SYSTEMATIC REVIEW. Introduction. L. Citrome

Transcription

1 SYSTEMATIC REVIEW for insomnia: a systematic review of the efficacy and safety profile for this newly approved hypnotic what is the number needed to treat, number needed to harm and likelihood to be helped or harmed? L. Citrome SUMMARY Objective: To describe the efficacy and safety of suvorexant for the treatment of insomnia. Data sources: The pivotal registration trials were accessed by querying and for the search terms suvorexant and MK4305. Briefing documents from the US Food and Drug Administration Peripheral & Central Nervous System Drugs Advisory Committee and product labelling, provided additional information. Study selection: All available clinical reports of studies were identified. Data extraction: Descriptions of the principal results and calculation of number needed to treat (NNT) and number needed to harm (NNH) for relevant dichotomous outcomes were extracted from the available study reports and other sources of information. Data synthesis: (MK4305) is the first orexin receptor antagonist approved for the treatment of insomnia. This approval was based in part on a Phase 3 clinical development programme that included two similarly designed, 3-month, randomised, double-blind, placebo-controlled, parallel-group studies examining suvorexant 40 and 20 mg in non-elderly adults (age < 65 years) and 30 and 15 mg in elderly patients (age 65 years). was superior to placebo for sleep latency as assessed both objectively by polysomnography and subjectively by patient-estimated sleep latency; suvorexant was also superior to placebo for sleep maintenance, as assessed both objectively by polysomnography and subjectively by patient-estimated total sleep time. NNT vs. placebo for response as measured by a 6 point improvement on the Insomnia Severity Index at month 3 was 8 (95% CI 6 14) for both the higher and lower dose regimens. The most commonly encountered adverse event (incidence 5% and at least twice the rate of placebo) as identified in product labelling is somnolence, with NNH values vs. placebo of 13 (95% CI 11 18) for suvorexant 40 and 30 mg, and 28 (95% CI 17 82) for suvorexant 20 and 15 mg. The efficacy and tolerability profile of suvorexant is similar for those < 65 and 65 years of age. Rebound insomnia and withdrawal effects were not observed when suvorexant was discontinued after 3 months or after 12 months of nightly use. Because of concerns about dose-related, next-day effects, including sedation, the recommended dose range is mg. Conclusions: appears efficacious and relatively tolerable. Its different mechanism of action and potentially different safety and tolerability profile compared with currently available hypnotics represents a new option for the pharmacological treatment of insomnia. Review criteria The pivotal registration trials were accessed by querying and for the search terms suvorexant and MK4305. Briefing documents from the US Food and Drug Administration Peripheral & Central Nervous System Drugs Advisory Committee meeting held on 22 May 2013, and product labelling, provided additional information. Message for the clinic is the first orexin receptor antagonist approved for the treatment of insomnia. In two similarly designed, 3-month, randomised, doubleblind, placebo-controlled, parallel-group studies of suvorexant 40 and 20 mg in adults < 65 years, and suvorexant 30 and 15 mg in patients 65 years, suvorexant was superior to placebo for sleep latency and for sleep maintenance. NNT vs. placebo for response as measured by a 6 point improvement on the Insomnia Severity Index at month 3 was 8 (95% CI 6 14) for both the higher and lower dose regimens. The most commonly encountered adverse event (incidence 5% and at least twice the rate of placebo) as identified in product labelling is somnolence, with NNH values vs. placebo of 13 (95% CI 11 18) for suvorexant 40 and 30 mg, and 28 (95% CI 17 82) for suvorexant 20 and 15 mg. The efficacy and tolerability profile of suvorexant is similar for those < 65 and 65 years of age. Rebound insomnia and withdrawal effects were not observed when suvorexant was discontinued after 3 months or after 12 months of nightly use. Because of concerns about dose-related, next-day effects, including sedation, the recommended dose range is mg. represents a new and different option for the pharmacological treatment of insomnia. New York Medical College, Valhalla, NY, USA Correspondence to: Leslie Citrome, 11 Medical Park Drive, Suite 106, Pomona, NY 10970, USA Tel.: Fax: citrome@cnsconsultant.com Disclosures No funding or writing assistance was utilised in the production of this article. In the past 36 months, Leslie Citrome has engaged in collaborative research with, or received consulting or speaking fees from: Alexza, Alkermes, AstraZeneca, Avanir, Bristol- Myers Squibb, Eli Lilly, Forest, Forum, Genentech, Janssen, Jazz, Lundbeck, Merck, Medivation, Mylan, Novartis, Noven, Otsuka, Pfizer, Reckitt Benckiser, Reviva, Shire, Sunovion, Takeda, Teva, and Valeant. As Editor-in-Chief for the Journal, Leslie Citrome withdrew from the review process and deferred all editorial decisions to Matt Rosenberg. Introduction Sleep problems are ubiquitous (1). Difficulty falling and staying asleep can be seen in persons with and without comorbid medical and psychiatric disorders, and it has been estimated that about a third of people have at least one insomnia symptom (2). Persons with hypertension, pulmonary disease, urinary disorders, chronic pain, and gastrointestinal problems have higher levels of insomnia than those without. doi: /ijcp

2 1430 for insomnia these medical disorders (3). In addition, sleep disturbances are part of the diagnostic criteria of several psychiatric disorders, including major depressive disorder, bipolar disorder, post-traumatic stress disorder, as well as others (4). People with insomnia have significantly higher levels of depression and anxiety than people not having insomnia; in one study, people with insomnia were 10 times more likely than people not having insomnia to have clinically significant depression and 17 times more likely to have clinically significant anxiety (5). Savvy clinicians have long used alterations in sleep as an early warning system regarding overall worsening of mental conditions. Pharmacological approaches to the treatment of insomnia have included GABA-A receptor agonists (for example, the benzodiazepine temazepam and the non-benzodiazepine zolpidem), melatonin receptor agonists (for example, ramelteon), and histamine H 1 receptor antagonists (for example, doxepin) (1,6). More recent attention has been paid to the orexin system (7). Orexin (hypocretin) is a peptide produced by about 50,000 80,000 neurons, all located in the posterior lateral hypothalamus, and these neurons project to the same targets that GABA inhibits (GABA is the most common inhibitory neurotransmitter in the human CNS). Orexin leads to diminished activity in sleep-promoting areas, thus enhancing wakefulness, as well as being involved in the stabilisation of the sleep wake cycle. Orexin neuropeptide release follows the circadian rhythm where levels rise with waking and decrease during night. The discovery of orexin being involved with sleep was prompted by studies of narcolepsy; 90% of people with narcolepsy are orexin-deficient (8). The orexin peptides A and B bind selectively to the orexin 1 receptor (OX1R) and orexin 2 receptor (OX2R) (7), and drugs that bind to both OX1R and OX2R are referred to as dual orexin receptor antagonists (DORAs) (7,9). Several DORAs have been investigated as potential treatments for insomnia (9 12), and one of them, suvorexant (MK4305), received approval by the US Food and Drug Administration (FDA) on 13 August 2014 for the treatment of insomnia, as defined by difficulties with sleep onset and/or sleep maintenance (13 15). This was followed on 28 August 2014 by action by the US Drug Enforcement Administration (DEA), placing suvorexant into Schedule IV of the Controlled Substances Act (16). The aim of this review was to synthesise the available data regarding the efficacy and safety of suvorexant for the treatment of insomnia and to place this new hypnotic into clinical perspective. Data sources Clinical trial information was accessed by online query on 14 August 2014 of nih.gov/pubmed/ and for the search terms suvorexant and MK4305. No date or language constraints were utilised. Briefing documents from the FDA Peripheral & Central Nervous System Drugs Advisory Committee meeting held on 22 May 2013 (17 19), and product labelling (14), provided additional information. Study selection Query of the PubMed literature database identified 33 records that contained the search term suvorexant or MK4305. Of these, two were original clinical trial reports for randomised controlled efficacy studies in patients (20,21), and an additional trial was found regarding sleep parameters in healthy men (22). As per the briefing documents available at the FDA website (17 19), three Phase 3 trials were conducted for assessing the efficacy and safety of suvorexant, with two of these trials providing 3-month treatment data for the primary efficacy assessment (unpublished at present) and the third contributing 12-month treatment data (21). Supportive data are available from a Phase 2 trial (20) and safety data are also available from 31 Phase 1 trials. The Phase 2 and Phase 3 suvorexant doubleblind randomised trials in adult patients with insomnia are summarised in Table 1. Query of clinicaltrials.gov revealed nine studies and those protocols not already included in Table 1 are summarised in Table 2; the registry (NCT) numbers are provided in the text at first mention of each relevant trial. All studies of suvorexant reported to date have been conducted by its manufacturer. Data extraction Principal findings are summarised, including time to sleep onset and assessment of sleep maintenance as measured by patient estimate and by polysomnography. Number needed to treat (NNT), and number needed to harm (NNH), comparing suvorexant with placebo, are calculated where appropriate. The methodology of this technique is described in detail elsewhere (23,24). NNT and NNH are measures of effect size and indicate how many patients would need to be treated with one agent instead of the comparator to encounter one additional outcome of interest. Lower NNTs are evidenced when there are large differences between the interventions in question. For example, a NNT of two would be a very large effect size, as a difference is encountered after treating just

3 for insomnia 1431 Table 1 Completed phase 2 and 3 double-blind randomised trials of suvorexant in adult patients with insomnia Author (Registry Identifier and Manufacturer s Identifiers) Phase Length (weeks) n dose (n) N Placebo Comments Herring et al. (20) (NCT , , 2008_583) Unpublished (18) (NCT , 2010_520, MK ) Unpublished (18) (NCT , 2010_521, MK ) Michelson et al. (21) (NCT , , MK , 2009_696) 2 4 weeks mg (62) 20 mg (61) 40 mg (59) 80 mg (61) 3 3 months mg for patients < 65 years old; 30 mg for patients 65 years old (383) 20 mg for patients < 65 years old; 15 mg for patients 65 years old (254) 3 3 months mg for patients < 65 years old; 30 mg for patients 65 years old (387) 20 mg for patients < 65 years old; 15 mg for patients 65 years old (239) 3 12 months mg for patients < 65 years old; 30 mg for patients 65 years old (521) 249 Start date: 11/2008. Completion date: 12/2009 This was a crossover design dose-finding study to evaluate efficacy using sleep efficiency as measured by polysomnography, and safety and tolerability, in non-elderly patients with chronic insomnia. Completion rate was 90% 384 Start date: 05/2010. Completion date: 12/2011 This was a parallel group design pivotal study to evaluate efficacy using patient reported outcomes and polysomnography, and safety and tolerability, in non-elderly and elderly patients with chronic insomnia. Completion rate was 90%. There was a 3-month extension phase that enroled 423 subjects 383 Start Date: 05/2010. Completion Date: 11/2011 This was a parallel group design pivotal study to evaluate efficacy using patient reported outcomes and polysomnography, and safety and tolerability, in non-elderly and elderly patients with chronic insomnia. Completion rate was 87% 258 Start Date: 12/2009. Completion Date: 08/2011 This was a parallel group design long-term safety study in non-elderly and elderly patients with chronic insomnia. Completion rate was 62%. There was a 2-month randomised discontinuation phase where out of 484 subjects, 156 on suvorexant continued suvorexant, 166 on suvorexant went on to placebo, and 162 on placebo remained on placebo

4 1432 for insomnia Table 2 clinical trials registered at (accessed 14 August 2014) Registry Identifier and Manufacturer s Identifiers Study title Length Phase n Interventions Comments NCT , NCT , NCT , NCT , , 2010_505 NCT , 2010_500, MK A Multiple Dose Study to Evaluate Next-Day Effects of MK-4305 on Driving Performance in Healthy Non-Elderly Subjects A Study to Evaluate the Effects of MK-4305 in Patients with Obstructive Sleep Apnea A Study to Evaluate the Effects of MK-4305 in Patients with Chronic Obstructive Pulmonary Disease An Open-Label, Single-Dose Study to Investigate the Pharmacokinetics of MK4305 in Patients with Impaired Renal Function A Single Dose Study to Investigate the Pharmacokinetics of MK4305 in Patients with Hepatic Insufficiency 8 days mg 40 mg Zopiclone 7.5 mg Placebo 4 days mg Placebo 4 days or 40 mg depending on participant age Placebo Single dose study Single dose study Crossover design Start date: 04/2011 Completion date: 11/2011 Crossover design Start date: 03/2011 Completion date: 08/2011 Crossover design Start date: 01/2011 Completion date: 02/ mg Start date: 05/2010 Completion date: 07/ mg Start date: 02/2010 Completion date: 04/2010 Ns as listed in the clinical registry record; does not include studies already listed in Table 1. two patients with one of the interventions vs. the other. A NNT of 50 would mean little difference between the two interventions, as it would take treating 50 patients to encounter a difference in outcome. NNH is used when referring to undesirable events. A useful medication is one with a low NNT and a high NNH when comparing it with another intervention; a low NNT and a high NNH would mean one is more likely to encounter a benefit than a harm. A rule of thumb is that single digit NNTs for efficacy measures suggests that the intervention has potentially useful advantages, and that double digit or higher NNHs for adverse outcomes indicate that the intervention is potentially tolerable (25). Where possible and applicable, for each NNT or NNH, 95% confidence intervals (CI) are computed. For where the NNT or NNH is not statistically significant at the p < 0.05 threshold, the notation ns is shown. Data synthesis Pharmacodynamics and mechanism of action is a highly selective reversible high affinity orexin receptor antagonist at OX1R (Ki 0.55 nm) and OX2R (Ki 0.35 nm) receptors with ~6,000- to 10,000-fold intrinsic selectivity for OX1R and OX2R when tested with over 165 receptors and enzymes in an in vitro screening panel (18). This antagonism of the orexin receptors dampens the activation of the downstream brainstem and cerebral nuclei that mediate wakefulness, thus enabling sleep to occur (18). is chemically unrelated to ethanol, GABA-A receptor agonists (either benzodiazepines or non-benzodiazepines), barbiturates, or other drugs with hypnotic properties, and suvorexant does not possess the intrinsic myorelaxant or anxiolytic properties accompanying interaction with the GABA system (18). In studies in rats, dogs and rhesus monkeys, suvorexant induced rapid sleep and increased non-rapid eye movement (NREM) and REM stage sleep, in contrast to GABA-ergic agonists which increase NREM sleep but decrease REM sleep (18). Pharmacokinetics The pharmacokinetics of suvorexant following single doses up to 240 mg and multiple doses up to 100 mg were evaluated (18). exposure

5 for insomnia 1433 increases in a less than strictly dose-proportional manner over the range of mg because of decreased absorption at higher doses (14), such that suvorexant exposures increase ~75% with a doubling of dose from 20 to 40 mg (18). The pharmacokinetics of suvorexant is similar in healthy subjects and patients with insomnia (14). Oral absorption of suvorexant is relatively rapid, with peak concentrations occurring at a median T max of 2 h (range 30 min to 6 h) under fasted conditions (14). Mean absolute bioavailability of suvorexant 10 mg is 82%. Although ingestion of suvorexant with a high-fat meal resulted in no meaningful change in AUC or C max, there was a delay in T max of approximately 1.5 h. Thus, suvorexant may be taken with or without food but for faster sleep onset, suvorexant should not be administered with or soon after a meal (14). The geometric mean terminal t 1/2 for suvorexant is 12.2 h (5th 95th percentile: 8 19 h) following 40 mg (18). The mean volume of distribution of suvorexant is approximately 49 l (14). is extensively protein bound (99.5%), primarily to serum albumin and a-1 acid glycoprotein, and does not distribute preferentially into red blood cells (18). The primary route of elimination is through the faeces, with approximately 66% of radiolabelled dose recovered in the faeces compared with 23% in the urine (14). is primarily metabolised by CYP3A4 with a minor contribution from CYP2C19 (14,18). M9, a circulating hydroxylated metabolite of suvorexant, is the major metabolite and is not active in vivo (18). Thus, strong inhibitors of CYP3A4 can increase systemic exposure of suvorexant, as observed with the co-administration of ketoconazole and suvorexant, resulting in a mean AUC that is almost twice greater than with suvorexant alone (18). Thus, suvorexant is not recommended for use in patients taking strong CYP3A4 inhibitors, and that the dose should be reduced in the presence of moderate CYP3A4 inhibitors (14). Strong inducers of CYP3A4, such as rifampin, can substantially reduce the systemic exposure of suvorexant. Although suvorexant has the potential to inhibit CYP3A4 and intestinal P-gp, suvorexant is unlikely to cause clinically significant inhibition of human CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19 or CYP2D6, and no clinically meaningful inhibition of OATP1B1, BCRP and OCT2 transporters is anticipated (14). Because of suvorexant s potential for the inhibition of intestinal P-gp, digoxin concentrations should be monitored when co-administering suvorexant with digoxin (14). No dose adjustment of suvorexant is required on the basis of age and race; however, exposure is somewhat higher in females than in males with an AUC and C max that is increased by 17% and 9%, respectively, following administration of 40 mg, with an average concentration of suvorexant 9 h after dosing that is 5% higher for females across the dose range studied (10 40 mg) (14). These small differences do not warrant a dose adjustment based on gender only but apparent oral clearance of suvorexant is also inversely related to body mass index (BMI). In obese patients, the AUC and C max are increased by 31% and 17%, respectively, and the average concentration of suvorexant approximately 9 h after a 20 mg dose is 15% higher in obese patients (BMI > 30 kg/m 2 ) relative to those with a normal BMI (BMI 25 kg/m 2 ). Thus, in obese females, the AUC and C max are increased by 46% and 25%, respectively, compared with non-obese females and this higher exposure to suvorexant in obese females should be considered before increasing dose (18). The effects of renal and hepatic impairment on the pharmacokinetics of suvorexant were evaluated in specific pharmacokinetic studies (18; NCT , NCT ). No dose adjustment is required in patients with mild and moderate hepatic impairment; however, suvorexant t 1/2 was increased to approximately 19 h (range h) in patients with moderate hepatic insufficiency (14). Because suvorexant has not been studied in patients with severe hepatic impairment, it is not recommended for these patients. No dose adjustment is required in patients with renal impairment. Efficacy in clinical trials In 2008, prior to the conduct of controlled clinical trials of suvorexant in persons with insomnia, a polysomnography study was conducted in a sleep laboratory that examined the effects of night-time administration of suvorexant in 22 healthy young men ages years old (22). This was a randomised, double-blind, placebo-controlled, 4-period crossover study, followed by an additional period to assess pharmacokinetics. Interventions tested were suvorexant 10, 50, and 100 mg, and placebo administered 1 h before night-time polysomnography recording. Next-morning residual effects were assessed by psychomotor performance tests and subjective assessments. Statistically significant sleep-promoting effects were observed with all doses of suvorexant compared with placebo. 50 and 100 mg significantly decreased latency to persistent sleep and wake after sleep onset time, and increased sleep efficiency. 10 mg significantly decreased wake after sleep onset time.

6 1434 for insomnia The above study was followed by a Phase 2b twoperiod, 4-week, crossover design, polysomnography study where 254 non-elderly adult insomnia patients received suvorexant at doses of 10, 20, 40, or 80 mg, in one period and placebo in the other period (20; NCT ). Polysomnography was performed on night 1 and at the end of week 4 of each period. All four doses of suvorexant showed significant improvements vs. placebo on the co-primary end-points of sleep efficiency at night 1 and end of week 4; however, these effects appeared dose-related. Dose-related effects were also observed for sleep induction (latency to persistent sleep) and maintenance (wake after sleep onset). The above studies, together with additional Phase 1 studies, led to the design and conduct of two similarly designed, 3-month, randomised, double-blind, placebo-controlled, parallel-group, pivotal Phase 3 studies in primary insomnia examining suvorexant 40 mg as the potential optimal dose for non-elderly adults (age < 65 years) and suvorexant 30 mg as the potential optimal dose for elderly patients (age 65 years) (18; NCT , NCT ). Lower doses were also examined (suvorexant 20 mg in non-elderly patients and suvorexant 15 mg in elderly patients) to further characterise dose response. In both studies, non-elderly and elderly patients were randomised separately (14). Demographics and baseline clinical characteristics are summarised in Table 3a (18). Results of the patient-estimated and polysomnographic assessments pooled across both studies are summarised in Table 3b (18). was superior to placebo for sleep latency as assessed both objectively by polysomnography and subjectively by patient-estimated sleep latency; suvorexant was also superior to placebo for sleep maintenance, as assessed both objectively by polysomnography and subjectively by patient-estimated total sleep time. The efficacy of suvorexant was noted to be similar between women and men and between Caucasians and non-caucasians (14). The magnitude of efficacy effects observed for suvorexant 15 and 20 mg was consistently less than for suvorexant 30 and 40 mg, confirming the presence of a modest dose response (18). However, because significantly more adverse reactions were reported at the higher doses the maximum FDA-approved dose of suvorexant is 20 mg once daily (14). When the results for those persons age 65 years were examined, they were similar to that found in the primary analyses (18). Results from pre-specified responder analyses of suvorexant vs. placebo from the pooled 3-month studies are also available. Table 3c provides the responder data together with NNTs and their respective 95% CIs. The Insomnia Severity Index (ISI) is a 7-item instrument that includes items related to sleep onset, sleep maintenance, and effects of sleep disturbance on daytime functioning, quality of life, and stress levels (26); a threshold of a 6-point improvement was used to categorise ISI-responders (18). ISI total scores range from 0 to 28, with scores 0 7 representing no clinically significant insomnia, 8 14 subthreshold insomnia, clinical insomnia of moderate severity, and clinical insomnia of severe intensity (26). Thus, a 6-point improvement corresponds to a potential one-step categorical improvement in the ISI. NNT vs. placebo for ISIresponse at month 3 was 8 for both suvorexant 15 or 20 mg and suvorexant 30 or 40 mg. Additional responder analyses used a threshold of a 15% improvement on subjective assessments of sleep onset and sleep maintenance, with NNTs vs. placebo at month 3 as robust as 8 (for suvorexant 30 or 40 mg on 15% improvement in mean subjective total sleep time). In general, effect sizes vs. placebo were greater (NNTs smaller) for suvorexant 30 or 40 mg compared with suvorexant 15 or 20 mg, with overlap of the 95% CIs. Supporting the above pivotal 3-month studies is a Phase 3 randomised, placebo-controlled, parallelgroup trial where adults aged 18 years or older with primary insomnia were randomised to receive nightly suvorexant (40 mg for patients younger < 65 years, 30 mg for patients aged 65 years) or placebo at a 2:1 ratio for 1 year with a subsequent 2-month randomised discontinuation phase in which patients on suvorexant either continued suvorexant or were abruptly switched to placebo while patients on placebo remained on placebo (21; NCT ). In this study, efficacy was observed to be maintained over time relative to baseline as well as to placebo. The additional 2-month randomised discontinuation phase of the trial demonstrated that insomnia symptoms returned in patients who were discontinued from suvorexant, but without clinically meaningful rebound or withdrawal, whereas those who continued to receive suvorexant maintained the gains they had made during the year of treatment, providing evidence for the continued benefit of treatment with suvorexant beyond 1 year (18). Tolerability and safety in clinical trials Safety data are available from over 30 Phase 1 studies conducted with suvorexant that assessed safety, tolerability, pharmacokinetics, pharmacodynamics, residual effects, abuse potential and respiratory safety. Additional safety and tolerability information was obtained from the Phase 2 dose-ranging clinical trial (20) and from the Phase 3 clinical programme (18).

7 for insomnia 1435 Table 3a Demographic and baseline clinical characteristics in the two 3-month Phase 3 pivotal studies of suvorexant Study NCT Study NCT Variable 15 or 20 mg (n = 254) 30 or 40 mg (n = 383) Placebo (n = 384) 15 or 20 mg (n = 239) 30 or 40 mg (n = 387) Placebo (n = 383) Male gender, n (%) 92 (36.2) 153 (39.9) 139 (36.2) 82 (34.3) 120 (31.0) 136 (35.5) Age (years) Mean Range Age < 65 years, n (%) 147 (57.9) 222 (58.0) 223 (58.1) 144 (60.3) 229 (59.2) 226 (59.0) Race/ethnicity, n (%) White 168 (66.1) 253 (66.1) 244 (63.5) 190 (79.5) 310 (80.1) 309 (80.7) Black 15 (5.9) 18 (4.7) 25 (6.5) 4 (1.7) 20 (5.2) 21 (5.5) Asian 66 (26.0) 98 (25.6) 99 (25.8) 27 (11.3) 26 (6.7) 25 (6.5) Other 5 (2.0) 14 (3.7) 16 (4.2) 18 (7.5) 31 (8.0) 28 (7.3) Mean subjective total sleep time (min) (n = 252) (n = 383) (n = 384) (n = 238) (n = 386) (n = 383) Mean subjective time to sleep onset (min) 63.3 (n = 252) 68.0 (n = 383) 66.9 (n = 384) 86.0 (n = 238) 74.4 (n = 386) 81.3 (n = 383) Mean subjective wake time after sleep onset (min) 73.9 (n = 252) 78.4 (n = 381) 78.2 (n = 384) 84.8 (n = 233) 82.1 (n = 382) 83.3 (n = 375) Wakefulness after persistent sleep onset (min) (n = 193) (n = 291) (n = 290) (n = 150) (n = 299) (n = 295) Latency to onset of persistent sleep (min) 68.9 (n = 193) 61.8 (n = 291) 66.2 (n = 290) 65.3 (n = 150) 67.3 (n = 299) 68.0 (n = 295)

8 1436 for insomnia Table 3b Patient-estimated and polysomnographic assessment of time to sleep onset and sleep maintenance in the two 3-month Phase 3 pivotal studies of suvorexant (pooled results) Measure Time point Placebo LS mean change from baseline 15 or 20 mg LS mean change from baseline 15 or 20 mg LS mean difference from placebo 30 or 40 mg LS mean change from baseline 30 or 40 mg LS mean difference from placebo Mean subjective total sleep time (min) Mean subjective wake time after sleep onset (min) Wakefulness after persistent sleep onset measured by polysomnography (min) Mean subjective time to sleep onset (min) Latency to onset of persistent sleep measured by polysomnography (min) Week *** *** Month *** *** Month *** *** Week ** *** Month ** *** Month * *** Night *** *** Month *** *** Month *** *** Week *** *** Month * *** Month ** *** Night *** *** Month *** *** Month ** LS, least squares. *p < 0.05, **p < 0.01, ***p < Table 3c Responder analyses for the Insomnia Severity Index total score and patient-estimated assessment of time to sleep onset and sleep maintenance in the two 3-month Phase 3 pivotal studies of suvorexant (pooled results) Measure Time point Placebo responders, %(n/n) 15 or 20 mg responders, %(n/n) 15 or 20 mg vs. placebo NNT (95% CI) 30 or 40 mg responders, %(n/n) 30 or 40 mg vs. Placebo NNT (95% CI) 6 point improvement in Insomnia Severity Index total score 15% improvement in mean subjective total sleep time 15% improvement in mean subjective wake time after sleep onset 15% improvement in mean subjective time to sleep onset Month (157/685) 33.9 (149/440) 10 (7 19) 39.9 (279/699) 6 (5 9) Month (269/638) 55.5 (228/411) 8 (6 14) 54.9 (360/656) 8 (6 14) Week (145/740) 31.3 (150/479) 9 (6 15) 34.8 (262/752) 7 (6 10) Month (210/715) 42.5 (197/463) 8 (6 14) 43.4 (316/728) 8 (6 11) Month (278/664) 50.1 (213/425) 13 (7 46) 54.7 (376/688) 8 (6 14) Week (350/729) 56.3 (267/474) 12 (8 39) 65.1 (486/746) 6 (5 9) Month (414/704) 67.2 (307/457) 12 (8 37) 70.8 (512/723) 9 (6 15) Month (458/660) 75.8 (322/425) 16 (9 102) 77.5 (529/683) 13 (8 30) Week (316/740) 55.7 (267/479) 8 (6 14) 60.0 (451/752) 6 (5 9) Month (384/715) 62.4 (289/463) 12 (7 34) 69.8 (508/728) 7 (5 9) Month (438/664) 69.9 (297/425) 26 (ns) 76.5 (526/688) 10 (7 18) CI, confidence interval; NNT, number needed to treat; ns, not significant. In the polysomnography study conducted in healthy young men there was no evidence of nextday residual effects for suvorexant 10 mg, however, suvorexant 50 mg statistically significantly reduced subjective alertness, and suvorexant 100 mg significantly increased reaction time and reduced subjective alertness (22). The Phase 2b polysomnography study of suvorexant in non-elderly persons with insomnia

9 for insomnia 1437 also demonstrated a dose response regarding the number of patients experiencing one or more adverse events, and for the specific adverse event of somnolence (20). Table 4 displays the number and percentage of subjects with common adverse events ( 2% incidence in any suvorexant group and frequency greater for suvorexant than for placebo) and NNH vs. placebo and 95% CIs, pooled from the 3-month pivotal trials and the first 3 months of the Phase month trial (18,21). All the NNH values are > 10 and most are not statistically significant. In most instances, NNH values are lower for suvorexant 30 or 40 mg compared with suvorexant 15 or 20 mg, with more values being statistically significant, supporting the notion that higher doses of suvorexant may be less tolerable than lower doses. As noted in product labelling, the most common adverse reaction (reported in 5% of patients treated with suvorexant and at least twice the placebo rate) with suvorexant at the recommended dose (15 or 20 mg) was somnolence (14). In the combined 0 3 month assessment of the three Phase 3 studies, somnolence was reported in 10.7% or 6.7% of patients who received suvorexant 30 or 40 mg, or suvorexant 15 or 20 mg, respectively, vs. 3.0% in those receiving placebo (18), yielding a NNH vs. placebo of 13 and 28, respectively. Somnolence typically resolved with continued treatment, was generally mild or moderate in severity, and in the majority of instances did not result in discontinuation of therapy (18). At suvorexant doses of 15 or 20 mg, the incidence of somnolence was higher in females (8%) than in males (3%) (14). s tolerability profile in elderly patients was generally consistent with non-elderly patients (14). The adverse reactions reported during longterm treatment up to 1 year were generally consistent with those observed during the first 3 months of treatment (14). In the combined 0 3 month assessment of the three Phase 3 studies, discontinuation because of any adverse event was 4.7% (48/1025) for placebo, 5.4% (70/1291) for suvorexant 30 or 40 mg, and 3.2% (16/493) for suvorexant 15 or 20 mg (18). No individual adverse reaction led to discontinuation at an incidence 1% (14). Additional safety information Product labelling (14) describes additional safety evaluations, including two trials regarding nextmorning driving performance, four trials examining next-day memory and balance, a middle of the night safety study in elderly people, rebound and withdrawal effects as assessed following discontinuation in the pivotal 3-month studies, three respiratory safety studies, and a dedicated ECG QT study. In the two randomised, double-blind, placeboand active-controlled, four-period crossover studies evaluating the effects of night-time administration of suvorexant on next-morning driving performance 9 h after dosing, 24 healthy elderly subjects ( 65 years old) received suvorexant 15 and 30 mg suvorexant, and 28 non-elderly subjects received suvorexant 20 and 40 mg (14,18; NCT ). Zopiclone 7.5 mg was included as an active control. In both studies, zopiclone demonstrated assay sensitivity, and analyses showed significant effects of zopiclone at all the time points. Clinically meaningful impaired driving performance was observed in some Table 4 Number and percentage of subjects with common adverse events ( 2% incidence in any suvorexant group and frequency greater for suvorexant than for placebo) and number needed to harm vs. placebo and 95% confidence intervals pooled from the 3-month pivotal trials and the first 3 months of the Phase 3 12-month trial Adverse event Placebo, n (%) (n = 1025) 15 or 20 mg, n (%) (n = 493) 15 or 20 mg vs. Placebo NNH (95% CI) 30 or 40 mg, n (%) (n = 1291) 30 or 40 mg vs. Placebo NNH (95% CI) Somnolence 31 (3.0) 33 (6.7) 28 (17 82) 138 (10.7) 13 (11 18) Headache 61 (6.0) 36 (7.3) 74 (ns) 85 (6.6) 158 (ns) Dizziness 29 (2.8) 15 (3.0) 469 (ns) 32 (2.5) Rate lower than for placebo Abnormal dreams 10 (1.0) 9 (1.8) 118 (ns) 27 (2.1) 90 (48 769) Diarrhoea 15 (1.5) 12 (2.4) 103 (ns) 21 (1.6) 613 (ns) Dry mouth 14 (1.4) 9 (1.8) 218 (ns) 36 (2.8) 71 (39 361) Nausea 16 (1.6) 7 (1.4) Rate lower than for placebo 27 (2.1) 189 (ns) Fatigue 18 (1.8) 11 (2.2) 211 (ns) 49 (3.8) 49 (30 139) Upper respiratory infection 12 (1.2) 8 (1.6) 222 (ns) 28 (2.2) 101 (ns) CI, confidence interval; NNH, number needed to harm; ns, not significant.

10 1438 for insomnia non-elderly subjects after either a 20 or 40 mg dose. In contrast, a statistically significant effect was not observed in elderly subjects after a 15 or 30 mg dose. Across these two studies, five subjects (two nonelderly women on suvorexant 40 mg and two nonelderly women on suvorexant 20 mg; one elderly woman on placebo) prematurely stopped their driving tests as a result of somnolence. Thus, product labelling notes that patients using suvorexant 20 mg should be cautioned against next-day driving and other activities requiring full mental alertness and that patients taking lower doses should also be cautioned about the potential for driving impairment because there is individual variation in sensitivity to suvorexant (14). Four placebo-controlled trials evaluated the effects of night-time administration of suvorexant on nextday memory and balance using word learning tests and body sway tests, respectively (14). Three trials showed no significant effects on memory or balance compared with placebo; however, in a fourth trial in healthy non-elderly subjects, there was a significant decrease in word recall after the words were presented to subjects in the morning following a single dose of suvorexant 40 mg, and there was a significant increase on body sway area in the morning following a single dose of suvorexant 20 or 40 mg. An additional double-blind, randomised, placebocontrolled trial evaluated the effect of a single dose of suvorexant on balance, memory and psychomotor performance in healthy elderly subjects after being awakened during the night (14). Zolpidem 5 mg served as an active control (18). As compared with placebo, suvorexant demonstrated minimal impairment on balance; there was a statistically significant increase in body sway area at 1.5 h, but not at 4 and 8 h post-dose. The impairment was less than that produced by zolpidem. increased choice reaction time at 1.5 h post-dose only, whereas no significant increase was observed for zolpidem. There were no statistically significant effects on immediate and delayed word recall with either suvorexant or zolpidem treatment, however, sample size was small (n = 12). Following participation in the two 3-month Phase 3 pivotal trials, rebound insomnia and withdrawal symptoms were assessed following discontinuation of suvorexant; no clear effects were observed on measures of sleep onset or maintenance, and no clear evidence of withdrawal was noted (14). Whether or not suvorexant would be associated with respiratory depression was assessed in a randomised, placebo-controlled, double-blind, crossover trial in healthy non-elderly subjects with one-time doses of suvorexant 40 and 150 mg; no respiratory depressant effect as measured by oxygen saturation or on the apnea/hypopnea index was observed, consistent with what is known about orexin receptor antagonists (14,18). Two additional randomised, placebo-controlled, 2-period, 4-day, crossover studies were conducted, one in patients with mild-to-moderate obstructive sleep apnea, and one in patients with mild-to-moderate chronic obstructive pulmonary disease (14; NCT , NCT ). In patients with obstructive sleep apnea, there were no effects on oxygen saturation observed following single and multiple doses of suvorexant 40 mg; however, a small mean increase in the apnea/hypopnea index was observed on Day 4. In patients with chronic obstructive pulmonary disease, suvorexant 30 or 40 mg had no respiratory depressant effects as measured by oxygen saturation. However, because of the wide inter- and intra-individual variability observed in both studies, clinically meaningful respiratory effects of suvorexant in obstructive sleep apnea or in chronic obstructive pulmonary disease cannot be excluded (14). has not been studied in patients with severe obstructive sleep apnea or severe chronic obstructive pulmonary disease (14). No untoward QTc interval effects were observed in both a thorough QT study with exposure to suvorexant at doses of 60, 150 and 240 mg (14,18), and in the Phase 2/3 trials of suvorexant (18). The suvorexant product label also warns of nighttime sleep-driving and other complex behaviours while out of bed and not fully awake, depression, sleep paralysis, hypnagogic/hypnopompic hallucinations, and cataplexy-like symptoms (14). Among all the patients participating in the Phase 3 clinical development programme, complex sleep behaviours were reported in 2 (0.2%) subjects receiving suvorexant 30 or 40 mg, but in none of the subjects receiving suvorexant 15 or 20 mg, or placebo (18). Hypnagogic/ hypnopompic hallucinations were reported in 2 (0.4%) subjects receiving suvorexant 15 or 20 mg, 5 (0.4%) subjects receiving suvorexant 30 or 40 mg, and in none of the subjects receiving placebo (18). Sleep paralysis was reported in 1 (0.2%) subject receiving suvorexant 15 or 20 mg, 5 (0.4%) subjects receiving suvorexant 30 or 40 mg, and in none of the subjects receiving placebo (18). Cataplexy was not reported for any of the subjects (18). Suicidal ideation was reported in 1 (0.2%) subject receiving suvorexant 15 or 20 mg, 9 (0.7%) subjects receiving suvorexant 30 or 40 mg, and in 1 (0.1%) subject receiving placebo; these events were confounded by a prior history of depression and suicidality, current depression, and/or precipitating life events and there were no reports of suicidal ideation that included intent with

11 for insomnia 1439 a formal plan or actual suicidal behaviours (18). Excessive daytime sleepiness was reported in 3 (0.6%) subjects receiving suvorexant 15 or 20 mg, 20 (1.5%) subjects receiving suvorexant 30 or 40 mg, and in 3 (0.3%) subjects receiving placebo (18). An additional caveat found in product labelling is that because sleep disturbances may be the presenting symptom of an underlying physical or mental condition, failure of insomnia to remit after 7 10 days of treatment will require a thoughtful and comprehensive re-evaluation (14). is contraindicated in persons with narcolepsy (14). Potential for abuse and dependence Because hypnotics used for the treatment of insomnia have been associated with misuse, including the potential for abuse, suvorexant was evaluated for abuse potential, even though suvorexant has a mechanism of action that is distinct from that of benzodiazepine and non-benzodiazepine sedative-hypnotic drugs and has a high degree of selectivity as an antagonist of the orexin receptors (18). Based on nonclinical abuse liability assessments for suvorexant in rats and in non-human primates, suvorexant does not appear to have a potential for abuse or dependence (18). This was further tested in a Phase 1 study where the relative abuse potential of suvorexant 40, 80, and 150 mg, as compared with placebo and zolpidem 15 and 30 mg (a Schedule IV drug in the US), was evaluated in healthy male and female subjects who are recreational poly-drug users with a history of sedative-hypnotic abuse (18). showed greater effects than placebo and similar effects as zolpidem on the primary end-point, Drug Liking on a Visual Analogue Scale. However, on other measures, all doses of suvorexant showed statistically significantly less effect than zolpidem 30 mg. In addition, there was no apparent dose response for suvorexant on positive measures of drug abuse potential, whereas higher dose of zolpidem appeared to have greater effects than the low dose on most measures. When pooled across all doses evaluated, the incidences of any abuse potential adverse event were 30.6% for suvorexant and 58.3% for zolpidem. In the suvorexant Phase 3 programme, abuse potential was assessed by spontaneously reported adverse events associated with the potential for abuse (e.g., depersonalisation, derealisation, dissociation, euphoric mood, hallucination, mania, potential study medication misuse). Abuse potential adverse events were reported in 20 (4.1%) subjects receiving suvorexant 15 or 20 mg, 34 (2.6%) subjects receiving suvorexant 30 or 40 mg, and in 31 (3.0%) of the subjects receiving placebo (18). Thus, based on the above and the lack of withdrawal, suvorexant does not appear to be associated with a significant risk for abuse in the treatment of insomnia for most patients (18). However, there remains a concern that individuals with a history of abuse or addiction to alcohol or other drugs may be at increased risk for abusing suvorexant (14), and thus suvorexant has been classified as a Schedule IV drug by the US DEA (16). Administration and optimal dose is available in 5, 10, 15, and 20 mg tablet strengths. As per an addendum to Merck s briefing document (19), the original dosing recommendation was to initiate suvorexant therapy at 40 mg (30 mg in elderly) with a reduction to 20 mg (15 mg in elderly) for patients based on individual tolerability. This recommendation was based on the design of the Phase 3 programme and the pre-specified statistical analysis plan but after further consideration, including the recent labelling change for zolpidem dosing, dosing recommendations were revised downwards and further refined in order to address FDA concerns about dose-related, next-day effects, including sedation. As per product labelling (14), the recommended dose for suvorexant is 10 mg, taken no more than once per night and within 30 min of going to bed, with at least 7 h remaining before the planned time of awakening. If the 10 mg dose is well-tolerated but not effective, the dose can be increased. The maximum recommended dose of suvorexant is 20 mg once daily. As noted earlier, time to effect of suvorexant may be delayed if taken with or soon after a meal. Because exposure to suvorexant is higher in persons who are obese as well as in women, this should be considered when determining the dose. In addition, there may be possible additive effects with other CNS depressant medications which would necessitate dose adjustments. Because suvorexant is metabolised primarily through CYP3A4, the recommended starting dose of suvorexant in the presence of moderate CYP3A4 inhibitors should be 5 mg, with a maximum dose of 10 mg, and suvorexant should be avoided in the presence of strong CYP3A4 inhibitors. Likelihood to be helped or harmed To be effective in real-world clinical practice, medications must be efficacious enough, tolerable enough and the patient has to take it consistently enough. Likelihood to be helped or harmed (LHH) can be calculated to quantify trade-offs between efficacy and tolerability, and is simply the ratio of NNH to NNT (25,27). For example, NNT vs. placebo for ISIresponse for suvorexant at the recommended maximum dose of 15 or 20 mg is 8 and the NNH vs. placebo for the adverse event of somnolence is

12 1440 for insomnia 28. LHH = NNH/NNT = 28/8 = 3.5. Thus, suvorexant 15 or 20 mg is 3.5 times more likely to result in a therapeutic response rather than an adverse event of somnolence. An assumption is that the two outcomes being compared in this manner are independent of each other, which may not be the case for ISI-response and somnolence. It is possible to calculate the LHH for any well matched and relevant outcomes for which one has statistically significant values for NNT and NNH. High LHH values are desirable. For harms that are time-limited and easily managed, lower LHH values may be acceptable (25). Discussion appears efficacious and relatively tolerable. Its mechanism of action differs from that of benzodiazepine and non-benzodiazepine hypnotics in that it has no effect on GABA rather than promoting sleep, suvorexant inactivates wakefulness via antagonism of the orexin receptors OX1R and OX2R (28,29). This different mechanism of action and potentially different safety and tolerability profile may facilitate the longer term pharmacological treatment of insomnia. Reassuringly, the efficacy and tolerability profile of suvorexant is similar for those < 65 and 65 years of age. Next-day drowsiness appears limited at the lower, approved, doses. Moreover, rebound insomnia and withdrawal effects were not observed when suvorexant was discontinued after 3 months or after 12 months of nightly use. s half-life is approximately 12 h and is thus considered to be on the borderline between intermediate and long; although half-life is a strong predictor of duration of action and hence residual effects, the relationship is not always consistent (30). Several limitations to this systematic review need to be made explicit. Available so far are the results of carefully conducted registration trials that enroled subjects who fulfilled restrictive inclusion/exclusion criteria. Such patients may differ from persons encountered in routine clinical practice. For example registration trials ordinarily exclude patients with current comorbid psychiatric disorders (including substance use disorders), untreated medical conditions, current suicidality, or who require concomitant use of other psychotropic agents. Thus, pragmatic clinical trials that are more generalisable will help place suvorexant into clinical perspective for its use in the real world. The outcomes used in the NNT calculations are all from subjective assessments and the pre-specified thresholds that defined response ( 6 point improvement on the ISI or 15% improvement on other subjective measures of time related to sleep onset and sleep maintenance) are arbitrary. Although a 6-point change on the ISI can represent a categorical shift in insomnia severity, it is not clear how important these differences are to patients and prescribers. The conduct of head-to-head randomised controlled trials of suvorexant vs. other hypnotics will be necessary in order to directly compare both efficacy and tolerability. At present there are no universally accepted dichotomous definitions of response to hypnotic medications that would make it simple to make indirect comparisons using NNT (vs. placebo) between suvorexant and medications such as benzodiazepine and non-benzodiazepine hypnotics, melatonin receptor agonists, antidepressants, and other agents (31 38). Unanswered questions include comparisons of suvorexant vs. other agents in terms of speed and durability of response, and effects on daytime fatigue, attention and mood irritability (39). Among options other than suvorexant, a meta-analysis of 31 randomised controlled hypnotic drug trials that included a total of 3820 subjects demonstrated higher effect sizes on continuous measures such as sleep onset latency and total sleep time for benzodiazepine receptor agonists and classical benzodiazepines compared with antidepressants (including low-dose doxepin) and for classical benzodiazepines compared with benzodiazepine receptor agonists (40); how suvorexant would compare is not yet known. At the preliminary stage are several meta-analytic reviews of zaleplon, zopiclone, zolpidem, and eszopiclone, where the protocols were published online in August 2013 in the Cochrane Database of Systematic Reviews, but results are not yet available. Comparisons should also be made with psychological and behavioural treatments for chronic insomnia, many of which are supported by a robust evidence-base (41). Conclusions is the first orexin receptor antagonist to receive regulatory approval for the treatment of insomnia. Efficacy is supported by the data collected by polysomnography and from subjective report. s different mechanism of action and potentially different safety and tolerability profile compared with currently available hypnotics represents a new option for the pharmacological treatment of insomnia.