regardless of whether or not patients were taking permitted anticataplectic medications during the trial.

Clinical Global Impression of Change on Cataplexy

In the HARMONY 1 trial, the mean final CGI-C score was 3.4 (SD = 1.4), 2.9 (SD = 1.5), 3.0 (SD = 1.6) in the placebo, pitolisant hydrochloride, and modafinil arms, respectively. The number of patients who improved compared to baseline was 6 (24.0%) in the placebo group, 9 (34.6%) in the pitolisant hydrochloride group, and 8 (28.6%) in the modafinil group. The number of patients who reported no change compared to baseline was 15 (57.7%) in the placebo group, 15 (57.7%) in the pitolisant hydrochloride group, and 16 (57.1%) in the modafinil group. There were 2 (8.0%) patients reporting worsened CGI-C scores in the placebo arm and 1 (3.6%) in the modafinil arm.

In the HARMONY 1bis trial, the number of patients who improved compared to baseline was |||||||||| in the placebo group, |||||||||| in the pitolisant hydrochloride group, and |||||||||| in the modafinil group. The number of patients who reported no change compared to baseline was |||||||||| in the placebo group, |||||||||| in the pitolisant hydrochloride group, and |||||||||| in the modafinil group. There were |||||||||| patients reporting worsened CGI-C scores in the placebo group, |||||||||| in the pitolisant hydrochloride group, and |||||||||| in the modafinil arm.

In the HARMONY CTP trial, the mean reduction in CGI-C score for pitolisant hydrochloride compared with placebo was –0.95 (95% CI, –1.36 to –0.54). The mean CGI-C score was 3.5 (SD = 1.1) with placebo and 2.6 (SD = 1.1) with pitolisant hydrochloride. Similar results were observed for the PP population, with a mean change of –0.86 (95% CI, –1.29 to –0.43).

Harms

In the HARMONY 1 trial, adverse events (AEs) after initiation of treatment were reported by 66.7% of patients in the placebo group, 64.5% in the pitolisant hydrochloride group, and 69.7% in the modafinil arm. In the HARMONY 1bis trial, approximately |||||| of patients in the pitolisant hydrochloride and modafinil groups reported AEs, as did |||||| of patients in the placebo group. In the HARMONY CTP trial, approximately 35% of patients experienced an AE. For HARMONY 1 ||||||||||||||||||||, there was a greater percentage of nervous system disorders in the pitolisant hydrochloride arm, but in in the HARMONY CTP trial, the placebo arm had more nervous system disorders .

In the HARMONY 1 trial, pyelonephritis and hemorrhoids were reported as serious adverse events (SAEs) in the pitolisant hydrochloride arm. |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||| HARMONY CTP reported 1 SAE in the pitolisant hydrochloride arm only.

In the HARMONY 1 trial, 1 patient in the pitolisant hydrochloride arm discontinued treatment because of pregnancy. Another patient in the pitolisant hydrochloride arm temporarily discontinued the study, but the study code was not broken and treatment was resumed so the study resumed. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||. In the HARMONY CTP trial, 1 patient receiving pitolisant hydrochloride discontinued due to severe nausea, which was characterized as a treatment-emergent adverse event (TEAE). No deaths were reported in any of the trials.

Critical Appraisal

All included trials were double-blinded, placebo-controlled studies with a short duration (7-week or 8-week treatment phase). All trials had a small sample size (between 96 to 164 patients), which can limit the power to detect significant changes in the efficacy outcomes. The allocation sequence was random and balanced for all trials and remained concealed for the duration of the trial. HARMONY 1 |||||||||||||||||||||||||||||| had between-group study differences for previous medication use and for the proportion of patients with cataplexy, which could suggest differences in disease severity. In the HARMONY 1bis trial, |||||||||| of patients had a history of cataplexy in the pitolisant hydrochloride group, as did |||||||||| in the placebo group. In the HARMONY 1 trial, patients taking at least 1 chronic medication in the 3 months before inclusion ranged from 70.0% (modafinil group) to 85.2% (placebo and pitolisant hydrochloride groups). The maximum daily dosage of pitolisant hydrochloride was 20 mg in the HARMONY 1bis trial and 40 mg in the HARMONY 1 and HARMONY CTP trials . Titration of the study drug was at the discretion of the study investigators, which could have had an effect on efficacy and could have potentially threatened blinding to treatment.

All studies authorized patients to remain on stable doses of anticataplectic medications. Patients on anticataplectic medications represented 35% of all patients in the HARMONY 1 trial, |||||||||| of all patients in the HARMONY 1bis trial, and 10% of all patients in the HARMONY CTP trial. In the HARMONY 1 and HARMONY CTP trials, there were between-group differences in the proportion of patients on anticataplectic medications during the trial. In the HARMONY 1 trial, 33.3% of placebo patients remained on authorized medications during the study, compared to 40.7% of patients in the pitolisant hydrochloride group and 56.7% of patients in the modafinil group. In the HARMONY CTP trial, 16% of patients in the placebo group remained on anticataplectic medication, compared with 7% in the pitolisant hydrochloride group. Inconsistency in concomitant anticataplectic medications between trials cannot be clearly explained. The interactions between pitolisant hydrochloride and the concomitant treatments are unknown. Although the trials were double-blinded, some patients who received modafinil previously may have recognized the study drug.

The primary efficacy outcome in the HARMONY 1 and HARMONY 1bis trials, change in EDS, was measured using the validated ESS. ESS is a subjective, self-administered questionnaire widely used in narcolepsy trials. The primary outcome in the HARMONY CTP trial was the weekly rate of cataplexy captured in patient diaries. All primary outcomes were assessed using unvalidated tools. Other secondary end points that assessed EDS were not validated, such as the CGI-C and patient global opinion (PGO) tools. The MWT and SART outcomes were validated, but the statistical analyses did not adjust for multiplicity. Patient diaries were completed daily and reviewed by the investigators for completion, which may have biased future outcome assessments. The primary outcome of the HARMONY CTP trial was the change in WCR, which was recorded in daily patient diaries. A reduction in cataplexy episodes was also reported in the placebo group. This could be related to the use of concomitant treatments or a placebo effect.

Missing values for all trials were imputed for ESS and cataplexy outcomes. Any missing values at the end of treatment were imputed using LOCF or BOCF. It is unclear whether these would be reflective of the true trajectory of the outcomes. Sensitivity analyses using the PP population were provided, which can minimize potential bias. In addition, for all outcomes other than the primary outcome in all trials, there was no adjustment for multiplicity which increases the risk of type I error and limits the ability to draw conclusions. Subgroups were outlined a priori. Conclusions could not be drawn for the subgroups due to the lack of adjustment for multiplicity and were therefore considered exploratory analyses.

The NIM was calculated using historical trials of ESS, which were not specified, that set the minimal important difference (MID) at 3. To remain less than the MID and the proportion of difference between placebo and pitolisant hydrochloride, a NIM of 2 was chosen. In addition, sample-size calculations assumed that the effects of pitolisant hydrochloride and modafinil were similar.

All trials noted protocol amendments. A major amendment in the HARMONY 1 trial included a change from the assessment of superiority of pitolisant hydrochloride over modafinil to a noninferiority analysis. The change in type of analysis would not bias the results because the noninferiority analysis was reported appropriately for both the intention-to-treat (ITT) and PP populations.

According to the clinical experts consulted for this review, the baseline characteristics of study patients are reflective of patients with narcolepsy in Canada seeking further treatment options. The drug titration would be reflective of clinical practice. The primary outcome measures in the trials are used by physicians in clinical practice and measured outcomes important to patients (EDS and cataplexy). Patients were allowed to combine conventional narcolepsy medication with the drug under study. The clinical experts noted that it is common for combination therapy to be used in clinical practice; however, the interactions between concomitant medications and pitolisant hydrochloride are unknown. On that note, TCAs were not allowed as concurrent medications, despite them being common anticataplectic drugs, the clinical experts reported. This may decrease the generalizability of the trial population. Adherence to treatment remained high, at more than 80%, in all trials.

Other Relevant Evidence

The open-label HARMONY III extension study^16,17 provides long-term safety and efficacy data that supplements evidence from the RCTs in the Systematic Review.

Description of Studies

The HARMONY III (NCT01399606) long-term, open-label, uncontrolled extension study was conducted to evaluate the efficacy and safety of pitolisant hydrochloride at daily doses of 5 mg, 10 mg, 20 mg, and 40 mg for the treatment of EDS in patients with narcolepsy with or without cataplexy for up to 5 years. Of the 102 patients enrolled in the HARMONY III extension trial, 86 were pitolisant hydrochloride–naive or secondary-naive (the naive group) and were not receiving pitolisant hydrochloride at the time of study enrolment and 16 were patients from a French compassionate use program (CUP) who were being treated with pitolisant hydrochloride in the 2 weeks preceding study enrolment. Of the 86 patients in the naive group, 73 had never been treated with pitolisant hydrochloride and 13 had been treated with pitolisant hydrochloride when they participated in single-blind or double-blind trials, including HARMONY 1,¹³ HARMONY II,¹⁸ and HARMONY 1bis.¹⁴

At study inclusion, CUP patients could continue at their established pitolisant hydrochloride dose (20 mg per day or 40 mg per day) without up-titration. The I patients began pitolisant hydrochloride treatment with a 1-month individual up-titration scheme that started at 5 mg per day and increased to up to 40 mg per day. Patients recruited from France who had received at least 1 dose of pitolisant hydrochloride and completed the initial 1-year period in the HARMONY III trial were eligible to continue treatment in a follow-up period for up to 5 more years.

A total of 102 patients with narcolepsy from 8 centres in France (n = 77) and Hungary (n = 25) were enrolled in the extension study, HARMONY III, with the first enrolment occurring in June 2011. After the initial 12-month treatment period, 48 patients from France continued in the 5-year follow-up period. Patients were required to have an ESS score of at least 12 to enrol in the extension study. Overall, the mean age of all participants was 38.0 (SD = 14.9) years, and slightly more than half were female (55.9%). About 75% of naive patients and CUP patients reported a history of cataplexy. Patients in the extension study could take concomitant medications for narcolepsy, including anticataplectics and/or psychostimulants. At inclusion, 35.3% of all patients were taking concomitant medications, with more CUP than naive patients taking concomitant medications (56.3% versus 31.4%). Overall, the baseline characteristics of the patients enrolled in the HARMONY III trial were generally consistent with the baseline characteristics of patients randomized in the pivotal trials. Characteristics of the French patients who continued in the 5-year follow-up period were similar to those of the total study population.

Efficacy Results

Sleepiness, Alertness, Severity of Daytime Sleepiness

In the HARMONY III extension study at year 1, the mean change from baseline in ESS score was –3.99 (SD = 4.56). Fifty-eight (58.2%) patients were considered responders, defined as having an ESS score no higher than 10 or a change from baseline of at least 3Its. Among naive patients, the mean change from baseline in ESS score was –4.30 (SD = 4.47). Forty-nine (59.8%) patients were considered responders. CUP patients, who were already receiving pitolisant hydrochloride treatment at inclusion, had a lower mean ESS score at baseline, and the mean change from baseline in ESS score was –2.38 (SD = 0.79). Eight (50.0%) patients were considered responders.

For patients taking concomitant narcolepsy treatments, the mean change from baseline in ESS score was –3.15 (SD = 4.01), –3.64 (SD = 4.55), and –4.00 (SD = 2.35) for patients taking psychostimulants (n = 26), anticataplectics (n = 14), and both psychostimulants and anticataplectics (n = 13), respectively. For the 45 patients taking pitolisant hydrochloride only (i.e., no concomitant treatments), the mean change from baseline was –4.67 (SD = 5.27). Thirteen (50.0%), 8 (57.1%), and 10 (76.9%) patients taking psychostimulants, anticataplectics, and psychostimulants plus anticataplectics, respectively, were considered responders. Twenty-six (57.8%) patients taking pitolisant hydrochloride only (i.e., no concomitant treatments) were considered responders.

The changes from baseline in ESS scores remained similar during the long-term follow-up in the French cohort. Among French patients who continued in the long term follow-up, the ESS mean change from baseline was –4.41 (SD = 5.38) at year 2 (n = 45), –4.45 (SD = 6.16) at year 3 (n = 38), –4.76 (SD = 5.73) at year 4 (n = 34), and –6.07 (SD = 7.19) at year 5 (n = 14). At 5 years, the mean change from baseline in ESS score was –8.17 (SD = 8.93) and –4.50 (SD = 5.71) for naive (n = 6) and CUP (n = 8) patients, respectively. Of the 14 patients remaining at 5 years, 10 (71.4%) were considered responders, 5 (83.3%) of whom were naive patients and 5 (62.5%) of whom were CUP patients.

For patients taking concomitant narcolepsy treatments, the mean change from baseline in ESS score after 5 years was –5.67 (SD = 6.11), –6.33 (SD = 7.77), and –5.50 (SD = 3.87) for patients taking psychostimulants (n = 3), anticataplectics (n = 3), and both psychostimulants and anticataplectics (n = 4), respectively. For the 4 patients taking pitolisant hydrochloride only (i.e., no concomitant treatments), the mean change from baseline was –6.75 (SD = 11.95). All patients remaining at 5 years, regardless of concomitant treatment, were considered responders.

A total of 71.7% of the 67 patients who completed the initial 1-year treatment period reported a CGI-C score of 1 (very much improved) or 2 (much improved), 22.4% reported a score of 3 (minimally improved), and 6.0% reported a score of 4 (no change). Three-quarters (73.1%) of naive patients and 66.7% of CUP patients were at least much improved, 21.2% and 26.7%, respectively, were minimally improved, and 5.8% and 6.7%, respectively, reported no change. Among French patients who continued in the long term follow-up, the proportion of patients who reported a much improved or very much improved CGI-C score compared to baseline was 77.3% at 2 years (n = 44), 84.2% at 3 years (n = 38), 73.5% at 4 years (n = 34), and 64.3% at 5 years (n = 14). At 5 years, 83.4% of naive patients (n = 5) and 50.0% of CUP patients (n = 4) were at least much improved, 16.7% of naive patients (n = 1) and 37.5% of CUP patients (n = 3) were minimally improved, 12.5% of CUP patients (n = 1) reported no change.

A total of 75.0% of patients (75.0% naive and 75.1% CUP) rated the effect of pitolisant hydrochloride as moderate or marked on the PGO test after 1 year of treatment. Among French patients who continued in the long term follow-up, the proportion of patients who reported a moderate or marked effect of pitolisant hydrochloride on the PGO test was 72.8% at 2 years (n = 44), 84.2% at 3 years (n = 38), 84.4% at 4 years (n = 32), and 64.3% at 5 years (n = 14) of treatment. At 5 years, 83.4% of naive and 50.0% of CUP patients rated the effect of pitolisant hydrochloride as moderate or marked.

Frequency and Severity of Cataplexy Attacks

At the end of the initial 1-year study period, the mean change in total cataplexy from baseline was a –0.25 (SD = 1.37) for all 44 patients who completed the sleep diary; for naive patients, the mean change was –0.25 (SD = 1.38), and for CUP patients it was 0.00 (SD = NA). The mean change in partial cataplexy from baseline was –0.49 (SD = 1.94) for all patients, –0.49 (SD = 1.96) for naive patients, and 0.53 (SD = NA) for CUP patients.

Health-Related Quality of Life

The mean EQ-5D visual analogue scale (VAS) score for all patients was 65.5 (SD = 16.1) at baseline and 72.4 (SD = 16.2) at 1 year, with a mean change of 6.8 (SD = 15.4) from baseline. For naive patients, the mean EQ-5D VAS score was 64.3 (SD = 15.9) at baseline and 73.5 (SD = 17.5) at 1 year; with a mean change of 9.2 (SD = 15.4) from baseline. For CUP patients, the EQ-5D VAS score was 69.6 (SD = 16.7) at baseline and 68.8 (SD = 11.4) at 1 year; with a mean change of –0.8 (SD = 12.7) from baseline.

Among French patients who continued in the long term follow-up, the mean of the EQ-5D VAS was 70.5 (SD = 15.9) at 2 years (n = 44), 69.5 (SD = 13.2) at 3 years (n = 38), 72.2 (SD = 13.3) at 4 years (n = 33), and 75.0 (SD = 12.2) at 5 years (n = 14). At 5 years, the EQ-5D VAS score was 80.5 (SD = 12.5) for naive patients and 70.9 (SD = 10.9) for CUP patients, with a change of 13.8 (SD = 15.5) and 2.4 (SD = 12.5) from baseline, respectively.

Sleep Attacks

At the end of the initial 1-year study period, among patients who completed the sleep diary (n = 44), the mean change in the daily number of sleep attacks from baseline was –0.37 (SD = 1.41) for all patients, –0.39 (SD = 1.42) for naive patients, and 0.47 (SD = NA) for CUP patients. The mean (SD) change in the duration of diurnal involuntary sleep attacks from baseline was –0.37 (SD = 1.41) minutes for all patients, –0.39 (SD = 1.42) minutes for naive patients, and 0.47 (SD = NA) minutes for CUP patients.

Nocturnal Sleep Properties

Among patients who completed the sleep diary (n = 44), the mean change in daily number of nocturnal awakenings from baseline to the 1-year visit was –0.42 (SD = 1.18) for all patients, –0.42 (SD = 1.19) for naive patients, and –0.14 (SD = NA) for CUP patients. The mean change in the duration of nocturnal awakening from baseline to the 1-year visit was –0.09 (SD = 0.73) hours for all patients, –0.10 (SD = 0.73) hours for naive patients, and 0.18 (SD = NA) hours for CUP patients. The mean change in the duration of nocturnal sleep from baseline to the 1-year visit was –0.10 (SD = 1.19) hours for all patients, –0.09 (SD = 1.21) hours for naive patients, and –0.37 (SD = NA) hours for CUP patients.

Number of Hallucinations

At the end of the initial 1-year study period, among patients who completed the sleep diary (n = 44), the mean change in the frequency of hallucinations from baseline was –0.06 (SD = 0.25) for all patients, –0.06 (SD = 0.20) for naive patients, and 0.0 (SD = NA) for CUP patients.

Concomitant Medication Use

The proportion of patients taking a concomitant treatment for narcolepsy or cataplexy changed from 35.3% at baseline to 52.9% over the course of the initial year. A total of 31.4% of naive and 56.3% of CUP patients were taking concomitant treatment at baseline and, over the course of the initial year, 51.2% of naive and 62.5% of CUP patients were taking concomitant medications. The most frequent treatments over the course of the study were methylphenidate (22.5%), modafinil (17.6%), and venlafaxine (13.7%). Eleven patients (10.8%) took sodium oxybate. In the French subset, the proportion of patients taking allowed concomitant treatment for narcolepsy or cataplexy in addition to pitolisant hydrochloride changed from 44.2% at baseline to 70.1% over the 5-year period. A total of 70.5% of naive and 68.8% of CUP patients were taking concomitant treatments over the 5-year period, respectively. The most frequent treatments were methylphenidate (31.2%), modafinil (29.9%), venlafaxine (19.5%), and sodium oxybate (16.9%).

Harms Results

All combinations of concomitant medications for narcolepsy or cataplexy were well tolerated, except there was a greater frequency of insomnia in the subgroup of patients taking concomitant modafinil (55%; n = 5) during the follow-up period in the French subset of patients.

During the initial year, 58 patients (56.9%) reported 168 TEAEs, the most common being headache (11.8% of patients), insomnia (8.8%), weight gain (7.8%), anxiety (6.9%), depression (4.9%), and nausea (4.9%). In the French subset of patients, over the 5-year period, 72.7% of patients reported 296 TEAEs, the most common being headache (19.5%), weight increase (18.2%), nausea (11.7%), anxiety (11.7%), insomnia (11.7%), and depression (11.7%).

A total of 16 patients reported SAEs during the 5-year period in the French subset, with the most common being depression (3.9%) and pregnancy (3.9%). All SAEs were considered unrelated to the study drug, except the 1 spontaneous abortion experienced by a patient who discontinued the study drug and permanently withdrew from the trial. One death was reported in the follow-up period. The clinical study report indicated that the death was determined to be not related to the study medication.

Among all patients, the mean 13-item Beck Depression Inventory-Short Form (BDI-SF-13) score was 4.1 (SD = 3.5) at baseline and 3.8 (SD = 4.1) at the 1-year visit. The mean BDI-SF-13 score among the 12 patients in the French subset at the year-5 visit was 2.4 (SD = 2.8). At each time point, no more than 1 patient experienced severe depression.

Critical Appraisal

The HARMONY III extension study allowed for investigation of the long-term efficacy and harms of pitolisant hydrochloride for up to 5 years. Limitations of the study include the absence of an active comparator, which limits causal conclusions. An additional limitation is the open-label study design and the unblinding of the study drug in the follow-up period, which could bias the reporting of end points. There was no sample-size calculation or statistical testing for changes from baseline, making it difficult to detect a clinically relevant treatment effect. All the end points in the HARMONY III trial were subjective, so it is possible that efficacy outcomes and known harms could have been overestimated. Findings are at a high risk of confounding because of the use of concomitant treatments and a lack of control for confounding variables. None of the P values were adjusted for multiplicity and should be considered hypothesis-generating.

Subgroup analyses were descriptive and often limited to few patients, reducing the chance of detecting a true effect. Interpretation of these patient-reported outcomes are also limited by the large amount of missing data due to attrition. More than one-third of patients discontinued the extension study in the first year, mainly because of AEs or a lack of perceived efficacy. This attrition could have resulted in a population of patients that were more tolerant of pitolisant hydrochloride, as those not responding to treatment may have been less likely to continue into the follow-up period. The presence of patients more tolerant of pitolisant hydrochloride can also lead to biased estimates of efficacy and AEs, potentially resulting in greater efficacy and fewer AEs being reported. The use of concomitant psychostimulant and/or anticataplectic drugs by patients during the HARMONY III extension study may have increased the risk of additional side effects not attributable to pitolisant hydrochloride alone. Furthermore, for the primary efficacy outcome of ESS score at the 1-year time point, LOCF was used for patients without final values, which may have biased the efficacy results, as these values may not be reflective of the true trajectory of this outcome.

External Validity

With respect to external validity, although no patients living in Canada were enrolled in the HARMONY III extension study, the characteristics of the patients enrolled in the trials were representative of patients with narcolepsy in Canada, according to the clinical experts consulted. Doses of pitolisant hydrochloride administered were in line with what would be expected in clinical practice.

Conclusions

Pitolisant hydrochloride has shown benefits over placebo in the treatment of EDS and cataplexy in patients with narcolepsy. However, because of limited comparisons and the failure of noninferiority testing, it is not possible to conclude that pitolisant hydrochloride is similar or noninferior to modafinil or other active drugs used for the treatment of EDS or cataplexy. In addition, the interaction between pitolisant hydrochloride and concomitant medications for narcolepsy is unclear. Pitolisant hydrochloride appeared to be well tolerated and no major safety concern was identified. In alignment with clinical expert input, pitolisant hydrochloride may provide an additional option for the treatment of narcolepsy because of its superiority to placebo. However, its place in therapy relative to other medications will be a challenge to elucidate from the trial results.

Introduction

Disease Background

Narcolepsy is a chronic neurologic condition that is caused by an imbalanced sleep-wake cycle or sleep-wake instability.¹ It is characterized by chronic, excessive episodes of drowsiness during the day, also known as EDS.² According to the International Classification of Sleep Disorders, Third Edition, diagnostic criteria, there are 2 types of narcolepsy: type 1 is classified as EDS with cataplexy, and type 2 is classified as EDS alone.¹ Cataplexy is defined as a sudden episode of partial or complete paralysis of voluntary muscles, triggered by strong (often positive) emotion.³ Approximately 60% to 70% of patients with narcolepsy have cataplexy (type 1 disease).⁴ Other symptoms, including automatic behaviours in which actions are performed in a semiconscious way without awareness, such as walking, eating, driving, hallucinations, and sleep paralysis. Symptoms can vary for each patient.¹⁹ Approximately 1 in 2,000 individuals in Canada are affected by narcolepsy.² The prevalence is likely underestimated because of misdiagnosis and the limited availability of health care providers with experience in narcolepsy.^5-7

Type 1 narcolepsy is thought to occur because of a loss hypocretinergic cells in the lateral hypothalamus, likely due to an autoimmune etiology, and 90% of these patients have evidence of hypocretin deficiency.¹ Type 2 narcolepsy is associated with significant EDS but no cataplexy or evidence of hypocretin deficiency.¹ There is likely significant overlap between type 2 narcolepsy and idiopathic hypersomnia.¹ The typical onset of symptoms happens in a person’s teens or early adulthood, but diagnosis can be delayed many years.^5-7 Hypocretin (or orexin) is thought to stabilize the state between sleep and wakefulness. Consequently, when such cells are lost, the boundary between sleep and wakefulness becomes more porous or fluid, leading to sleep phenomena intruding on daytime functioning.¹ Sleep paralysis involves the paralysis of REM sleep intruding on wakefulness at sleep onset or offset).¹ Hypnagogic (sleep onset) and/or hypnopompic (sleep offset) hallucinations are dream-like phenomena that occur during the transition between sleep and wake, that intrude on daytime functioning. Insomnia is wakefulness of the day that intrudes on sleep at night.¹

Narcolepsy can affect a patient’s HRQoL by reducing productivity and the ability to work and function in day-to-day life and social settings.⁸ Patients can experience EDS during common situations in the day, such as work or driving, and often during sedentary periods.³ Narcolepsy is also associated with an increased risk for comorbid conditions, including depression, anxiety, obesity, cardiovascular disease, and an increased overall mortality risk⁸

Updated diagnostic criteria include the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) and International Classification of Sleep Disorders, Third Edition.^1,20-22 DSM-5 criteria require EDS in association with any 1 of the following: cataplexy; cerebrospinal fluid (CSF) hypocretin deficiency; REM sleep latency of no more than 15 minutes on nocturnal polysomnography; or mean sleep latency of no more than 8 minutes on multiple sleep latency testing (MSLT) with at least 2 sleep-onset REM periods (SOREMPs).²¹ International Classification of Sleep Disorders, Third Edition relies on objective data in addition to EDS, somewhat complicating the diagnostic criteria: cataplexy and either positive MSLT or polysomnography findings or CSF hypocretin deficiency; MSLT criteria similar to DSM-5, except that a SOREMP on polysomnography may count as 1 of the SOREMPs required on MSLT; and distinct division of narcolepsy into type 1, which requires the presence of cataplexy or documented CSF hypocretin deficiency, and type 2, in which cataplexy is absent and CSF hypocretin levels are either normal or undocumented.²³

Standards of Therapy

Based on input from the clinicians consulted by CADTH for this review, the treatment of narcolepsy includes a biopsychosocial approach. Education about the illness is important for patients, who need to understand it is a primary sleep disorder. According to the clinical experts, strategic naps and/or the appropriate use of countermeasures for sleepiness, such as caffeine, can offer modest benefit.

According to the most recent (2021) guidelines from the AASM for the treatment of hypersomnolence disorders, initial pharmacologic treatments for EDS associated with narcolepsy include modafinil and the stimulant methylphenidate (Table 4).^12,24 If response is suboptimal, more potent stimulants, including lisdexamfetamine, dextroamphetamine, and/or dextroamphetamine salts can be considered. Solriamfetol is also a recommended drug for treatment of narcolepsy.^12,24 For symptoms related to REM intrusion, including cataplexy, SSRIs and SNRIs are generally considered first-line therapies (e.g., fluoxetine and venlafaxine).^12,24 TCAs, such as protriptyline, imipramine, and clomipramine, can also be considered for such symptoms.^12,24 Sodium oxybate exits as an alternative treatment option; however, it received a Do Not List recommendation from CADTH and is not reimbursed by any public drug plans participating in the CADTH process. It is a powerful gamma-aminobutyric acid-B agonist that improves symptoms of EDS and cataplexy but has significant potential for side effects.^12,24 All of the above-mentioned medications are thought to treat the underlying symptoms of narcolepsy, including primarily daytime sleepiness and cataplexy, if present. The AASM recommends pitolisant hydrochloride, sodium oxybate, and dextroamphetamine as the only medications indicated for cataplexy.^12,24

According to the clinical experts consulted for this review, the current treatment standard in Canada for EDS in patients with narcolepsy is modafinil, which is thought to improve wakefulness by reducing dopamine reuptake. Stimulants such as methylphenidate (as well as dextroamphetamine and similar amphetamines) have been used off-label for narcolepsy; however, abuse and diversion of these medications does occur. SSRIs are also used off-label for the treatment of cataplexy in patients with narcolepsy.

Drug

Pitolisant hydrochloride is an inverse H₃ receptor antagonist/inverse agonist. The human H₃ receptor functions as a presynaptic autoreceptor on histamine-containing neurons.^9,11,25 H₃ antagonists promote wakefulness by increasing histamine synthesis and release. By binding competitively to H₃ autoreceptors on presynaptic histaminergic neurons, pitolisant hydrochloride blocks the normal negative-feedback mechanisms for histamine release, increasing histaminergic transmission and resulting in enhanced histamine synthesis and release.¹⁰ Pitolisant hydrochloride is administered orally in doses up to 40 mg daily and is available as 5 mg and 20 mg tablets.^26,27 It is indicated for the treatment of EDS or cataplexy in adults with narcolepsy. It received a Notice of Compliance from Health Canada on May 25, 2021, after undergoing standard review. The reimbursement request is per the indication.²⁷ Pitolisant hydrochloride has not previously been reviewed by CADTH for any indication.

Stakeholder Perspectives

Patient Group Input

This section was prepared by CADTH staff based on the input provided by patient groups. The full patient group submission can be found at the end of this report in the Stakeholder Feedback section.

One patient group, WUN, submitted patient input for this review. WUN is a nonprofit patient advocacy organization established in 2008 that aims to accelerate research, increase awareness of narcolepsy, and provide supportive services to patients. The input was based on a survey of 19 individuals: 18 in Canada who have a narcolepsy diagnosis or are undiagnosed but living with narcolepsy symptoms, and 1 caregiver. Most patients were aged 18 to 44 years (77%) and female (72%), and none had experience with the treatment under review.

Respondents reported EDS to be the most troubling symptom of narcolepsy, with 39% of respondents giving it a rating of 6 on a scale of 1 (not at all bothersome) to 7 (completely bothersome). The second-most troublesome symptom was DNS, followed by hallucinations when falling asleep or waking up, cataplexy, and sleep paralysis. The negative impacts of narcolepsy on respondents’ lives include mental health and emotional symptoms (mood swings, anger, depression, and anxiety), missing out on social activities, difficulty managing career and job tasks, depending on others for support for daily activities, and difficulty maintaining physical health and wellness (weight gain). Treatments that respondents were currently using for their narcolepsy include stimulants (56%), antidepressants (33%), sodium oxybate (13%), and modafinil or armodafinil (13%). Some respondents reported that the physical side effects (28%) and mental side effects (39%) of their current treatment options were moderately or extremely challenging.

Respondents would like a new drug or treatment to be more effective in treating symptoms of sleepiness, cataplexy, and DNS. Respondents indicated a desire to have a treatment that is easy to swallow and does not cause nausea, weight gain, or affect mood or personality. Respondents want a treatment with an extended-release formulation that allows them to stay awake longer during the day without having to take additional doses.

Clinician Input

Input From Clinical Experts Consulted by CADTH

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

Unmet Needs

Although existing medications treat the underlying symptoms of narcolepsy, including primarily daytime sleepiness and cataplexy, it is believed that none of the available treatment options address the fundamental underlying neurochemical abnormality of loss of hypocretin cells or the secondary absence or reduction of available CNS hypocretin associated with narcolepsy.

According to the clinical experts consulted for this report, treatment goals are primarily to improve quality of life. Although narcolepsy is not lethal, symptoms of EDS and cataplexy can be debilitating if left uncontrolled. In severe circumstances, sleep attacks can occur when a person is eating or even talking to someone.^3,7,31 If left uncontrolled, these symptoms limit the ability of patients to perform basic daily activities, such as driving, working, and interacting with people. Cataplexy (which occurs in 60% to 70% of patients with narcolepsy) is equally, if not more, debilitating when left uncontrolled. Patients cannot drive or walk outside safely because surprises could trigger a cataplectic attack. Basic daily activities, such as showering and bathing, dressing, and eating can be dangerous and/or challenging when patients are untreated.^3,7,31 Without treatment, most patients have very limited, if any, work options, and may not be able to attend school.^3,7,31 These symptoms can lead to isolation, anxiety, and depression.^3,7,31 Treatment is aimed at reducing EDS and cataplexy potential, so that patients are not dependent on caregivers for support and can interact and be functional members of society. Treatment can significantly improve alertness and the ability of patients to be functional members of society during the day. Diagnosis is often delayed, often occurring 10 years or longer after symptom onset, potentially leading to significant suffering, but if appropriate treatment is initiated, tolerated, and maintained, up to 80% of functional capacity could be retained.³²

According to the clinical experts consulted for this review, several problems persist with existing treatment options. Not all patients respond to treatment with SSRIs, SNRIs, or TCAs, and some can become tolerant to treatment. Tolerance to the REM-suppressing effects of SSRI, SNRI, and TCA medications frequently occurs, leading to persistent cataplexy. Side effects can be problematic and can include stomach upset, night sweats, sexual side effects, and headaches. Side effects can include excessive sedation during the day, which can be a problem despite anticataplectic effects. Stimulants may not fully resolve daytime sleepiness and drugs may wear off at inopportune times, leading to EDS in the evenings and/or insomnia difficulties at night. Problematic side effects can include appetite suppression, anxiety, increased blood pressure, cardiac effects, allergic reaction, reduced seizure threshold, fetal defects, inactivation of birth control, and hair loss. There can be the potential for abuse or sequestration; usually patients with narcolepsy have a low abuse potential even though they may require high doses, but some patients could experience sequestration temptation.

Place in Therapy

Based on the efficacy of pitolisant hydrochloride in early studies, its novel mechanism of action as a H₃ antagonist/inverse agonist, and its relatively favourable side-effect profile, it is likely to become an early treatment option. It received a strong recommendation in the most recent AASM (2021) guidelines for the treatment of hypersomnolence disorders.¹² Although pitolisant hydrochloride does not address the underlying abnormality associated with most cases of narcolepsy related to the loss of hypocretinergic neurons, no other drugs available for the treatment of narcolepsy work through histamine. Hypocretin neurons are known to project to the tuberomammillary nucleus and contribute to wakefulness, and patients with narcolepsy have been shown to have lower histamine levels.

It will be an early drug to consider for the treatment of narcolepsy. It may also find a niche as an adjunct treatment to boost the efficacy of other drugs. It may also become a drug of choice for patients in whom stimulant and/or other therapies are contraindicated because it has no effect on birth control efficacy (unlike modafinil) and no significant known cardiovascular effects (unlike other stimulants). Because efficacy studies have shown pitolisant hydrochloride to be not visibly superior to modafinil, this drug will not change the trajectory of the treatment algorithm for narcolepsy. The main reason other medications may be used first is increased familiarity and the availability of longer-term data for other medications, such as modafinil and methylphenidate. However, it also seems quite expensive compared to methylphenidate and modafinil, which may limit its utility as an early drug. Solriamfetol might also be considered an early treatment option, but in some ways has issues similar to pitolisant hydrochloride, including cost, lack of long-term data, and minimal doctor familiarity with the medication. Patients most in need of intervention include those who cannot tolerate stimulant therapies, those who are concerned about getting pregnant, and those with a history of drug abuse.

A trial of SOC would not be required to initiate a trial of pitolisant hydrochloride. Pitolisant hydrochloride will be able to be considered as a first-line or near first-line therapy. If SOC were required, then yes, it would be an issue for jurisdictions that would only have such medications available for attention-deficit/hyperactivity disorder. Jurisdictions should continue to provide coverage for prior therapies that are currently considered SOC when used in combination with pitolisant hydrochloride because the mechanism of action of pitolisant hydrochloride is quite different than any currently available drug, which represents an exciting prospect for patients living with this debilitating condition.

The consistent use of ongoing anticataplectic treatments in combination with pitolisant hydrochloride may mask and/or minimize the potential benefits of pitolisant hydrochloride for cataplexy. However, if pitolisant hydrochloride has minimal benefits for cataplexy, it this would be difficult to assess. In short, it will be difficult to properly assess the potential benefits of pitolisant hydrochloride for cataplexy when it is used in combination with ongoing anticataplectic treatments.

It is difficult to predict which symptoms or disease characteristics will best respond to pitolisant hydrochloride. A patient with mild narcolepsy with cataplexy is most likely to respond because the hypocretin neurons are connected directly to the tuberomammillary nucleus to help promote wakefulness. Because the overall effect of pitolisant hydrochloride is comparable to that of modafinil, patients with more severe symptoms of narcolepsy will require more aggressive therapies; however, even those patients may benefit significantly from pitolisant hydrochloride as an adjunct.

Patient Population

A number of factors make diagnosis extremely challenging. Patients often first visit a family doctor or pediatrician, who may not immediately recognize this condition. Patients are frequently misdiagnosed, and narcolepsy is undiagnosed in more than 70% of patients.

Specific expertise and testing are needed to make the diagnosis, which often can occur 10 or more years after symptom onset. The MSLT administered by a specialist with training in sleep medicine is most commonly used to diagnose narcolepsy. One MSLT is thought to have a sensitivity of 85%; running it twice will bring the sensitivity to more than 95%. Patients should be off all psychiatric and stimulant medications for at least 2 weeks (some for > 1 month) before undergoing testing and should have completed sleep logs before the study. Ideally, diagnosis would involve an analysis of CSF to assess hypocretin levels, but this is technically challenging and not routinely available. Antigen testing can also be used for the diagnosis of narcolepsy with cataplexy.

Sleep medicine was recognized in Canada as an Area of Focused Competence in 2021. Most current sleep medicine specialists obtained subspecialty certification training in the US, which contributes to the shortage of physicians trained in sleep medicine. Furthermore, Quebec does not recognize sleep medicine practised by those outside of respirology, except in very few circumstances, further contributing to the probable underrecognition of this condition. Pitolisant hydrochloride will be considered for off-label use for mood disorders associated with hypersomnolence or hypersomnolence related to other conditions.

At this time, pitolisant hydrochloride may not be suitable for patients who wish to become pregnant or breastfeed. In addition, because of the lack of data on its use in children and in older adults, it should be used with caution in those populations. Pitolisant hydrochloride may not be ideal for patients who are on multiple medications (with more potential for drug interactions, particularly with those that affect the QTc interval or that are significant CYP2D6 inhibitors) or for patients who have a history of significant kidney or liver failure (with difficult-to-predict metabolism). Patients who have had adverse reactions to opioids (including hives), or who have a history of urticaria or other skin conditions, might be predisposed to an allergic reaction to pitolisant hydrochloride.

Assessing Response to Treatment

It is unclear how to select patients who might be more or most likely to respond to pitolisant hydrochloride. The primary outcomes in clinical practice will likely be a reduction in EDS, a reduction in the frequency, intensity, and duration of cataplexy episodes, and the ability to predict episodes. Clinically meaningful responses to treatment include a reduction in frequency, severity, and intensity of cataplexy episodes. Although frequency is easier to assess systematically, the intensity and severity of spells, as well as the perceived predictability of episodes, are more of a clinical assessment. For example, patients may describe certain emotions that no longer trigger episodes in the way they had before. Other parameters for assessment could include a reduction in other REM intrusion phenomena, if present, and the degree to which patients can resume normal functioning and daily activities.

Outcomes typically assessed in most clinical trials include the degree of reduction of EDS and possibly the degree of reduction in the frequency of cataplexy spells. The use of ESS scores in clinical practice to determine coverage of pitolisant hydrochloride may not be ideal. The ESS is very subjective and could easily manipulate their scores. In addition, there can be significant differences in the way male and female patients score their results, further skewing potential for the determination of coverage. In research trials, it’s ideal if patients are blinded to what they are being offered and there is no incentive to report better or worse scores. A score of 10 or lower on the ESS would indicate that sleepiness is no different than in the general population. As a comparison, an ESS score for patients with narcolepsy would typically be above 18, indicating severe sleepiness; a score of 15 to 17 indicates moderate sleepiness and a score of 11 to 14 indicates mild sleepiness.

There are few data to define an effective reduction of cataplexy. Trials of sodium oxybate demonstrated a reduction in cataplexy episodes of more than 90%. Driving is not recommended if a patient has experienced a cataplexy episode in the previous year. A minimum reduction in cataplexy episodes of 50% would be meaningful. Depending on severity and frequency, fewer than 1 episode per week would be a reasonable standard.

As with other drugs for narcolepsy, there should be close follow-up of the patient in the first months of therapy. The first follow-up should occur 1 month after the initiation of pitolisant hydrochloride, then every 1 to 2 months for the next several months, and then intermittent follow-up after that, with at least yearly follow-up, at a minimum, in the long-term.

Discontinuing Treatment

Ongoing treatment will be determined by the lack of response and/or by excess adverse effects, as with most medications. Whether it will continue to be used as an adjunct therapy if abandoned as monotherapy is unclear. Excess adverse effects, drug interactions, or a patient’s wish to become pregnant may necessitate withdrawal.

Prescribing Conditions

Medical supervision in an outpatient setting with a physician trained in sleep medicine would be appropriate for patients with narcolepsy being treated pitolisant hydrochloride. In the future, psychiatrists will likely become interested in using this medication for conditions and symptoms outside of narcolepsy. At this time, because the indication for pitolisant hydrochloride is only for narcolepsy, with a conditional recommendation for idiopathic hypersomnia, prescribing should be limited to those with specialty training or certification in sleep medicine.

Drug Program Input

The drug programs provide input on each drug being reviewed through CADTH’s reimbursement review processes by identifying issues that may affect their ability to implement a recommendation. 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

EDS = excessive daytime sleepiness; ESS = Epworth Sleepiness Scale; MSLT = Multiple Sleep Latency Test; SNRI = selective noradrenergic reuptake inhibitor (e.g., fluoxetine, venlafaxine, atomoxetine); SSRI = selective serotonin reuptake inhibitors (e.g., escitalopram, sertraline).

Clinical Evidence

The clinical evidence included in the review of pitolisant hydrochloride is presented in 3 sections. The first section, the Systematic Review, includes pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as studies that were selected according to an a priori protocol. The second section includes indirect evidence from the sponsor and indirect evidence selected from the literature that met the selection criteria specified in the review. The third section includes sponsor-submitted long-term extension studies and additional relevant studies that were considered to address important gaps in the evidence included in the Systematic Review.

Systematic Review (Pivotal and Protocol-Selected Studies)

Objectives

The objective was to perform a systematic review of the beneficial and harmful effects of oral pitolisant hydrochloride (available in 5 mg and 20 mg tablets) at daily doses of up to 40 mg for the treatment of EDS or cataplexy in adults with narcolepsy.

Methods

Studies selected for inclusion in the Systematic Review included pivotal studies provided in the sponsor’s submission to CADTH and Health Canada, as well as those meeting the selection criteria presented in Table 6. Outcomes included in the CADTH review protocol reflect outcomes considered to be important to patients, clinicians, and drug plans.

Table 6: Inclusion Criteria for the Systematic Review

AE = adverse events; CGI-C = Clinical Global Impression of Change; CGI-S = Clinical Global Impression of Severity; EDS = excessive daytime sleepiness; ESS = Epworth Sleepiness Scale; FOSQ = Functional Outcomes of Sleep Questionnaire; HRQoL = health-related quality of life; LSEQ = Leeds Sleep Evaluation Questionnaire; MSLT = Multiple Sleep Latency Test; MWT = Maintenance of Wakefulness Test; RCT = randomized controlled trial; SAE = serious adverse event; SART = Sustained Attention to Response Task; SNRI = selective noradrenergic reuptake inhibitor; SSRI = selective serotonin reuptake inhibitors; TCA = tricyclic antidepressant; WDAE = withdrawal due to adverse event.

The literature search for clinical studies was performed by an information specialist using a peer-reviewed search strategy in accordance with the Peer Review of Electronic Search Strategies (PRESS) checklist.³⁴

Published literature was identified by searching the following bibliographic databases: MEDLINE All (1946–) via Ovid and Embase (1974–) via Ovid. All Ovid searches were run simultaneously as a multifile search. Duplicates were removed using Ovid deduplication for multifile searches, followed by manual deduplication in Endnote. The search strategy comprised both controlled vocabulary, such as the National Library of Medicine’s MeSH (Medical Subject Headings), and keywords. The main search concept was Wakix (pitolisant hydrochloride). The following clinical trials registries were searched: the US National Institutes of Health’s clinicaltrials.gov, WHO’s International Clinical Trials Registry Platform (ICTRP) search portal, Health Canada’s Clinical Trials Database, and the European Union Clinical Trials Register.

No filters were applied to limit the retrieval by study type. Where possible, retrieval was limited to the human population. Retrieval was not limited by publication date or by language. Conference abstracts were excluded from the search results. Refer to Appendix 1 for the detailed search strategies.

The initial search was completed on February 4, 2022. Regular alerts updated the search until the meeting of the CADTH Canadian Drug Expert Committee (CDEC) on June 22, 2022.

Grey literature (literature that is not commercially published) was identified by searching relevant websites from the Grey Matters: A Practical Tool For Searching Health-Related Grey Literature checklist.³⁵ Included in this search were the websites of regulatory agencies (FDA and European Medicines Agency). Google was used to search for additional internet-based materials. Refer to Appendix 1 for more information on the grey literature search strategy.

These searches were supplemented by reviewing bibliographies of key papers and through contacts with appropriate experts. In addition, the manufacturer of the drug was contacted for information regarding unpublished studies. A focused literature search for network meta-analyses dealing with Wakix (pitolisant hydrochloride) and narcolepsy was run in MEDLINE All (1946–) on February 4, 2022. No limits were applied to the search.

Two CADTH clinical reviewers independently selected studies for inclusion in the review based on titles and abstracts, according to the predetermined protocol. Full-text articles of all citations considered potentially relevant by at least 1 reviewer were acquired. Reviewers independently made the final selection of studies to be included in the review, and differences were resolved through discussion.

Indirect Evidence

In addition to the sponsor-provided indirect evidence, additional indirect evidence that includes the patients, interventions, comparators, and outcomes specified in Table 6 may be summarized and critically appraised, if considered relevant by CADTH.

Other Relevant Studies

Additional studies not meeting the selection criteria of the protocol may be considered for inclusion in the report in this section.

Findings From the Literature

A total of 5 reports from 3 studies were identified from the literature for inclusion in the Systematic Review (Figure 1).^13-15,36,37 The included studies are summarized in Table 7. A list of excluded studies is presented in Appendix 2.

Figure 1: Flow Diagram for Inclusion and Exclusion of Studies

Table 7: Details of Included Studies

BDI-SF-13 = 13-item Beck Depression Inventory-Short Form; CGI-C = Clinical Global Impression of Change; CGI-S = Clinical Global Impression of Severity; CNS = central nervous system; DB = double-blind; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; EDS = excessive daytime sleepiness; ESS = Epworth Sleepiness Scale; HRQoL = health-related quality of life; MWT = Maintenance of Wakefulness Test; PGO = patient global opinion; RCT = randomized controlled trial; SART = Sustained Attention to Response Task; SSRI = selective serotonin reuptake inhibitors; TCA = tricyclic antidepressant.

Sources: Clinical Study Reports for HARMONY 1, HARMONY 1bis, and HARMONY CTP.^13-15

Description of Studies

Three double-blind, placebo-controlled, phase III, RCTs met the inclusion criteria for the Systematic Review. In all 3 trials, patients were included if they had narcolepsy with cataplexy. The HARMONY 1 and HARMONY 1bis trials also included patients without cataplexy (type 2 narcolepsy).

The HARMONY 1 trial was a 12-week, double-blind, RCT that assessed the superiority of pitolisant hydrochloride to placebo with regard to EDS (based on the ESS) in patients with narcolepsy. An additional efficacy objective was to evaluate the noninferiority of pitolisant hydrochloride to modafinil for EDS (based on the ESS). Patients were enrolled from 24 sites in Europe, beginning on May 26, 2009. The trial consisted of a 2-week washout period (during which patients discontinued treatments for EDS) and a 1-week baseline period, after which patients who remained eligible were randomized to pitolisant hydrochloride (10 mg to 40 mg), modafinil (100 mg to 400 mg), or placebo. Patients underwent dose titration for 3 weeks and a final 5-week maximum dose period (highest dose reached and tolerated) (Figure 2). The treatment period was followed by a 1-week withdrawal period, during which all patients received placebo. The study randomized 95 patients (32 to pitolisant hydrochloride, 30 to placebo, and 33 to modafinil).

The HARMONY 1bis trial was similar to the HARMONY 1 trial in design, objectives, and end points. The double-blind, RCT assessed the superiority of pitolisant hydrochloride to placebo with respect to EDS (based on the ESS) in patients with narcolepsy. An additional efficacy objective was to evaluate the noninferiority of pitolisant hydrochloride to modafinil for EDS (based on the ESS). Patients were enrolled from 32 sites in Europe beginning on October 25, 2021. There was a 2-week washout period during which prohibited treatments (particularly psychostimulants) were discontinued and a 1-week baseline period, after which patients who met the selection criteria were randomized into 1 of 3 treatment groups: placebo, pitolisant hydrochloride (5 mg to 20 mg), or modafinil (100 mg to 400 mg). Patients underwent dose titration for 3 weeks and a final 5-week maximum dose period (highest dose reached and tolerated) (Figure 3). The treatment period was followed by a 1-week withdrawal period, during which all patients received placebo. A total of 166 patients were randomized (33 to placebo, 67 to pitolisant hydrochloride, and 66 to modafinil).

The HARMONY CTP trial was an 11-week, double-blind, RCT comparing pitolisant hydrochloride to placebo. It focused on the safety of pitolisant hydrochloride and its efficacy in decreasing the frequency of cataplexy attacks in patients who had narcolepsy with cataplexy. Secondary objectives included exploring the efficacy of pitolisant hydrochloride in reducing EDS and analyzing the tolerance profile of pitolisant hydrochloride in patients who had narcolepsy with cataplexy. Patients were enrolled from 16 sites in Europe and Turkey, beginning on April 19, 2013. There was a 1-week washout period, during which prohibited treatments were discontinued (Figure 4). This was followed by a 2-week period during which baseline tests were conducted, after which patients were randomized to pitolisant hydrochloride or placebo. There was 7-week treatment period that consisted of 3 weeks of dose titration up to a maximum of 40 mg and 4 weeks of stable dosing. All patients then received 1 week of placebo as the withdrawal phase. A total of 105 patients were randomized to receive either placebo (n = 51) or pitolisant hydrochloride (n = 54). A total of 13 sites were enrolled outside of Canada. The first patient was enrolled on April 19, 2013, and the last patient completed the study on January 28, 2015.

Figure 2: Study Design of the HARMONY 1 Trial

BF2.649 = pitolisant hydrochloride; D = day; ESS = Epworth Sleepiness Scale; t = telephone contact; V = visit; W = week.

Source: Clinical Study Report for HARMONY 1.¹³

Figure 3: Study Design of the HARMONY 1bis Trial

BF2.649 = pitolisant hydrochloride; D = day; ESS = Epworth Sleepiness Scale; t = telephone contact; V = visit; W = week.

Source: Clinical Study Report for HARMONY 1bis.¹⁴

Figure 4: Study Design of the HARMONY CTP Trial

BF2.649 = pitolisant hydrochloride; D = day; ESS = Epworth Sleepiness Scale; Tel = telephone contact; V = visit; W = week.

Source: Clinical Study Report for HARMONY CTP.^13,15

Populations

Inclusion and Exclusion Criteria

In all 3 trials, patients were included if they had narcolepsy with cataplexy. The HARMONY 1 and HARMONY 1bis trials also included patients without cataplexy (type 2 narcolepsy). The HARMONY 1 and HARMONY 1bis trials required patients to have an ESS score of at least 14 during the baseline period, whereas the HARMONY CTP trial required an ESS score of at least 12. The HARMONY CTP trial included patients with at least 3 cataplexy attacks weekly. In all trials, patients were required to be free of drugs or to have discontinued psychostimulants for at least 7 days (HARMONY 1 and HARMONY 1bis) or 14 days (HARMONY CTP) during baseline. Patients with severe cataplexy were permitted stable doses of anticataplectic medications (except TCAs) that were administrated for at least 1 month before the trial. Doses of anticataplectic medications remained stable throughout the trials.

Patients with narcolepsy were excluded if they had other conditions that could have been the primary cause of EDS. Patients with a history of substance abuse or dependence disorder in the previous 1 year, psychiatric and neurologic disorders, prior severe adverse reactions to CNS stimulants, inability to continue daily activities safely without treatment for EDS, or any clinically significant illness that would interfere with the completion of the study were excluded.

Baseline Characteristics

In the HARMONY 1 trial, there were no clinically relevant differences between groups for any of the characteristics examined (Table 8). The median age of the patients ranged from 33 to 40 years across the 3 treatment groups. Approximately half of the patients were male (43% to 65% across groups). More than 90% of the patients in each group were white. The median time since diagnosis of narcolepsy was 15.2 (interquartile range [IQR], 9.2 to 25.3) years in the placebo group, 11.2 (IQR, 8.2 to 18.0) years in the pitolisant hydrochloride group, and 12.2 (IQR, 5.7 to 20.3) years in the modafinil group. Overall, approximately 80% of the patients in each group reported a history of cataplexy. More than half of all patients reported histories of sleep paralysis (48% to 67% across groups), hallucinations (58% to 64% across groups), or dyssomnia (47% to 61% across groups). Less than half of all patients (30% to 49% across groups) reported automatic behaviour. The MSWT was performed on 54 of the 98 patients at baseline, and the mean latency time appeared to be shorter in the pitolisant hydrochloride group (3.7 [SD = 2.6] minutes) than in the placebo (5.4 [SD = 2.0] minutes) and modafinil (4.9 [SD = 2.4] minutes) groups. The number of patients who took at least 1 chronic medication for the treatment of narcolepsy during the 3-month period before inclusion was 85% in the placebo and pitolisant hydrochloride groups and 70% in the modafinil group. Thirty-nine percent of patients were treated with modafinil before baseline (43% in the placebo group, 42% in the pitolisant hydrochloride group, and 33% in the modafinil group). Across all treatment arms, 33 patients with severe cataplexy were permitted to remain on their anticataplectic medications at stable doses for the duration of the trial (8 taking sodium oxybate and 25 taking antidepressants). The ESS scores at baseline ranged from 17.8 to 18.9 across treatment groups. At baseline across groups, patients with cataplexy reported approximately 1 complete or partial cataplexy episode per day.

Table 8: Summary of Baseline Characteristics — HARMONY 1 (ITT Population) Trial

CI = confidence interval; ESS = Epworth Sleepiness Scale; GMT = geometric mean based on natural logarithm; ITT = intention-to-treat; MWT = Maintenance of Wakefulness Test; SART = Sustained Attention to Response Task; SD = standard deviation.

Note: Screening is visit 1. ESS baseline is defined as (visit 2 + visit 3) / 2.

^aThere were missing data for some baseline variables. Sample sizes for the placebo, pitolisant hydrochloride, and modafinil groups, respectively, were 17, 11, and 15 for the percent of postmenopause or sterilized females; 18, 20, and 20 for the MSLT; 30, 29, and 33 for SART-NOGO; 30, 30, and 33 for SART-TOTAL; 14, 20, and 23 for the number of cataplexy episodes; 9, 8, and 14 for the number of sleep paralysis episodes, and 13, 11, and 15 for the number of hallucination episodes.

^bGMT is the geometric mean based on natural logarithm. The geometric mean was used as the data were of a log-normal distribution. The mean values of log were compared between treatment groups with a Student t-test, and converted as a geometric mean to estimate the geometric mean ratio with the corresponding 95% CI.

^cBaseline = (all episodes at visit 2 and visit 3) / (number of days at visit 2 and visit 3).

Source: Clinical Study Report for HARMONY 1.¹³

In the HARMONY 1bis trial, the 3 treatment groups were comparable for the baseline demographic characteristics of age, weight, height, body mass index (BMI), sex, and ethnicity (Table 9). The median age ranged from 37 to 43 years across the 3 treatment groups. Nearly half of patients were male (47% to 48% across groups). The median time since diagnosis of narcolepsy was 11 (range, 0 to 62) years in the placebo arm, 15 (range, 0 to 47) years in the pitolisant hydrochloride arm, and 10 (range, 0 to 59) years in the modafinil arm. Overall, approximately 80% of the patients reported a history of cataplexy (75% to 81% across groups). Patients reported histories of sleep paralysis (45% to 69% across groups), hallucinations (52% to 63% across groups), or dyssomnia (25% to 40% across groups). MSLT was performed on 129 of the 163 patients at baseline, and the mean latency time appeared comparable. The mean baseline ESS score was approximately 18 in each group at baseline. Eighty-one percent of patients in the placebo group and 75% of patients in each of the pitolisant hydrochloride and modafinil groups had a history of cataplexy. The proportion of patients who took at least 1 chronic medication before inclusion was 52% in the placebo group, 56% in the pitolisant hydrochloride group, and 61% in the modafinil group.

In the HARMONY CTP trial, both treatment groups had comparable baseline demographic characteristics (Table 10). The median age was 39 (range, 18 to 66) years in the placebo group and 34 (range, 18 to 64) years in the pitolisant hydrochloride group. Approximately half of the patients in each group were male. The mean time since diagnosis of narcolepsy was not reported. At baseline, the mean number of cataplexy attacks per week was 9.2 (SD = 8.8) and 11.0 (SD = 8.9) in the placebo and pitolisant hydrochloride groups, respectively. The severity of cataplexy was rated on the CGI-S scale as moderate to severe by most patients (92% in the placebo group and 89% in the pitolisant hydrochloride group). The proportion of patients with associated symptoms (sleep paralysis, hallucinations, automatic behaviour, and dyssomnia) were similar in the treatment groups, except for ongoing sleep paralysis (44% in pitolisant hydrochloride group and 59% in placebo group). In the 3 months preceding inclusion, 66 (63%) patients had received anticataplectic medication (69% in the placebo group and 52% in the pitolisant hydrochloride group).

Table 9: Redacted

Note that some redacted rows have been deleted.

||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||

Table 10: Summary of Baseline Characteristics— HARMONY CTP (ITT Population) Trial

BDI-SF-13 = 13-item Beck Depression Inventory-Short Form; CGI-S = Clinical Global Impression of Severity; CI = confidence interval; EDS = excessive daytime sleepiness; ESS = Epworth Sleepiness Scale; GMT = geometric mean based on natural logarithm; ITT = intention-to-treat; MWT = Maintenance of Wakefulness Test; SD = standard deviation.

Source: Clinical Study Report for HARMONY CTP.¹⁵

Interventions

All 3 trials were double-blind and investigational treatments were provided in capsules that were identical in appearance for all treatments and provided in blister packs. In the HARMONY 1 and HARMONY 1bis trials, patients were randomized to receive placebo, pitolisant hydrochloride, or modafinil, whereas in the HARMONY CTP trial, patients were randomized to placebo or pitolisant hydrochloride. For the HARMONY 1 trial, patients were instructed to take 4 capsules per day, orally, with a glass of water (2 in the morning and 2 at noon). For the HARMONY 1bis trial, patients were instructed to take 2 capsules per day, orally, with a glass of water (1 in the morning and 1 at noon). For the HARMONY CTP trial, patients were instructed to take 1 capsule per day, orally, with a glass of water before breakfast, at around 8:00 a.m. For all trials, the individual dosage of pitolisant hydrochloride and modafinil was determined by investigators, according to clinical efficacy and tolerance criteria (individual dose titration). The maximum dosage of pitolisant hydrochloride was 40 mg per day in the HARMONY 1 trial (range, 10 mg to 40 mg) and the HARMONY CTP trial (range, 5 mg to 40 mg), whereas in the HARMONY 1bis trial, the maximum dosage was 20 mg daily (range, 5 mg to 20 mg). The maximum dosage of modafinil in the HARMONY 1 and HARMONY 1bis trials was 400 mg (range, 100 mg to 400 mg).

Both the HARMONY 1 and HARMONY 1bis trials had a dose-titration period of 3 weeks and a stable-dose period of 5 weeks, whereas the HARMONY CTP trial had a dose-titration period of 3 weeks and a stable-dose period of 4 weeks. In the HARMONY 1 trial, patients were asked to take a low dose of pitolisant hydrochloride (10 mg) or modafinil (100 mg) for the first 7 days and a medium dose of pitolisant hydrochloride (20 mg) or modafinil (200 mg) for the next 7 days. At the beginning of the maximal-dose period, doses were adjusted for each patient after assessment of efficacy and tolerance by the investigator, according to a predefined dose adjustment of up to 20 mg daily for pitolisant hydrochloride. Patients continued at their assigned stable dose for an additional 4 weeks. From the 1-week withdrawal phase up to the final study visit, all patients received placebo. In the HARMONY 1bis trial, patients took a low dose (5 mg of pitolisant hydrochloride or 100 mg of modafinil or placebo) for 7 days, then a middle dose (10 mg of pitolisant hydrochloride or 200 mg of modafinil or placebo) for the next 7 days. At the beginning of the maximal-dose period, doses were adjusted for each patient after assessment of efficacy and tolerance by the investigator, according to a predefined dose adjustment of up to 20 mg daily for pitolisant hydrochloride. The dose remained stable for a 5-week period. Treatment was stopped at day 56. From the 1-week withdrawal phase up to the final study visit, all patients received placebo. During the first week of the HARMONY CTP trial, all patients took a low dose of pitolisant hydrochloride (5 mg), followed by a second week of a middle dose of pitolisant hydrochloride (10 mg) or placebo. At day 14, doses were adjusted for each patient after assessment of clinical efficacy and safety by investigators, and each patient was assigned an individual optimum dose for the following 4 weeks.

Concomitant Medications Permitted

Across all trials, patients discontinued psychostimulant medications before the start of the baseline period (14 days prior for HARMONY 1 and HARMONY 1bis; 7 days prior for HARMONY CTP). Patients with severe cataplexy who were on stable doses for at least 1 month before the trial were authorized to continue anticataplectic treatments at stable doses throughout the trial. TCAs were not authorized. In the HARMONY 1 trial, no definitive list of authorized medications was provided; however, sodium oxybate and antidepressive drugs like SSRIs and SNRIs were indicated. In the HARMONY 1bis and HARMONY CTP trials, authorized medications included sodium oxybate, sertraline, fluoxetine, venlafaxine, atomoxetine, fluvoxamine, femoxetine, citalopram, paroxetine, viloxazine, reboxetine, and any other SSRI or SNRIs.

Outcomes

A list of efficacy end points identified in the CADTH review protocol that were assessed in the clinical trials included in this review is provided in Table 11 and summarized after Table 14. The timing of efficacy evaluations in the HARMONY 1, HARMONY 1bis, and HARMONY CTP trials are shown in Table 12, Table 13, and Table 14, respectively. A detailed discussion and critical appraisal of the outcome measures is provided in Appendix 4.

Table 11: Summary of Outcomes of Interest Identified in the CADTH Review Protocol

BDI-SF-13 = 13-item Beck Depression Inventory-Short Form; CGI-S = Clinical Global Impression of Severity; ESS = Epworth Sleepiness Scale; MWT = Maintenance of Wakefulness Test; NR = not reported; PGO = patient global opinion; SART = Sustained Attention to Response Task; VAS = visual analogue scale.

Sleepiness, Alertness, and Severity of Daytime Sleepiness

Epworth Sleepiness Scale

In the HARMONY 1 and HARMONY 1bis trials, the primary outcome of interest was EDS, assessed with the ESS. EDS assessed with the ESS was a secondary outcome of interest in the HARMONY CTP trial. The ESS is a self-administered questionnaire that evaluates the chance of dozing in 8 different situations often encountered in daily life, ranging from “would never doze” to “high chance of dozing.”^38,39 The scores are summed to yield a total score from 0 to 24, with higher scores representing greater sleepiness. A score greater than 10 is considered to indicate abnormal sleepiness. The ESS was completed by the patient at each visit and reviewed by the investigator.

In all 3 trials, the baseline value was calculated as the mean of the first 2 weeks of treatment, whereas the end point value was calculated as the mean of the final 2 weeks on treatment. Responders were defined as those with an ESS score of 10 or lower at the end of treatment.

Maintenance of Wakefulness Test

The MWT is used to assess a patient’s ability to stay awake while resisting pressure to fall asleep. For all trials, patients were required to have sufficient nocturnal sleep (minimum of 6 hours) and not drink alcohol the night before the test. Patients were required to recline in a quiet, dimly lit room and instructed to stay awake as long as possible. Sleep onset was defined as either 3 consecutive 30-second epochs of stage 1 sleep or any single 30-second epoch of stage 2, 3, 4, or REM sleep. In all 3 trials, the MWT was performed 4 times at 2-hour intervals at the inclusion visit and at end point visit.

Sustained Attention to Response Task

The SART is used to evaluate a patient’s vigilance and attention. Patients were required to have sufficient nocturnal sleep (minimum of 6 hours) and not drink alcohol the night before the SART.⁴⁰ While seated in front of a computer screen, patients were shown numbers, from 1 to 9, 225 times in quasi-random order. Patients were asked to respond to the appearance of each number by pressing a button, except when the number was 3. The SART error score is the total number of errors (i.e., the button was pressed when 3 was presented or the button was not pressed when it should have been). In the HARMONY 1 and HARMONY 1bis trials, SART was performed 4 times at 2-hour intervals at the inclusion visit and at the end point visit.

Patient Diaries

Patient diaries (electronic or paper) were used in the HARMONY 1 and HARMONY 1bis trials to capture EDS using the ESS score and the number and duration of diurnal sleep and sleepiness episodes, and in the HARMONY CTP trial to capture sleepiness more generally. During the initial visit, patients were instructed by the investigators on the use of the diary and provided a detailed review of definitions for the information to be collected. Patients were required to fill in the sleep diary every morning or evening for each 24-hour period in the 7 or 14 days preceding the scheduled visit.

Clinical Global Impression

The severity of EDS was measured by the investigator with the CGI-S and CGI-C. In the HARMONY 1 and HARMONY 1bis trials, during the 2 baseline visits before randomization, CGI-S was rated by the investigator using a 6-grade scale (no sign of illness, borderline ill, slightly ill, moderately ill, markedly ill, among the most extremely ill patients). At each post-baseline visit, a patient’s change in EDS from baseline was rated by the same investigator using the CGI-C, a 7-grade scale (very much improved, much improved, minimally improved, no change, minimally worse, much worse, very much worse).

In the HARMONY CTP trial, during the 2 baseline visits before randomization, CGI-S was rated by the investigator using a 7-point scale (no sign of illness, borderline ill, mildly ill, moderately ill, markedly ill, severely ill, among the most extremely ill patients). At each post-baseline visit, a patient’s change in EDS from baseline was rated by the same investigator with the CGI-C using a 7-grade scale (very much improved, much improved, minimally improved, no change, minimally worse, much worse, very much worse).

Number, Frequency, and Severity of Cataplexy Attacks

The frequency and severity of cataplexy attacks was a primary end point in the HARMONY CTP trial and a secondary end point in the HARMONY 1 and HARMONY 1bis trials. In all 3 trials, the frequency and severity of complete and partial cataplexy attacks were recorded by patients in sleep diaries, as previously described for EDS. The definitions of total and partial cataplexy attacks were provided to the patient. Cataplexy was defined as a sudden muscle weakness triggered by emotional factors; the patient must have remained fully lucid and aware during the attack. Total cataplexy was defined as a cataplexy attack where all striated muscles were affected, with a loss of posture and falling to the ground. Partial cataplexy was defined as a cataplexy attack that was limited to facial muscles or to the upper or lower limbs, leading to head drop, jaw opening, knees unblocking, or dropping of objects.^14,15 In all 3 trials, the severity of cataplexy was also rated by the investigator using the CGI-S and CGI-C, as previously described for EDS.

Health-Related Quality of Life

In all 3 trials, the 5-Level EQ-5D was used to measured HRQoL. The 5-Level EQ-5D is a validated, self-reported, generic questionnaire of HRQoL.⁴¹ Aspects of HRQoL were assessed with the 5 following dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. A VAS was used by patients to rate their own health state (the best state was rated 100 and the worst state was rated 0). In the HARMONY 1 trial, the 5-Level EQ-5D was completed at the baseline visit, the inclusion visit, and the end point visit. In the HARMONY 1bis trial, the 5-Level EQ-5D was completed at the baseline visit, the inclusion visit, the stable-dose visit, the end point visit, and the withdrawal visit. In the HARMONY CTP trial, the 5-Level EQ-5D was completed at treatment start, the stable-dose visit, and the end point visit.

Nocturnal Sleep Properties

Duration of Nocturnal Awakenings, Diurnal Involuntary Sleep Attacks and Severe Sleepiness, Sleep Paralysis

In the HARMONY 1 and HARMONY 1bis trials, the number and duration of nocturnal awakenings, diurnal involuntary sleep attacks and severe sleepiness, total duration of nocturnal sleep, and incidence of sleep paralysis (defined as being unable to move) were assessed with patients’ sleep diaries, as previously described for EDS. In the HARMONY CTP trial, bedtime and wake-up times were the only sleep properties investigated, also with patients’ sleep diaries.

Polysomnography

In the HARMONY 1bis trial, polysomnography (a standard full-night sleep study) was performed at baseline for the first 20 patients enrolled in 3 selected centres and at the end of the study to evaluate the effect of the interventions on various sleep parameters. The polysomnography was scored manually, according to standard criteria. Episodes of apnea were defined as a complete cessation of airflow for more than 10 seconds. Episodes of hypopnea were defined as a decrease in oronasal airflow of more than 50% lasting for at least 10 seconds, a decrease of more than 30% associated with a decrease in oxygen saturation of more than 3%, or a microarousal.¹⁴

Number of Hallucinations

In all 3 trials, the frequency of hallucinations was evaluated with the patients’ sleep diaries, as previously described for EDS. Patients were asked to record and estimate the number of hallucinations that occurred in the previous 24 hours, defined as “episodes when the patient imagines seeing or hearing people, animals, objects, or frightening events.”^13-15

Patient Satisfaction, Ease of Use

For all trials, at each visit after the onset of the treatment, patients evaluated the global effect of their treatment by comparing the 1-week period before that visit with their condition before study entry. The following 6-level scale was used: marked effect (complete or nearly complete remission of EDS), moderate effect (partial remission of EDS), minimal effect (slight decrease in EDS that does not substantially alter the status of the patient), no change, minimally worse (slight increase in EDS), much worse (substantial increase in EDS). The PGO tool is not validated.

Adherence

In all 3 trials, compliance with treatment was evaluated at each visit during the treatment period. Compliance was not considered an efficacy outcome. Patients were asked to bring back all their used blister packs, including empty ones, at each visit. Details of the quantities of medication dispensed were entered into an accountability form and patients were asked whether they had taken the treatment as prescribed. Any diversions from the prescribed treatment (as reported by patients or evidenced by the blister packs) were recorded.

Harms

The following safety criteria were monitored during the course of all trials: AEs, SAEs, and adverse drug reactions (including frequency, severity, relationship to study drug, incidence, and occurrence). Safety was monitored throughout the trials via changes in vital signs (heart rate, blood pressure, body weight), physical examinations, sleep quality and narcolepsy symptoms based on sleep-diary analysis, electrocardiogram (ECG) parameters, and laboratory abnormalities.

In all 3 trials, the BDI-SF-13 was used to measure depressive symptoms as part of the safety evaluation upon request of the German and Swiss sites to minimize the risk of suicidality.⁴² The protocol was amended accordingly. In the HARMONY 1 trial, patients completed the BDI-SF-13 at all visits. In the HARMONY 1bis trial, patients completed the BDI-SF-13 at the screening visit, baseline visit, inclusion visit, during treatment, at the end of treatment, and after the withdrawal period. In the HARMONY CTP trial, patients completed the BDI-SF-13 at the screening visit, at randomization, before the stable-dose period, at the end of treatment, and after the withdrawal period. In all 3 trials, withdrawal symptoms were evaluated after the withdrawal period using a trial-specific questionnaire. Withdrawal was defined as dysphoria and at least 2 other symptoms, including fatigue, vivid and unpleasant dreams, insomnia or hypersomnia, increased appetite, and psychomotor retardation or agitation.^13-15,23

Statistical Analysis

HARMONY 1 Trial

The primary efficacy criterion was daytime somnolence, assessed using the ESS. To reduce intra-individual variability, ESS score at baseline was calculated as the mean of the pre-baseline visit and end-of-baseline visit values. The final ESS value was calculated as the summary mean of the final 2 visits on treatment. For patients with a premature interruption, the LOCF method was used to impute the missing values. The final ESS value was imputed as the mean of the final 2 available measures, or as the baseline value when no post-baseline values were available. The primary efficacy analysis was the comparison, using a linear mixed-effect model, of the difference in final ESS value between the pitolisant hydrochloride and placebo groups, adjusted for ESS at baseline and using treatment and centre as fixed and random effects, respectively. The significance of pitolisant hydrochloride compared with placebo was assessed by analysis of covariance (ANCOVA) on the final ESS adjusted for ESS at baseline. ANCOVA was conducted with a mixed linear model that took centre heterogeneity into account.

Owing to multiple comparisons of treatments, the multiplicity of type I error for the evaluation of EDS was taken into account using a step-down approach. The test for noninferiority was performed in 2 steps: the null hypothesis (i.e., no difference) for pitolisant hydrochloride versus placebo had to be rejected at an alpha level of 0.025 and the noninferiority of pitolisant hydrochloride to modafinil was only tested if the null hypothesis was rejected; and the noninferiority of pitolisant hydrochloride to modafinil was tested at an alpha level of 0.025. The main ANCOVA model assumed no treatment-baseline interaction term.

For secondary efficacy outcomes, a descriptive analysis for each variable was performed for each treatment group separately. Two-way comparisons between treatment groups were evaluated with ANCOVA, with baseline adjustment on associated baseline values as appropriate. For outcomes involving duration of time, standard survival analysis was undertaken, with adjustment on baseline values as appropriate. For MWT and SART, the significance of treatment difference was tested using the Mann–Whitney U test.

Mean changes from baseline in the daily cataplexy rate were calculated as the ratio of the geometric mean change of the final period from baseline. For patients who had no cataplexy episodes during the baseline or treatment periods, the daily cataplexy rate was imputed as the worst-case value, defined as the reciprocal of the number of days of exposure. The significance of the differences among placebo, pitolisant hydrochloride, and modafinil was tested by comparing the geometric mean of the daily rate of cataplexy among the 3 treatments, and the ratio was tested with a t-test on log-transformed values.

The clinical relevance of the difference between placebo and pitolisant hydrochloride, and between pitolisant hydrochloride and modafinil, for efficacy variables other than EDS was tested by calculating the proportion of patients for whom the increase from baseline to the end of treatment exceeded a predetermined minimum clinical relevance using the absolute risk difference (95% CI). Because BMI, sex, age, and duration of narcolepsy could affect the outcomes, these variables were entered in a stepwise model to identify significant predictors of the studied end point. Any significant predictors identified were added to the main analysis to test the treatment effect.

The determination of sample size was designed under the following hypotheses derived from unspecified historical trials: the minimum clinically relevant difference on ESS was 3, with a SD assumed to be 5 and estimated coefficient of correlation of 0.65, and compound symmetry for the repeated measurements. The NIM was estimated as a small proportion of the difference between the reference and placebo and less than the minimum clinically important difference. The difference, based on meta-analytical results on historical trials of modafinil, was 4.12 (95% CI, 0.14 to 7.09). The NIM of 2 was chosen. Using the power function of ANCOVA, the sample size was determined to achieve a difference of 3 (with the following parameters: 2-sided alpha = 0.05; pre-visits = 2; post-visits = 2; r = 0.65), with a power of at least 95% detected once the sample size exceeded 30 patients per group. Assuming that modafinil and pitolisant hydrochloride have the same efficacy, the probability to reject at a predetermined fixed NIM of 2 (associated with the following parameters: alpha = 0.025; pre-visits = 2; post-visits = 2; r = 0.65) will be at least 80% once the sample size exceeds 30 patients per group. Thus, to satisfy the requirement of the 2 tests, 30 patients per group were planned.

HARMONY 1bis Trial

As in the HARMONY 1 trial, the primary efficacy criterion was daytime somnolence assessed using the ESS, with the same calculations for baseline and final ESS values. The primary efficacy analysis was the comparison, using a linear mixed-effect model, of the difference in final ESS value between the pitolisant hydrochloride and placebo groups, adjusted for ESS at baseline and using treatment and centre as fixed and random effects, respectively. The significance of pitolisant hydrochloride compared with placebo was assessed with ANCOVA on the final ESS adjusted for ESS at baseline. ANCOVA was conducted with a mixed linear model that took centre heterogeneity into account.

Owing to multiple comparisons of treatments, the multiplicity of type I errors for the evaluation of EDS was taken into account using a step-down approach. The test for noninferiority was performed in 2 steps: the null hypothesis (i.e., no difference) for pitolisant hydrochloride versus placebo had to be rejected at an alpha level of 0.05 and the noninferiority of pitolisant hydrochloride to modafinil was only tested if the null hypothesis was rejected; and the noninferiority of pitolisant hydrochloride to modafinil was tested at an alpha level of 0.05. The main ANCOVA model assumed no treatment-baseline interaction term.

For secondary efficacy outcomes, a descriptive analysis of each variable was performed for each treatment group separately. Two-way comparisons between treatment groups were evaluated with ANCOVA, with baseline adjustment on associated baseline values as appropriate. For outcomes involving duration of time, standard survival analysis was undertaken, with adjustment on baseline values as appropriate. For MWT and SART, the significance of treatment difference was calculated as the ratio of the geometric mean change of the final period from baseline. The geometric mean at baseline and the study end was calculated for all treatment groups. The geometric mean was used, as the data were of a log-normal distribution. The mean values of log were compared between treatment groups with a Student t-test and converted as a geometric mean to estimate the geometric mean ratio with the corresponding 95% CI.

Mean changes from baseline in the daily cataplexy rate were calculated as the ratio of the geometric mean change of the final period from baseline. For patients who had no cataplexy episodes during the baseline or treatment periods, the daily cataplexy rate was imputed as the worst-case value, defined as the reciprocal of the number of days of exposure. The significance of the differences among placebo, pitolisant hydrochloride, and modafinil was tested by comparing the geometric mean of the daily rate of cataplexy among the 3 treatments, and the ratio was tested with a t-test on log-transformed values.

The clinical relevance of the difference between placebo and pitolisant hydrochloride, and between pitolisant hydrochloride and modafinil, for efficacy variables other than EDS was tested by calculating the proportion of patients for whom the increase from baseline to the end of treatment exceeded a predetermined minimum clinical relevance using the absolute risk difference (95% CI). Because BMI, sex, age, and duration of narcolepsy could affect the outcomes, these variables were entered in a stepwise model to identify significant predictors of the studied end point. Any significant predictors identified were added to the main analysis to test the treatment effect.

The determination of sample size was informed by the HARMONY 1 trial. The sample-size calculation initially took into account the strategy of ANCOVA under a step-down analysis, using 2 pre-treatment and 2 post-treatment measures. The sample size was determined by separately examining the superiority of pitolisant hydrochloride to placebo with a difference of at least 3, with the following parameters: 2-sided alpha = 0.05; pre-visits = 2; post-visits = 2; r = 0.7; sigma = 5. It would be detected with a power of 95% once the sample size exceeds 20 patients per group. The noninferiority of pitolisant hydrochloride to modafinil was calculated assuming that the 2 drugs had the same efficacy. The probability to refuse noninferiority at a predetermined fixed NIM of 2 (associated with the following parameters: alpha = 0.05; pre-visits = 2; post-visits = 2; r = 0.7; sigma = 5) would be at least 80% once the sample size exceeded 40 patients per group. To satisfy the requirement of the 2 tests, suggested sample sizes of 20 for placebo, 40 for modafinil, and 40 for pitolisant hydrochloride were suggested. The choice of an initial 1:2:2 randomization ratio was reduced because the noninferiority test requires more patients. However, as noninferiority had to be concluded both on ITT and PP bases, it was necessary to increase the sample sizes by 20%. Under these conditions, the sample sizes were initially 25 for placebo, 50 for pitolisant hydrochloride, and 50 for modafinil. An increase in the sample size was necessary after preliminary blind analyses were carried out on the first 50 patients in the HARMONY 1 trial.

HARMONY CTP Trial

The main end point was the measure of anticataplectic efficacy assessed by the change in the average number of cataplexy attacks per week from the 2-week baseline period to the 4-week stable-treatment period. The 2 groups were compared on the WCR. For patients who terminated the study with missing final values, the final value was calculated as the mean of the 2 last known periods. For patients without post-baseline values, the final value was assimilated with baseline (LOCF). For sensitivity purposes, baseline carried forward allocation was also used, meaning that the unknown period for WCR was imputed from the baseline value. The significance of pitolisant hydrochloride compared with placebo was assessed with an ANCOVA on WCR during the 4-week stable-treatment phase adjusted for baseline. ANCOVA was conducted with a mixed nonlinear model featuring a possibly overdispersed Poisson distribution and taking into account centre heterogeneity by using centre as a random factor. In case of overdispersion (a coefficient greater than 2), use of a quasi-Poisson correction was planned. The null hypothesis (i.e., no difference) for pitolisant hydrochloride versus placebo had to be rejected at a 2-sided alpha level of 0.025.

For secondary efficacy outcomes, a descriptive analysis of each variable was performed for each treatment group separately. A 2-way comparison between treatment groups was evaluated with ANCOVA, with baseline adjustment on associated baseline values as appropriate. For outcomes involving duration of time, standard survival analysis was undertaken with adjustment on baseline values as appropriate. For ESS, the treatment effect was tested with ANCOVA after adjustment for baseline severity and centre heterogeneity. The difference between the mean value during baseline (average of first and second visits) and the mean value during the end of treatment period (average of fifth and sixth visits) was calculated. For MWT, the significance of the treatment difference was calculated as the ratio of the geometric mean change of the final period from baseline. The geometric mean at baseline and the study end was calculated for all treatment groups. The geometric mean was used as the data were of a log-normal distribution. The mean values of log were compared between treatment groups with a Student t-test and were converted as a geometric mean to estimate the geometric mean ratio with the corresponding 95% CI.

The sample-size calculation was based on the assumption of a Poisson regression where a rate ratio (ratio of the mean number of cataplexy episodes in pitolisant hydrochloride compared with placebo) was fixed to a rate ratio of 0.75. As the selection of patients is based on a minimum WCR of 3 and because of the expected strong effect of placebo, a placebo effect leading to a reduction of 50% of the baseline value was assumed resulting in a WCR of 1.5 and the corresponding expected mean number of 6 cataplexy crises during the month. Following the sample size calculation, we calculated that an effect of the studied drug was calculated as large as a rate ratio of 0.75 and a mean number of cataplexies of 6 during the studied period (1 month) in the control group should be detected at a 0.05 2-sided level with a power of at least 1 – beta = 0.9 from an equal sample size per group of at least 47.

Subgroup Analysis

Subgroups of patients with a history of cataplexy were reported in the HARMONY 1 trial. Subgroups with a high frequency of cataplexy episodes (> 15) and patients on anticataplectic treatments during the study were reported in the HARMONY CTP trial for the cataplexy frequency outcome.

Analysis Populations

For the HARMONY 1bis and HARMONY CTP trials, the ITT population consisted of all randomized patients who received at least 1 treatment dose and had at least 1 post-dose value to compute the primary end point. However, the definition of ITT used in the trials does not fit the classic definition, as it should include all patients who were randomized, regardless of actual treatment, and was reflected in the extended intention-to-treat (EIT) populations in the trials. For the HARMONY 1 trial, the ITT population consisted of all randomized patients who received at least 1 treatment dose and had missing post-baseline values computed from the baseline value. The ITT population was used for both the primary and secondary analyses in the HARMONY 1 trial. The HARMONY 1 and HARMONY 1bis trials included an EIT population that consisted of all randomized patients, regardless of whether treatment was initiated and irrespective of outcome. For the HARMONY 1 and HARMONY 1bis trials, the PP population consisted of all patients in the ITT population who remained in the study until at least visit 6 (second-to-last visit on treatment) and who had no major protocol deviations related to primary end point. In the HARMONY CTP trial, the PP population was defined as a subset of the ITT sample that included all patients who finished the trial on time according to the protocol, with no major deviations, and in conformity with the prescribed regimen. For all trials, the safety population was composed of all randomized patients who took at least 1 dose of the study drug.

Results

Patient Disposition

Patient disposition in the 3 studies is provided in Table 15 and Table 16. In the HARMONY 1 trial, patients were randomized to placebo, pitolisant hydrochloride, or modafinil for 56 (HARMONY 1 and HARMONY 1bis) and 49 (HARMONY CTP) days. Eighty-six percent of screened patients were randomized. Across all treatment arms, approximately 16% of patients discontinued the study, with AEs and lack of efficacy being the most common reasons. In the HARMONY 1 trial, the 94 patients (31 in the pitolisant hydrochloride group, 30 in the placebo group, and 33 in the modafinil group) in the ITT analysis and the safety population had at least 1 dose of the study drug and provided at least 1 post-baseline value. The EIT population in the HARMONY 1 trial included an additional patient in the pitolisant hydrochloride arm who did not take the study treatment and did not go to the visits after randomization. In the HARMONY 1bis trial, 91% of screened patients were randomized. There were 165 patients in the safety population (33 in the placebo group, 67 in the pitolisant hydrochloride group, and 65 in the modafinil group) and 163 patients in the ITT population (32 in the placebo group, 66 in the pitolisant hydrochloride group, and 65 in the modafinil group). A total of 153 patients completed the study. More patients in the pitolisant hydrochloride group than in the modafinil and placebo groups discontinued the study, with AEs being the most common reason. The EIT population in the HARMONY 1bis trial included an additional patient in the pitolisant hydrochloride arm who had premature withdrawal at day 14 after only 1 treatment. In the HARMONY CTP trial, 91% of screened patients were randomized. A total of 105 patients were treated in the ITT population and the safety population, 54 in the pitolisant hydrochloride group and 51 in the placebo group. Ninety-eight patients completed the study. Discontinuation from the trial was low, at 6% in the placebo group and 8% in the pitolisant hydrochloride group.

Table 15: Patient Disposition in the HARMONY 1 and HARMONY 1bis Trials

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

Sources: Clinical Study Reports for HARMONY 1 and HARMONY 1bis.^13,14

Table 16: Patient Disposition in the HARMONY CTP Trial

ITT = intention-to-treat; PP = per protocol.

Source: Clinical Study Report for HARMONY CTP.¹⁵

Exposure to Study Treatments

In the HARMONY 1 trial, among the 31 patients in the pitolisant hydrochloride group, 8 (26%) were treated for the final 6 weeks at the stable medium dose of 20 mg and 19 (61%) at the stable high dose of 40 mg. Among the 33 patients of the modafinil group, 4 (12%) were treated for the final 6 weeks at the stable medium dose and 24 (73%) at the stable high dose. No evaluation of drug dose, drug concentration, or relationship to response was planned in the HARMONY 1bis trial. The average dose taken in each treatment arm was not reported, with a presumed maximum dose of 20 mg daily. In the HARMONY CTP trial, by the end of the study, 32 (59.3%) of 54 patients in the pitolisant hydrochloride group were receiving 40 mg, 11 (20.4%) were receiving 20 mg, and 7 (13.0%) were receiving 10 mg. The stable dose during the stable-dose period was 40 mg for 35 (64.8%) of 54 patients in the pitolisant hydrochloride group, 20 mg for 9 (16.7%) patients, and 10 mg for 7 (13.0%) patients.

In the HARMONY 1 trial, 33 patients with severe cataplexy were permitted to remain on their anticataplectic medications at stable doses for the duration of the trial; 8 were taking sodium oxybate at the previous stable dosage and the others were taking antidepressants used as anticataplectic medication. Prohibited treatments were continued during the trial in 4 cases. ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||. In the HARMONY CTP trial, 8 (15.7%) patients in the placebo group were treated with an anticataplectic during the study (5 with venlafaxine, 1 with sodium oxybate, 1 with reboxetine, and 1 with escitalopram). There were 4 patients (7.4%) in the pitolisant hydrochloride group treated with an anticataplectic during the study (1 with citalopram, 2 with fluoxetine, 1 with sodium oxybate). Three of 51 patients (5.9%) took at least 1 authorized medication in the placebo group, as did 5 of 54 patients (9.3%) in the pitolisant hydrochloride group.

Table 17: Concomitant Medication Use — HARMONY 1 (ITT Population) and HARMONY 1bis (ITT Population) Trials

ITT = intention-to-treat.

Sources: Clinical Study Reports for HARMONY 1, HARMONY 1bis, and HARMONY CTP.^13-15

Efficacy

Only efficacy outcomes and analyses of subgroups identified in the review protocol are reported here. Refer to Appendix 3 for detailed efficacy data.

Sleepiness, Alertness, and Severity of Daytime Sleepiness

Epworth Sleepiness Scale

HARMONY 1 Trial: The adjusted mean difference in the final ESS score between pitolisant hydrochloride and placebo was –3.10 (95% CI, –5.73 to –0.46; P = 0.022) (Table 18, Figure 5). Sensitivity analyses on the PP population and without accounting for the centre effect showed similar results. Because the superiority of pitolisant hydrochloride to placebo for EDS was demonstrated at the a priori alpha level of 0.025, the noninferiority of pitolisant hydrochloride to modafinil was tested. The adjusted mean difference in the final ESS score between pitolisant hydrochloride and modafinil was 0.09 (95% CI, –2.31 to 2.30); thus, pitolisant hydrochloride was judged to not be noninferior to modafinil at the prespecified NIM of 2.

A patient was considered a responder when the final ESS score was below 10. On this basis, the responder rates were 13.3% in the placebo group, 45.2% in the pitolisant hydrochloride group, and 45.3% in the modafinil group. The adjusted OR of response for pitolisant hydrochloride compared with placebo was 7.86 (95% CI, 1.59 to 38.86). The adjusted OR of response for pitolisant hydrochloride compared with modafinil was 1.09 (95% CI, 0.31 to 3.81).

Maintenance of Wakefulness Test

In the HARMONY 1 trial, the adjusted mean difference in final score between placebo and pitolisant hydrochloride was 1.47 (95% CI, 1.01 to 2.14) and the adjusted mean difference in final score between pitolisant hydrochloride and modafinil was 0.77 (95% CI, 0.52 to 1.13). This was consistent with the findings of the HARMONY 1bis trial, where the adjusted mean difference between placebo and pitolisant hydrochloride was 1.46 (95% CI, 1.069 to 2.010) and the adjusted mean difference in final score between pitolisant hydrochloride and modafinil was ||||||||||||||||||||||||||||||||||||||||||. In the HARMONY CTP trial, the geometric mean of ratios (final divided by baseline) was 1.78 (95% CI, 1.22 to 2.60). Sensitivity analyses for all trials using the PP population were consistent with the main analysis.

Sustained Attention to Response Task

In the HARMONY 1 trial, the adjusted mean difference between the pitolisant hydrochloride and placebo treatment arms was 0.82 (95% CI, 0.67 to 0.99) for NOGO, 0.80 (95% CI, 0.57 to 1.13) for GO, and 0.79 (95% CI, 0.64 to 0.99) for TOTAL SART scores. The adjusted mean difference between the pitolisant hydrochloride and modafinil treatment arms was 1.03 (95% CI, 0.83 to 1.28) for NOGO, 1.03 (95% CI, 0.56 to 1.15) for GO, and 0.90 (95% CI, 0.70 to 1.14) for TOTAL SART scores. Sensitivity analyses using the PP population were consistent with the main analysis.

In the HARMONY 1bis trial, the ratio of mean change between the pitolisant hydrochloride and placebo groups was significant, at 0.83 (95% CI, 0.69 to 0.99; P = 0.043), whereas the difference between the pitolisant hydrochloride and modafinil groups was ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.

CGI-S and CGI-C on EDS

In the HARMONY 1 and HARMONY 1bis trials, CGI-C scores for EDS improved in a higher proportion of patients in the pitolisant hydrochloride and modafinil groups than in the placebo group. However, the change in CGI-C scores was similar in the pitolisant hydrochloride and modafinil arms. In the HARMONY 1 trial, CGI-C scores for EDS improved in the subgroup of patients with a history of cataplexy, but a greater proportion reported an improvement in the modafinil arm. In the HARMONY CTP trial, the mean reduction in CGI-score for pitolisant hydrochloride compared with placebo was –0.95 (95% CI, –1.36 to –0.54; P < 0.0001). Mean CGI-C score was 3.5 (SD = 1.1) with placebo and 2.6 (SD = 1.1) with pitolisant hydrochloride. Similar results were observed in the PP population, with a mean reduction of –0.86 (95% CI, –1.29 to –0.43; P < 0.0001).

Patient Diaries

In the HARMONY 1 trial, the mean drowsiness episodes per day was 0.58 (SD = 1.36), 0.88 (SD = 2.13), and 0.46 (SD = 1.20) at the final visit for placebo, pitolisant hydrochloride, and modafinil, respectively. Drowsiness was not reported in the HARMONY 1bis or HARMONY CTP trials.

Frequency and Severity of Cataplexy Attacks

In the HARMONY 1 trial, the final mean of complete and partial cataplexy episodes (episodes per day) was 0.68 (SD = 1.66), 0.28 (SD = 1.11), 0.65 (SD = 1.62) in the placebo, pitolisant hydrochloride, and modafinil groups, respectively. In the exposed population, the RR of daily rates of complete and partial cataplexy episodes at the end of treatment for pitolisant hydrochloride compared to placebo was 0.38 (95% CI, 0.15 to 0.93). The RR of daily rates of complete and partial cataplexy episodes at the end of treatment for pitolisant hydrochloride compared to modafinil was 0.54 (95% CI, 0.24 to 1.23). In the HARMONY 1bis trial, the mean least squares of daily cataplexy rate for those with cataplexy between the final 7 days of treatment and baseline was |||||||||||||||||||||||||||||||||||||||||||||||||||||||| for pitolisant hydrochloride compared to placebo.

Figure 5: ESS Change in the HARMONY 1 Trial

ESS = Epworth Sleepiness Scale; ITT = intention-to-treat; SD = standard deviation; V = visit.

Source: Clinical Study Report for HARMONY 1.¹³

HARMONY 1bis Trial: By the study end, mean ESS score reductions from baseline (SD) were |||||||||| in the placebo group, |||||||||| in the pitolisant hydrochloride group, and |||||||||| in the modafinil group (Table 18). The adjusted mean difference in the final ESS score for pitolisant hydrochloride compared with placebo was –2.19 (95% CI, –4.17 to –0.22; P = 0.030). Sensitivity analyses without reallocation by centre, without adjustment for baseline ESS, or after adjustment for baseline following the mean change, and the mean change from baseline methods showed similar results. Because the superiority of pitolisant hydrochloride to placebo for EDS was demonstrated at the a priori alpha level of 0.05, the noninferiority of pitolisant hydrochloride to modafinil was tested. The adjusted mean difference in the final ESS score for pitolisant hydrochloride compared with modafinil was 2.75 (95% CI, 1.02 to 4.48); thus, pitolisant hydrochloride was judged to not be noninferior to modafinil at the prespecified NIM of 2.

A patient was considered a responder when the final ESS was 10 or lower or the change from baseline was at least 3 points. The response proportions were |||||||||||||||||||||||||||| for the placebo, pitolisant hydrochloride, and modafinil groups, respectively. The adjusted RR for the difference between pitolisant hydrochloride and placebo ||||||||||||||||||||||||||||||||||||||||||. The adjusted RR for the difference between pitolisant hydrochloride and placebo was ||||||||||||||||||||||||||||||||||||||||||.

HARMONY CTP Trial: The observed mean changes in ESS from baseline were –1.9 (SD = 4.3) and –5.4 (SD = 4.3) in the placebo and pitolisant hydrochloride arms, respectively (Table 19). The adjusted mean difference in the change from baseline for pitolisant hydrochloride compared with placebo was –3.42 (95% CI, –4.96 to –1.87). Sensitivity analyses using LOCF and BOCF and the PP population were consistent with the main analysis. A patient was considered a responder when the final ESS score was 10 or lower or the change from baseline was at least 3 points. The response proportions were 34.0% and 68.6% for placebo and pitolisant hydrochloride, respectively. The adjusted OR for the difference between pitolisant hydrochloride and placebo was 4.26 (95% CI, 1.72 to 10.52).

Table 19: Sleepiness — HARMONY CTP (ITT Population) Trial

CGI-C = Clinical Global Impression of Change; CGI-S = Clinical Global Impression of Severity; CI = confidence interval; EDS = excessive daytime sleepiness; ESS = Epworth Sleepiness Scale; ITT = intention-to-treat; MWT = Maintenance of Wakefulness Test; OR = odds ratio; SD = standard deviation.

Note: P values were not controlled for type I errors (i.e., multiple testing).

^aMean of values at visits 1 and 2.

^bMean of values at visits 5 and 6. For missing values, the LOCF approach was used.

^cLinear mixed model (ANCOVA) adjusted for baseline ESS and for centre heterogeneity (i.e., including centre as a random factor).

^dResponse was defined as a final ESS score ≤ 10 or a change from baseline of ≥ 3 points.

^eLogistic regression model adjusted for ESS at baseline and featured by a mixed nonlinear model, taking into account centre heterogeneity.

^fMissing values were estimated using the linear relationship between MWT and ESS.

^gCalculated as the ratio or geometric means (final/baseline) for pitolisant hydrochloride / ratio of geometric means (final/baseline) for placebo.

^hMeasured at visit 2.

ⁱMeasured at visit 6.

^jLinear mixed-effects model.

^kResponders were patients with CGI-C ≤ 3 at visit 6.

^lNonlinear mixed model taking into account centre heterogeneity.

Source: Clinical Study Report for HARMONY CTP.¹⁵

Table 20: Drowsiness From Sleep Diary — HARMONY 1 (ITT Population) Trial

ITT = intention-to-treat; SD = standard deviation.

^aAll episodes at visit 2 and visit 3 divided by number of days at visit 2 and visit 3.

^bAll episodes at visit 4 to visit 7 and at visit 9 divided by number of days at visit 4 to visit 7 and at visit 9.

Source: Clinical Study Report for HARMONY 1.¹³

Table 21: Cataplexy — HARMONY 1 (ITT Population) and HARMONY 1bis (ITT Population) Trials

CGI-C = Clinical Global Impression of Change; CI = confidence interval; ITT = intention-to-treat; NR = not reported; OR = odds ratio; RR = relative risk; SD = standard deviation.

Note: P values were not controlled for type I error (i.e., multiple testing).

^aBaseline = (all episodes at visit 2 and visit 3) / (number of days at visit 2 and visit 3). For patients with no cataplexy at baseline or during treatment period, imputation value was determined by 0.5/number of days.

^bFinal = (all episodes at visit 7 and visit 9) / (number of days at visit 7 and visit 9).

^cDaily cataplexy rate was calculated as the ratio of the number of crises during 1 period on the number of days of this period. For these calculations, we accounted for the shortness of the exposures in the baseline, treatment, and final periods in the following way: for patients characterized as having no observed crisis during these periods, the rate is, at the most, the reciprocal of the duration (1 / number of days) and 0, at the least; thus, the imputation value was approximated from the mean between the 2 extremes (0.5 / number of days).

^dAnalysis conducted on patients who had at least 1 cataplexy episode at baseline or during study treatment.

^eGeometric mean based on base 10 logarithm of titre.

^fDaily rate of cataplexy for patients with cataplexy at baseline or during treatment, final 7 days: visit 6 to visit 7. Sums of cataplexy equal to 0 have been replaced with 0.1.

^gQuasi-Poisson model on daily cataplexy rate (the ratio final/baseline in geometric mean based on natural logarithm [GMT] of the number of cataplexy episodes on the number of exposed days), final 7 days, adjusted on DCR baseline with treatment considered as a fixed factor and reallocated centre as a random effect. For all the tests, pitolisant hydrochloride was compared with placebo and modafinil with a superiority test.

^hAs measured at the end of the treatment period (visit 7).

^jImproved defined as “very much improved, much improved, or minimally improved.” Worsened defined as “very much worse, much worse, or minimally worse.”

^jCochran-Mantel-Haenszel test (ANOVA statistic) with modified ridit scores.

Sources: Clinical Study Reports for HARMONY 1 and HARMONY 1bis.^13,14

Table 22: Cataplexy — HARMONY CTP (ITT Population) Trial

CGI-C = Clinical Global Impression of Change; CGI-S = Clinical Global Impression of Severity; CI = confidence interval ITT = intention-to-treat; OR = odds ratio; SD = standard deviation; SSRI = selective serotonin reuptake inhibitor.

Note: Aside from the WCR outcome, P values were not controlled for type I errors (i.e., multiple testing).

^aMean of week 1 and week 2 values.

^bMean of values during the stable-dose period (i.e., week 6 through week 9). For patients terminating the trial before completion, the final value was calculated as the mean of the 2 last known periods (LOCF).

^cANCOVA via mixed nonlinear model featuring a possibly overdispersed Poisson distribution and taking into account centre heterogeneity by using centre as a random factor.

^dMeasured at visit 2.

^eMeasured at visit 6.

^fNonlinear mixed model taking into account centre heterogeneity.

^gMeasured at visit 2.

^hMeasured at visit 6.

ⁱLinear mixed-effects model.

^jResponders were patients with CGI-C ≤ 3 at visit 6.

^kNonlinear mixed model taking into account centre heterogeneity.

^lMean ratio of cataplexy rate final / baseline was compared between study treatment and concomitant anticataplectic permitted medication.

Source: Clinical Study Report for HARMONY CTP.¹⁵

The primary end point of the HARMONY CTP trial was the measure of anticataplectic efficacy. The geometric means of the WCR at the end of treatment decreased to 4.51 (95% CI, 2.90 to 7.02) in the placebo group and 2.27 (95% CI, 1.51 to 3.41) in the pitolisant hydrochloride group during the stable-dose period (Figure 6). The ratio of geometric means during the stable-dose period was 0.51 (95% CI, 0.43 to 0.60; P < 0.0001) for pitolisant hydrochloride compared to placebo. Similar results were observed in the PP population, with a ratio of 0.50 (95% CI, 0.34 to 0.74; P < 0.0001) for pitolisant hydrochloride compared to placebo. The effect of pitolisant hydrochloride on the WCR remained consistent at 20 mg and 40 mg doses. The proportion of patients with a high frequency of weekly cataplexy episodes (> 15) during the stable-dose period was 5.6% in the pitolisant hydrochloride group and 17.6% in the placebo group (OR, 0.035; 95% CI, 0.0035 to 0.352). The effect remained consistent regardless of whether patients were taking permitted anticataplectic medications during the trial.

Figure 6: Weekly Cataplexy Rates — HARMONY CTP Trial

CI = confidence interval.

Source: Clinical Study Report for HARMONY CTP.¹⁵

CGI-S and CGI-C on Cataplexy

In the HARMONY 1 trial, the mean final CGI-C score was 3.4 (SD = 1.4), 2.9 (SD = 1.5), 3.0 (SD = 1.6) in the placebo, pitolisant hydrochloride, and modafinil arms, respectively. The number of patients who improved compared to baseline was 6 (24.0%) in the placebo group, 9 (34.6%) in the pitolisant hydrochloride group, and 8 (28.6%) in the modafinil group. The number of patients who reported no change from baseline was 15 (57.7%) in the placebo group, 15 (57.7%) in the pitolisant hydrochloride group, and 16 (57.1%) in the modafinil group. There were 2 (8.0%) patients reporting worsened CGI-C scores in the placebo arm and 1 (3.6%) patient in the modafinil arm.

In the HARMONY 1bis trial, the number of patients who improved from baseline was |||||||||||||| for placebo, |||||||||||||| for pitolisant hydrochloride, and |||||||||||||| for modafinil. The number of patients who reported no change from baseline was |||||||||||||| for placebo, |||||||||||||| for pitolisant hydrochloride, and |||||||||||||| for modafinil. There were |||||||||||||| patients reporting worsened CGI-C scores in the placebo arm, |||||||||||||| in the pitolisant hydrochloride arm, and |||||||||||||| in the modafinil arm.

In the HARMONY CTP group, the mean reduction in CGI-C score for pitolisant hydrochloride compared with placebo was –0.95 (95% CI, –1.36 to –0.54). Mean CGI-C score was 3.5 (SD = 1.1) with placebo versus 2.6 (SD = 1.1) with pitolisant hydrochloride. Similar results were observed for the PP population, with a mean reduction of –0.86 (95% CI, –1.29 to –0.43).

Health-Related Quality of Life

In the HARMONY 1 trial, mean 5-Level EQ-5D scores at baseline were 64.0 (SD = 19.2), 65.3 (SD = 21.3), and 58.7 (SD = 19.4) in the placebo, pitolisant hydrochloride, and modafinil arms, respectively (Table 23). By the end of the treatment period, mean 5-Level EQ-5D scores were 70.2 (SD = 17.7), 43.8 (SD = 17.8), and 72.6 (SD = 16.5) in the placebo, pitolisant hydrochloride, and modafinil groups, respectively. However, there were no differences among the 3 treatment arms and modafinil demonstrated a greater improvement from baseline. In the HARMONY 1bis trial, mean 5-Level EQ-5D scores were |||||||||||||||||||||||||||||||||||||||||||||||||||||||| in the placebo, pitolisant hydrochloride, and modafinil groups, respectively. In the HARMONY CTP trial, no difference between groups was demonstrated for the overall score of this questionnaire (sum of the 5 dimensions) or the corresponding VAS (Table 24).

Table 23: HRQoL — HARMONY 1 (ITT Population) and HARMONY 1bis (ITT Population) Trials

HRQoL = health-related quality of life; ITT = intention-to-treat; SD = standard deviation; VAS = visual analogue scale.

^aMeasured at visit 3.

^bMeasured at visit 7.

Sources: Clinical Study Reports for HARMONY 1 and HARMONY 1bis.^13,14

Table 24: HRQoL — HARMONY CTP (ITT Population) Trial

CI = confidence interval; HRQoL = health-related quality of life; ITT = intention-to-treat; SD = standard deviation; VAS = visual analogue scale; WCR = weekly cataplexy rate.

Note: P values were not controlled for type I errors (i.e., multiple testing).

^aMeasured at visit 2.

^bMeasured at visit 6.

^cLinear mixed-effects model.

Source: Clinical Study Report for HARMONY CTP.¹⁵

Functional Activity

This outcome was not reported outside of the PGO score (refer to the Patient Satisfaction, Ease of Use section).

Mental Health

Refer to the Harms section for the mental health outcomes reported.

Sexual Function

This outcome was not reported.

Sleep Attacks

This outcome was reported as part of the Patient Diaries section, immediately preceding Table 20.

Nocturnal Sleep Properties

The duration of nocturnal awakening was reported in the HARMONY 1 trial (Table 25). All treatment arms reported a final duration of 0.36 hours, or 31.6 minutes (SD = 0.03, 0.05, and 0.08 in placebo, pitolisant hydrochloride, and modafinil groups, respectively). Nocturnal sleep properties were not reported in the HARMONY 1bis trial. In the HARMONY CTP trial, the mean stable-dose period duration was 47.0 (SD = 44.6) minutes for placebo and 41.5 (SD = 58.6) minutes for pitolisant hydrochloride (Table 26). The mean change from baseline to the stable-dose period was –20.3 (SD = 94.6) minutes in the placebo group and –8.2 (SD = 31.5) minutes in the pitolisant hydrochloride group. The adjusted difference in change from baseline was –0.631 (95% CI, –21.939 to 20.677).

Sleep paralysis was decreased across all treatment arms in the HARMONY 1 trial. The final mean number of patients reporting sleep paralysis was 0.03 (SD = 0.11), 0.01 (SD = 0.05), and 0.08 (SD = 0.34) in the placebo, pitolisant hydrochloride, and modafinil groups, respectively.

Polysomnography results are provided in Appendix 3.

Number of Hallucinations

The number of days with hallucinations in the HARMONY 1 trial significantly decreased with pitolisant hydrochloride in comparison to placebo (Table 25). The adjusted ratio rate was 0.46 (95% CI, 0.27 to 0.79). The number of days of hallucinations for patients with any reported hallucinations in the HARMONY CTP trial was significantly lower in the pitolisant hydrochloride arm than in the placebo arm, with an adjusted difference of –0.7766 (–1.3143 to –0.2389).

Table 25: Nocturnal Sleep Properties and Hallucinations — HARMONY 1 (ITT Population) Trial

CI = confidence interval; ITT = intention-to-treat; SD = standard deviation.

^aAll episodes at visit 2 and visit 3 / number of days at visit 2 and visit 3.

^bAll episodes at visit 4 to visit 7 and at visit 9 / number of days at visit 4 to visit 7 and at visit 9.

Sources: Clinical Study Reports for HARMONY 1 and HARMONY 1bis.^13,14

Table 26: Sleepiness and Nocturnal Sleep Properties From Sleep Diary — HARMONY CTP (ITT Population) Trial

CI = confidence interval; ITT = intention-to-treat; SD = standard deviation.

Note: P values were not controlled for type I errors (i.e., multiple testing).

^aMeasured at visit 2.

^bMeasured at visit 7.

^cLinear mixed-effect model adjusted for baseline value.

Source: Clinical Study Report for HARMONY CTP.¹⁵

Concomitant Medication Use

Concomitant medication use was reported for all trials and described in the Exposure to Study Treatments section, which follows Table 35.

Patient Satisfaction, Ease of Use

In the HARMONY 1 trial, patients were classified as improving (PGO score ≥ 3) or not improving (PGO score < 3) on efficacy after 8 weeks of treatment, and improvement in symptoms was reported by 56.0% of patients in the placebo group, 80.8% in the pitolisant hydrochloride group, and 85.7% in the modafinil group (Table 27). The mean final PGO score was 3.1 (SD = 1.4), 2.3 (SD = 1.2), 2.0 (SD = 1.3) in the placebo, pitolisant hydrochloride, and modafinil arms, respectively. The number of patients who reported improvement from baseline was 14 (56.0%) in the placebo group, 21 (80.8%) in the pitolisant hydrochloride group, and 24 (85.7%) in the modafinil group. The number of patients who reported no change from baseline was 7 (28.0%) in the placebo group, 4 (15.4%) in the pitolisant hydrochloride group, and 2 (7.1%) in the modafinil group. Worsened PGO scores were reported by 4 (16.0%) patients in the placebo arm, 1 (3.8%) patient in the pitolisant hydrochloride arm, and 2 (7.1%) patients in the modafinil arm.

||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||.

In the HARMONY CTP trial, more patients noted an effect for efficacy (marked, moderate, or minimal) in the pitolisant hydrochloride group (39 of 50 patients, or 78%) than in the placebo group (13 of 48 patients, or 27%) (Table 28).

Adherence

Overall, more than 91% of the patients in the ITT population had adherence to treatment of at least 80% (Table 29). There was no significant difference between treatment groups with respect to adherence. In the HARMONY 1bis trial, almost all the patients were 80% to 120% adherent, with the exception of 5 (7.6%) patients in the pitolisant hydrochloride group (1 patient who took more than 120% of prescribed treatment and 4 patients who took less than 80%). In the HARMONY CTP trial, only 1 patient in the pitolisant hydrochloride group took less than 80% or more than 120% of the prescribed treatment (Table 30).

Health Care Utilization

This outcome was not reported.

Table 27: PGO — HARMONY 1 (ITT Population) and HARMONY 1bis (ITT Population) Trials