Narcolepsy with or without cataplexy in adults: pitolisant

Clinical effectiveness

This evidence summary is based on 2 double-blind RCTs.

The first 8‑week RCT compared pitolisant with modafinil or placebo in 95 adults with narcolepsy with or without cataplexy; 81% had a history of cataplexy (HARMONY I; Dauvilliers et al. 2013). Participants had to have an Epworth Sleepiness Scale score of 14 or more (mean baseline score was about 18) and were withdrawn from any previous psychostimulants (45% had taken these previously). Participants could remain on sodium oxybate or non-tricyclic antidepressants at stable doses (35% remained on these). See evidence tables for details.

Pitolisant (as hydrochloride) was given at a flexible dose of 10 mg, 20 mg or 40 mg per day (equivalent to 9 mg, 18 mg or 36 mg per day of pitolisant) for 3 weeks, followed by 5 weeks of stable dosing. Modafinil was given at a flexible dose of 100 mg, 200 mg or 400 mg per day for 3 weeks, then again at a stable dose for 5 weeks. Following the 8‑week treatment phase all participants received placebo for 1 week (the withdrawal phase). During the stable dose phase of the trial, 61% of participants were taking pitolisant 40 mg per day, 26% were taking 20 mg per day and 7% were taking 10 mg per day. In the modafinil treatment arm, 73% were taking 400 mg per day, 12% were taking 200 mg per day and 6% were taking 100 mg per day (EPAR for pitolisant).

The second 7‑week RCT compared pitolisant with placebo in 106 adults with a diagnosis of narcolepsy with cataplexy (HARMONY-CTP; Szakacs et al. 2017). Participants had to have 3 or more cataplexy attacks per week (mean baseline was 11 per week in the pitolisant group and 9 per week in the placebo group) and an Epworth Sleepiness Scale score of 12 or more (mean baseline score was 17 in both groups). Any previous psychostimulants were withdrawn, but participants could remain on sodium oxybate or antidepressants at stable doses for cataplexy. In the pitolisant group, 41% of people were taking cataplexy medication before the trial and 7% continued with this throughout the trial. In the placebo group, 80% of people were taking cataplexy medication before the trial and 16% continued this. See evidence tables for details. Pitolisant (as hydrochloride) was given at a flexible dose of 5 mg, 10 mg or 20 mg per day (equivalent to 4.5 mg, 9 mg or 18 mg per day of pitolisant) for 3 weeks, followed by 4 weeks of stable dosing at 5 mg, 10 mg, 20 mg or 40 mg per day. Following the 7‑week treatment phase all participants received placebo for 1 week (the withdrawal phase). During the stable dose phase of the trial, 35 participants were taking pitolisant 40 mg per day, 9 participants were taking 20 mg per day and 7 participants were taking 10 mg per day.

Safety and tolerability

Pitolisant is contraindicated in people with severe hepatic impairment and should be administered with caution in people with moderate hepatic impairment or renal impairment, with the dosing regimen adapted (see dosing information; pitolisant summary of product characteristics [SPC]).

Pitolisant should be administered with caution in people with a history of psychiatric disorders (such as severe anxiety or severe depression with suicidal ideation risk); in people with acid-related gastric disorders or when co-administered with gastric irritants (such as corticosteroids or NSAIDs); in people with severe obesity or severe anorexia; and in people with severe epilepsy. There is a risk of mild to moderate prolongation of QTc interval with supra-therapeutic doses of pitolisant, therefore caution and monitoring is required in people with cardiac disease; those taking other QT-prolonging medicines; those taking medicines that are known to increase pitolisant levels (for example CYP2D6 inhibitors, such as paroxetine); or people with severe renal or moderate hepatic impairment. No rebound effect was reported during clinical trials of pitolisant but treatment discontinuation should be monitored. See the pitolisant SPC for details.

Women of childbearing potential have to use effective contraception during treatment and for at least 21 days after treatment discontinuation. Pitolisant may reduce the effectiveness of hormonal contraceptives, therefore an alternative method of effective contraception should be used. Pitolisant may induce CYP3A4 and CYP2B6 enzymes, therefore its use with substrates of these that have a narrow therapeutic margin should be avoided. Pitolisant levels may also be decreased when it is co-administered with potent CYP3A4 inducers (pitolisant SPC).

The most frequent adverse drug reactions reported in the pitolisant SPC are insomnia (8.4%), headache (7.7%), nausea (4.8%), anxiety (2.1%), irritability (1.8%), dizziness (1.4%), depression (1.3%), tremor (1.2%), sleep disorders (1.1%), fatigue (1.1%), vomiting (1.0%), vertigo (1.0%), dyspepsia (1.0%), weight increase (0.9%), and upper abdominal pain (0.9%). The most serious adverse drug reactions are abnormal weight decrease (0.09%) and spontaneous abortion (0.09%).

The EPAR discusses that the long-term safety data for pitolisant in people with narcolepsy are limited. In HARMONY III (NCT01399606) an unpublished open-label, 12‑month safety study, only 10 people received a 20 mg daily dose and 87 people received a 40 mg daily dose. As narcolepsy is an orphan disease, clinical studies include small numbers of people and their ability to detect rare adverse reactions or adverse reactions due to prolonged exposure is low. The EPAR states that some uncertainties remain with regard to the effects of pitolisant on depression, weight and appetite, ulcer formation, and more generally on adverse events that might occur after long-term exposure.

An overview of the results for the safety and tolerability of pitolisant in HARMONY I and HARMONY-CTP can be found in the results tables.

In HARMONY I, adverse events occurred in 71% (22/31) of the pitolisant group, 79% (26/33) of the modafinil group and 33% (10/30) of the placebo group. The most frequent adverse events were headache for all 3 groups; insomnia, abdominal discomfort and nausea for the pitolisant group; and abdominal discomfort, nausea, diarrhoea, dizziness, anxiety and irritability for the modafinil group. Serious adverse events occurred in 1 person in the pitolisant group (abdominal pain), 5 people in the modafinil group (1 case each of abdominal pain, abnormal behaviour, withdrawal symptoms, lymphoadenopathy and inner ear disorder), and no people in the placebo group. No participants in the placebo or pitolisant group had withdrawal syndrome during the withdrawal phase, compared with 3 participants in the modafinil group.

In HARMONY‑CTP, adverse events occurred in 35% (19/54) of the pitolisant group and 31% (16/51) of the placebo group (p=0.528). These were considered treatment related in 28% (15/54) of the pitolisant group and 12% (6/51) of the placebo group (p=0.048). The most frequent adverse events were headache for both groups; irritability, anxiety and nausea for the pitolisant group; and somnolence for the placebo group. Severe adverse events occurred in 1 person in the pitolisant group (nausea) and no people in the placebo group. No participants in the pitolisant group had withdrawal syndrome during the withdrawal phase, compared with 1 participant in the placebo group.

Evidence strengths and limitations

An overview of the quality assessment of the included studies HARMONY I (Dauvilliers et al. 2013) and HARMONY-CTP (Szakacs et al. 2017) can be found in the evidence tables. These RCTs had several limitations including small patient numbers (n=95 and n=106) and a short duration (8‑week or 7‑week treatment phase), which limits findings related to long-term efficacy and safety in particular. The authors of both studies state that flexible dosage and multiple visits could have affected efficacy, with less responsive people being titrated to the highest dose. There is also a concern about the limited generalisability of the findings because children, those with unstable comorbidities, and those who declined to participate because they didn't want to receive placebo were excluded. In both studies, people had to discontinue psychostimulants before baseline, but could remain on stable doses of sodium oxybate or antidepressants for treating cataplexy. About 35% of people continued on these medicines in HARMONY I; and 7% (in the pitolisant group) and 16% (in the placebo group) in HARMONY-CTP, therefore the findings may not be generalisable to people who are not taking any other medicines for narcolepsy or cataplexy.

The method of randomisation in the double-blind RCTs suggests allocation was concealed and a double-dummy design was used to ensure HARMONY I remained double-blind. However, people who had received modafinil before in HARMONY I (between 33% and 43% of participants) may have realised they were taking it because of its effects.

HARMONY I was designed to assess the superiority of pitolisant compared with placebo, and the non-inferiority of pitolisant compared with modafinil. However, the EPAR states that the trial protocol was amended from an original design to assess the superiority of pitolisant compared with modafinil.

The primary outcome of HARMONY I was the change in excessive daytime sleepiness measured by the Epworth Sleepiness Scale. This is a subjective, self-administered questionnaire, but it is widely used in narcolepsy trials, and the EPAR states it is suitable when combined with an objective secondary end point such as the maintenance of wakefulness test. Another of the secondary end points, the modified CGI‑C is a non-validated measure in narcolepsy. The occurrence of withdrawal syndrome after treatment ended was assessed, but the authors state that early withdrawal effects might have been missed if they were not recalled or reported by participants and if the scale used was not sensitive enough.

The primary outcome of HARMONY-CTP was the change in weekly cataplexy rate, as reported in patient diaries. Participants were asked to report all cataplexy attacks defined as sudden and transient episodes (ranging from several seconds to a few minutes) of partial or generalised loss of muscle tone triggered by emotion. A high placebo effect was found on cataplexy rate reduction, which might reflect subjective and emotional triggers.

The authors of HARMONY-CTP state that baseline characteristics were similar in the 2 groups. However, there were numerical differences in the mean number of cataplexy episodes per week (11 in the pitolisant group; 9 in the placebo group) and proportion of people who had taken cataplexy medicines in the previous 3 months (41% in the pitolisant group at baseline with 7% continuing during the trial; 80% in the placebo group at baseline with 16% continuing during the trial); no analysis reported.

A commentary accompanying the publication of HARMONY‑CTP (Baumann 2017) suggests the main limitation of this study was the use of non-specific criteria for the diagnosis of narcolepsy. This limitation would also apply to HARMONY I. In both RCTs, narcolepsy was diagnosed using the International Classification of Sleep Disorders‑2 (ICSD‑2), rather than the more recent ICDS‑3, because the trials were started before this was published. See Sateia 2014 for more information on diagnostic criteria.

This evidence summary is based on HARMONY I and HARMONY‑CTP because these studies have been published. However, the EPAR for pitolisant states that HARMONY I and the unpublished study HARMONY Ibis were considered pivotal for the indication in the treatment of narcolepsy with or without cataplexy.

According to the EPAR, HARMONY Ibis was an RCT with a similar trial design to HARMONY I, with an 8‑week treatment phase and a 1‑week withdrawal phase. However, the dose of pitolisant was lower at a range of 5 mg to 20 mg per day rather than 10 mg to 40 mg per day. In HARMONY Ibis, 166 people were randomised to either placebo (n=33), pitolisant 5 mg to 20 mg per day (n=67) or modafinil 100 mg to 400 mg per day (n=66). For the primary outcome, the change in excessive daytime sleepiness measured by the Epworth Sleepiness Scale, pitolisant was not superior to placebo (mean difference −1.94 points; 95% CI −4.05 to 0.07, p=0.065). There was a mean decrease of 4.6 points from a baseline of 18.3 points in the pitolisant group, a mean decrease of 3.6 points from a baseline of 18.2 points in the placebo group and a mean decrease of 7.8 points from baseline of 18.1 points in the modafinil group.

When non-inferiority compared with modafinil was considered, pitolisant was not established as non-inferior to modafinil based on the pre-specified non-inferiority margin of −2 points (mean difference −2.75 points; 95% CI −4.48 to −1.02). The EPAR states that HARMONY Ibis was most likely underpowered, and the low dose may not have been sufficient for many people. In HARMONY I, 61% of participants were taking pitolisant 40 mg per day during the stable dose phase of the trial.