sign in sign up
<<
Home , Niaprazine

Accepted Manuscript

Efficacy and Safety of Pediatric Prolonged-Release for in Children With Autism Spectrum Disorder

Paul Gringras, MD, MBA, Tali Nir, DVM, John Breddy, MSc, Anat Frydman-Marom, PhD, Robert L. Findling, MD, MBA PII: S0890-8567(17)31672-6 DOI: 10.1016/j.jaac.2017.09.414 Reference: JAAC 1959

To appear in: Journal of the American Academy of Child & Adolescent Psychiatry

Received Date: 13 July 2017 Revised Date: 28 August 2017 Accepted Date: 3 September 2017

Please cite this article as: Gringras P, Nir T, Breddy J, Frydman-Marom A, Findling RL, Efficacy and Safety of Pediatric Prolonged-Release Melatonin for Insomnia in Children With Autism Spectrum Disorder, Journal of the American Academy of Child & Adolescent Psychiatry (2017), doi: 10.1016/ j.jaac.2017.09.414.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT

Efficacy and Safety of Pediatric Prolonged-Release Melatonin for Insomnia in Children With Autism Spectrum Disorder RH: Pediatric PR Melatonin for Sleep in ASD

Paul Gringras, MD, MBA; Tali Nir, DVM; John Breddy , MSc; Anat Frydman-Marom, PhD; Robert L. Findling, MD, MBA Supplemental material cited in this article is available online. Accepted September 11, 2017 Dr. Gringras is with Children's Sleep Medicine, Evelina London Children's Hospital, Guy's and St Thomas', London. Drs. Nir and Frydman-Marom are with Neurim Pharmaceuticals Ltd, Tel Aviv, Israel. Mr. Breddy is with Pharmastat Consulting Ltd, Canterbury, UK. Dr. Findling is with Kennedy Krieger Institute/Johns Hopkins University, Baltimore, MD. The study was funded by Neurim Pharmaceuticals. Mr. Breddy served as the statistical expert for this research and was paid by Neurim Pharmaceuticals. The authors would like to thank the study participants and their families for their cooperation and commitment. They thank all members of the trial team for their active contribution. They express their sincere gratitude to Amnon Katz, PhD, of the MANUSCRIPTUniversity of Jerusalem, for his help with the study design and research and Shiri Shahmoon, PhD, of the University of Tel-Aviv, for her help with the conduct of the study, data collection, and analysis. Moreover, a sincere gratitude to all investigators that participated in the study: Beth Malow, MD, MS, of Vanderbilt University Department of Neurology, Nashville, TN; Athanasios Maras, MD, PhD, of Yulius Mental Health Organization, Barendrecht; Roshan Raja, DO, of the Clinical Research Center of Nevada; Markku Partinen, MD, PhD, of Päijät-Häme Central Hospital, Helsinki; Amanda Bennett, MD, MPH, of Children's Hospital of Philadelphia; Samantha Bostrom, MD, of Ericksen Research & Development, Clinton, UT; Karen Hillock, MD, Lake Mary Pediatrics, FL; Carmen Schroder, MD, PhD, of Strasbourg University Hospital Department of Child Psychiatry and Neurology; Raun Melmed, ACCEPTED MD, of Southwest Autism Research & Resource Center, Scottsdale, AZ; Lawrence Ginsberg, MD, Red Oak Psychiatry Association, Houston, TX; Thomas Challman, MD, Geisinger Health Systems, Union County, PA; Catherine Hill, MD, University Hospital Southampton NHS Foundation Trust; Megan Thomas, MD, PhD, Blackpool Victoria Hospital NHS Trust; Marcel G. Smits, MD, PhD, Ziekenhuis Gelderse Vallei, Willy Brandtlaan; Jerry ACCEPTED MANUSCRIPT

Tomasovic, MD, Road Runner Research, San Antonio, TX; Ralph Gallo, MD, Clinical Research Center, NJ; Laszlo Mate, MD, of Semmelweis University, FL; Angel Rico, MD, Crystal Biomedical Research, FL; Rajinder Shiwach, MD, Insite Clinical Research, Dallas, TX; Jean- Marc Pinard, MD, Hôpital Raymond Poincaré, Paris; Chetana Kallappa, MD, Birmingham Children’s Hospital NHS Trust. Disclosure: Dr. Gringras has acted as chief investigator and consultant for Neurim Pharmaceuticals. Dr. Nir is an employee of Neurim Pharmaceuticals. Dr. Frydman-Marom is an employee of Neurim Pharmaceuticals. Dr. Findling has received research support from, acted as a consultant for, and/or served on a speaker’s bureau for Akili, Alcobra, the American Academy of Child and Adolescent Psychiatry, American Psychiatric Press, Bracket, Epharma Solutions, Forest, Genentech, Guilford Press, Ironshore, Johns Hopkins University Press, KemPharm, Lundbeck, Medgenics, Merck, the National Institutes of Health, Neurim Pharmaceuticals (also chief investigator), the Patient-Centered Outcomes Research Institute, Pfizer, Physicians Postgraduate Press, Purdue, Roche, Sage, Shire, Sunovion, Supernus Pharmaceuticals, Syneurx, Takeda, Teva, TouchPoint, Tris, Validus, and WebMD. Mr. Breddy has served as statistician for the work on this manuscript and was paid by Neurim Pharmaceuticals. Correspondence to Paul Gringras, MD, MBA, Children'sMANUSCRIPT Sleep Medicine, Evelina London Children's Hospital, Guy's and St Thomas', Westminster Bridge Road, London, UK; email: [email protected]

ACCEPTED

ACCEPTED MANUSCRIPT ABSTRACT Objective: To assess the efficacy and safety of novel pediatric-appropriate prolonged-release melatonin minitablets (PedPRM) vs. placebo for insomnia in children with autism spectrum disorder (ASD), with or without attention-deficit/hyperactivity disorder (ADHD) comorbidity, and neurogenetic disorders (NGD). Method: 125 children (2-17.5 years; 96.8% ASD, 3.2% Smith-Magenis syndrome [SMS]) whose sleep failed to improve on behavioral intervention alone were randomized (1:1 ratio), double-blind, to receive PedPRM (2mg escalated to 5mg) or placebo for 13 weeks. Sleep measures included the validated caregivers’ Sleep and Nap Diary (SND) and Composite Sleep Disturbance Index (CSDI). The a priori primary endpoint was SND-reported total sleep time (TST) after 13 weeks of treatment. Results: The study met the primary endpoint: after 13 weeks of double-blind treatment, participants slept on average 57.5 minutes longer at night with PedPRM compared to 9.14 with placebo (adjusted mean treatment difference PedPRM–placebo -32.43 minutes; p= .034). Sleep latency (SL) decreased by 39.6 minutes on average with PedPRM and 12.5 with placebo (adjusted mean treatment difference -25.30 minutes; p= .011) without causing earlier wakeup time. Rate of participants attaining clinically meaningful responses in TST and/or SL was significantly higher with PedPRM than placebo (68.9% vs 39.3% respectively; p= .001) corresponding to number needed to treat (NNT) 3.38. Overall sleep disturbance (CSDI) tended to decrease. PedPRM was generally safe; MANUSCRIPTsomnolence was more commonly reported with PedPRM than placebo. Conclusion: PedPRM was efficacious and safe for treatment of insomnia in children with ASD with/without ADHD and NGD. The acceptability of this pediatric formulation in a population who usually experience significant difficulties in swallowing was remarkably high. Clinical trial registration information—Efficacy and Safety of Circadin in the Treatment of Sleep Disturbances in Children With Neurodevelopment Disabilities; http://clinicaltrials.gov/; NCT01906866. Key words: melatonin, sleep, autism, neurogenetic, pediatric ACCEPTED

1 1

ACCEPTED MANUSCRIPT INTRODUCTION Pediatric insomnia is a widespread problem with prevalence of 1-6% in the general pediatric population and 50–75% in children with neurodevelopmental or psychiatric comorbidities, specifically autism spectrum disorder (ASD; including autistic disorder, Asperger’s disorder and pervasive developmental disorder) and neurogenetic disorders (NGD, e.g. Rett's disorder, Tuberous Sclerosis, Smith-Magenis syndrome [SMS] and Angelman syndrome).1-3 The sleep disturbances exacerbate cognitive performance deficits and behavioral problems and subsequently entire family distress.4 Current practices recommend parent-directed behavioral sleep interventions as first-line treatment for pediatric insomnia in ASD/NGD with reportedly 25% response rate.5,6 Pharmacotherapy is often provided when the behavioral intervention fails. However, despite the severity of sleep problems, there are no medications with regulatory approval for the treatment of insomnia in children and adolescents.7,8 Consequently, physicians often prescribe off-label (e.g., , alpha- agonists [], , ) for their side effects without proven safety, efficacy, or dosing regimen in children.7 Moreover, unlicensed melatonin preparations or food supplements are used despite considerable concerns over the quality and potential safety hazards. 8,9 Melatonin (N-acetyl-5-methoxytryptamine) is the primary hormone produced by the pineal gland during nocturnal periods to properly time circadian sleep/wake rhythms and enhance sleepiness.10 A growing body of evidence MANUSCRIPT indicates abnormal melatonin secretion and circadian rhythmicity in children with neurodevelopmental disorders, specifically ASD, which may explain the abnormal development of sleep/wake cycles, noted since the first year of life; such abnormalities justify the use of melatonin for insomnia in these populations.11-13 The use of melatonin for treating chronic sleep–wake cycle disorders of children with ASD/NGD is increasing.5,12,14 A recent meta-synthesis concluded that melatonin, behavioral interventions, and parent education/interventions appear the most effective at ameliorating multiple domains of sleep problems. 14 A recent study compared the effect of behavioral intervention and controlled-release melatonin 3 mg on sleep in children with ASD.15 Whereas this study demonstrated that each treatment alone was beneficial in improving SL and TST, their combination tended to be more effectiveACCEPTED although not significantly. In another study, children who failed to improve on behavioral intervention alone, immediate release (IR)-melatonin demonstrated beneficial effects on sleep latency (SL) and to a lesser extent on total sleep time (TST).6 Similarly, a recent open-label, dose escalation study of IR-melatonin reported significant improvement in SL but not TST.16 As melatonin has a very short half-life (40 minutes), a prolonged-release (PR) melatonin (PRM) formulation, designed to mimic the endogenous profile by releasing

2 2

ACCEPTED MANUSCRIPT melatonin throughout the night, may help improve both sleep initiation and maintenance.13,17 Despite the increasing use of melatonin in this pediatric population, short- and long- term safety and efficacy controlled studies of melatonin as well as dosing regimen data are lacking. Therefore, there is an unmet need for appropriately tested and approved drugs proven to be effective and safe for insomnia in this population. A 2mg PRM prescription is available in the European Union and other countries for primary insomnia in patients aged 55 years and older (Circadin, Neurim Pharmaceuticals Ltd).18 If crushed to facilitate swallowing, these tablets lose the controlled-release properties, as the active ingredient is immediately released.19 A pediatric-appropriate 3mm diameter prolonged release melatonin minitablet (PedPRM) was developed (Neurim Pharmaceuticals Ltd), which can be more easily swallowed by young children.20 The aim of this study was to assess the efficacy and safety of PedPRM vs placebo for insomnia in children with ASD and NGDs who did not improve after 4 weeks of sleep behavioral intervention. We hypothesized that PedPRM will effectively increase TST, an important concern for families of the children. 2,21 The dosage of melatonin used in the study (dose titration, 2-5 mg/day and up to 10 mg/day) was well within the range reported in children with ASD and NGD. 6,12,15-17 METHOD Study Design A randomized, double-blind, placebo-controlled MANUSCRIPT trial was conducted at 14 centers in the United States (US) (196 participants, 73.4%) and 10 centers in Europe: United Kingdom, France, Netherlands and Finland (71 participants, 26.6%). The double-blind phase was conducted from December 2013 to May 2016. The protocol and informed consent form were reviewed and approved by the institutional review boards of participating institutions. The trial complied with the principles of the Declaration of Helsinki (1989) and standards of good clinical practices; participants and their caregivers gave written, signed, informed consent prior to participation. Participants Participants were children (2 -17.5 years) with: 1) physician-diagnosed ASD according to International Classification of Diseases-10 th Revision or DSM-5/-IV criteria, or NGDs and 2)ACCEPTED sleep problems (minimum 3 months of impaired sleep defined as ≤6 hours of continuous sleep and/or ≥0.5-hour SL from lights-off on 3 of 5 nights based on parent reports and patient medical history. Exclusion criteria included other sleep disorders (e.g., moderate to severe sleep apnea), use of prohibited medication (see Table S1, available online) or melatonin within 2 weeks prior to screening, allergy to melatonin or lactose, or unresponsive to previous Circadin therapy participation in a clinical trial within the last 3 months prior to the study. Females not

3 3

ACCEPTED MANUSCRIPT using contraceptives who were sexually active, pregnant, and/or breastfeeding were excluded. Procedures The study comprised 2 weeks single-blind placebo run-in, a randomized double-blind efficacy and safety study of 13 weeks PedPRM/placebo treatment (completed), followed by open-label treatment (ongoing) comprising 13 weeks PedPRM at final dose (or placebo equivalent dose), 78 weeks PedPRM treatment with optional dose escalation to 10 mg, and 2 weeks single-blind placebo period (withdrawal effects), altogether 2.2 years. Here we report on the double-blind period. Children without documented history of sleep behavioral intervention at screening underwent 4 weeks of advice booklet-assisted basic parent-led sleep behavioral intervention based on a previously studied and standardized sleep behavior treatment).22 This period ensured that children whose sleep disorder was amenable to treatment with non- pharmacological intervention were not randomized, and also served as wash-out from any . Eligible children entered a 2-week single-blind placebo run-in period after which, if they still had impaired sleep defined as ≤6 hours of continuous sleep and/or ≥0.5 hour SL from lights-off in 3 out of 5 nights in the last two weeks, based on parent-reported Sleep and Nap Diary (SND) 23 , they were randomly assigned (1:1), via permuted block randomization (block size of four) using an electronic clinical record form (eCRF) system, to receive either PedPRM or placebo for 13-week double-blind treatment period. Placebo mini-tablets were identical in appearance and formulation to PedPRM MANUSCRIPT minitablets without melatonin. The starting dose of PedPRM (or placebo) was 2 mg once-daily. After 3 weeks of double-blind treatment, sleep variables were assessed. If the patient did not improve from baseline by at least 1 hour as measured by shortening of SL and/or increase in TST, dose was escalated to 5 mg. Children then continued double-blind on 2 or 5 mg of PedPRM or placebo for the remaining 10 weeks, with an efficacy assessment by the end of the 13-week double-blind treatment period. Study Endpoints Sleep variables, reported by parent/caregiver, were assessed using a validated SND that has been used in previous trials, including a previous pediatric IR-melatonin trial.6,24 The SND was to be completed every morning by the parent/caregiver at home for 14 days prior to each visit. TheACCEPTED a priori-defined primary efficacy endpoint was the changes from baseline in mean TST over the 14 days. The first secondary endpoint was the change from baseline in mean SL. Other secondary sleep variables were the changes from baseline in mean duration of wake after sleep onset, mean number of awakenings and mean longest sleep episode (LSE), all from the SND, change from baseline in Composite Sleep Disturbance Index (CSDI) score and sub-scores, and number of dropouts during the 13-week double-blind treatment period. The CSDI is a validated tool scoring the frequency and duration of sleep

4 4

ACCEPTED MANUSCRIPT problems reported by parents.6 Safety was monitored throughout the study using standard clinical trials methods and definitions (Treatment-Emergent Signs and Symptoms [TESS], adverse events [AEs], vital signs, and physical examination). Standard age-appropriate methods (Tanner scale [pubertal stage], body mass index [BMI] percentiles, Z scores [growth], epilepsy [assessment]) were used to assess the children’s development and health status. Actigraphy monitors (Actiwatch) were included in the study as a secondary objective measurement tool for the primary (TST) and first secondary (SL) sleep parameters. Actigraphs were dispensed to participants and were worn from drug intake at night until lights on in the morning for 14 nights before Visit 2 and Visit 4. Statistical Methods Efficacy analyses are presented for the full analysis subset (FAS), comprising all participants in the safety analysis set (who take at least one dose of study medication) who satisfy all major entry criteria and who have valid assessments of mean TST at baseline and at least one during the double-blind phase and for the per protocol population (PPP), comprising all FAS participants without major protocol violations. The primary and secondary variables were analyzed using a mixed-effects model for repeated measures (MMRM) that included fixed effects for visit, mean baseline value, randomized treatment, and mean baseline value and randomized treatment both nested within visit. Visit-to-visit repeated measures assumed unstructured covariance structure. Multiple imputat MANUSCRIPTion (MI; with withdrawals mimicking placebo participants) and analysis of covariance (ANCOVA) using baseline observation carried forward (BOCF) analyses were also undertaken to assess effect of missing data and unbalanced withdrawal. Sample size was calculated assuming change (mean±SD) in TST of 0.72± 0.69 hours in the PedPRM and 0.27 ±0.45 in the placebo arm, 2-sided significance level of .05, power of 0.95, 45 participants per treatment group. Assuming 25% attrition rate, approximately 120 participants were needed. RESULTS Participants 267 participants were enrolled, of whom 125 (46.8%) were randomized in the double- blind phase, ACCEPTED and 119 (44.6%) treated (58 participants in the PedPRM group and 61 participants in the placebo group) (Figure 1). A total of 95 participants completed the double- blind phase; the completion rate was higher in the PedPRM group (85.0%) than in the placebo group (67.7%). Withdrawal of parent consent and loss to follow-up were higher among participants in the placebo than the PedPRM group. Dropout rates were significantly higher in the placebo than in the PedPRM group (21 participants, 32.3% compared to 9 participants, 15.0% Chi-Square p=.040)

5 5

ACCEPTED MANUSCRIPT Randomized participants treatment groups had similar demographic characteristics (Table 1). Of randomized participants, 3.8% had <6h continuous sleep, 40.2% had SL<30 min, and 56% had both. Most (83.2%) participants had a documented history of sleep behavioural intervention, 28.8% of participants had ADHD, and 12.8% participants had epilepsy as determined by patient medical history. Prior ADHD treatment was reported by 10.4% participants, by 8.0% participants, anti-epileptics by 6.4% participants, clonidine by 5.6%, and beta-blockers by 0.8% participants. Prior melatonin use was reported by 65.6% of the participants (IR 57.6%, CR 4.0%, other 4.0%), but 52.0% had stopped because the melatonin had no effect (22.4%) or for other reasons (29.6%). The rest of the 17 participants (13.6%) were washed out during the behavioural intervention period. No participants reported prior Circadin use. There were no notable differences between the groups for baseline disease characteristics or medications. Treatment compliance, calculated by returned mini-tablets after 13 weeks of double-blind treatment, was 105.4 ± 21.98% in the PedPRM and 99.7 ± 9.25% in the placebo group. Principal investigators reported that there was no need to crush the mini-tablets and children were able to swallow the tablets, thus confirming recent study results.20 Efficacy The study met the primary endpoint demonstrating statistically significant effects of PedPRM vs. placebo on change from baseline in mean SND-assessed TST after 13 weeks of double-blind treatment (Table 2). At baseline, mean MANUSCRIPT TST was 457.2 minutes in the PedPRM and 459.9 in the placebo group. By the end of the 13 weeks, children slept on average 57.50 minutes longer in the PedPRM-treated (n=52), compared to 9.14 minutes in the placebo- treated group (n=48). The adjusted mean (SE) change from baseline TST in PedPRM-treated participants was 51.16 (10.46) min vs 18.73 (10.82) min in the placebo-treated group (p=.034); a similar effect was demonstrated in the PP population (n=88, 44 in each group, p=.009). Results of the MI (p=.026) and BOCF (p= .039) analyses (see Table S2, available online) confirmed that MMRM results are robust to different assumptions regarding patient attrition. Although many patients with SMS have comorbid ASD,25 when analysing the ASD subpopulation alone, the adjusted mean (SE) change from baseline TST in the PedPRM- treated was 52.6 (10.7) min (n=50) vs 17.63 (11.1) min in the placebo-treated group (p=.026). Cohen’s d effectACCEPTED sizes26 for the FAS, ASD, and PPP were 0.43, 0.46, and 0.58, respectively (Table 2). MMRM analysis including ADHD comorbidity as a factor demonstrated the same improvement in participants with (adjusted mean [95% CI] difference of 32.92 minutes [- 22.61, 88.46]) and without (32.96 minutes [-2.98, 68.89]) ADHD comorbidity. Using improvement in duration of mean TST by 45 minutes or more as a criterion of clinical response,27-29 the percentage of TST responders after 13 weeks of double-blind

6 6

ACCEPTED MANUSCRIPT treatment in the PedPRM group was 37.9%, significantly higher than in the placebo-treated group 16.4% (Odds ratio 5.75; p=.003; see Table S3, available online), thus providing evidence for the clinical meaningfulness of the treatment effect. The 21.5% difference in percentage between groups equals NNT=4.7 for any additional responder above placebo. Change from baseline in mean TST during the double-blind phase is shown in Figure 2A. The improvement in TST and the significant differences between PedPRM and placebo treatment effects were already evident at week 3 (p=.006). In participants who improved on the 2-mg dose after 3 weeks, efficacy was maintained throughout the rest of the 13-week period (Table 3). The increase in TST between weeks 3 and 13 of the double-blind treatment is ascribed to the dose escalation in participants who did not respond to the initial dose. SND-assessed SL after 13 weeks of double-blind treatment (first secondary outcome) also improved (i.e., decreased) significantly more with PedPRM 2/5 mg than placebo treatment (Table 2). At baseline mean SL was 95.2 minutes in the PedPRM and 98.8 in the placebo group. By the end of the 13-week treatment period, children fell asleep on average 39.6 minutes faster in the PedPRM-treated (n=52), compared to 12.51 minutes in the placebo- treated group (n=48). Mean (SE) adjusted change from baseline in SL in the PedPRM-treated was -37.88 (6.82) min vs -12.58 (7.01) min in the placebo-treated group (MMRM: p= .011, MI: p=.012 , BOCF: p=.024 ); same for the PP population (n=88, 44 in each group, p= .018). Effect size of the treatment (over placebo; FAS, PPP) was 0.52 for both (Table 2). Using reduction in SL by ≥15 minutes as MANUSCRIPT a criterion of clinical response 27,29 after 13 weeks, 63.8% of the PedPRM-treated, compared to 32.8% in the placebo-treated group, were responders (Odds ratio 3.83; p=.001, see Table S3, available online). The difference of 31% between groups corresponds to an NNT of 3.2. The improvement (decrease) in SL and the significant differences between PedPRM and placebo effects were already evident at week 3 (p=.001, Figure 2B). In participants who improved on the 2mg dose after 3 weeks, efficacy was maintained throughout the rest of the 13-week period (Table 3). The improvement in SL between weeks 3 and 13 of the double- blind treatment is ascribed to the dose escalation in participants who did not respond to the initial dose. Altogether, by the end of the double-blind period, the percentage of participants with TST and/or SLACCEPTED response rate was significantly higher with PedPRM than placebo (68.9% vs 39.3%, p=.001). Besides shortening of SL, the observed increase in TST could be best explained by a greater improvement (increase) in the LSE compared to placebo (Table 2). By the end of the 13-week double-blind period, mean LSE increased on average by 77.93 minutes in the PedPRM-treated, compared to 25.45 in the placebo-treated group (Table 2). The adjusted mean (SE) change from baseline in LSE in the PedPRM-treated (FAS) was

7 7

ACCEPTED MANUSCRIPT 72.18 (14.76) minutes vs 30.02 (15.49) in the placebo-treated group; and estimated treatment differences of 42.16 minutes in the FAS (MMRM: p=.052 MI: p=.039 ; BOCF: p=.044 ) and 55.41 minutes in the PPP (n=88, 44 in each group, p=.019). The effect sizes of the treatment over placebo were 0.41 (FAS) and 0.55 (PPP) (Table 2). The increase in LSE with PedPRM was present after 3 weeks of treatment and significantly greater than placebo (p=.034). There were no other statistically significant differences in SND-assessed treatment effects. By the end of the double-blind period, mean bedtime in the PedPRM-treated group was somewhat earlier than in the placebo-treated group but not significantly. Importantly, PedPRM treatment did not induce earlier wakeup; children woke up somewhat later in the PedPRM-treated compared to those in the placebo-treated group but not significantly (Table 2). Subgroup analyses showed similar responses by geographical region (US and Europe) and age groups (2-5 and over 8 years of age). A somewhat higher response in smaller, 6-8 years age group is hard to interpret due to participant small sample size. Dose Escalation After 3 weeks of double-blind treatment, 41% (23 of 56) of children in the PedPRM- treated group improved on the 2-mg dose by 60 minutes or more in mean TST and/or SL, vs 20% (12) in the placebo-treated group. The rest, 59% (33 of 56) in the PedPRM-treated and 80% (49 of 61) in the placebo-treated group, were escalated (after 3 weeks) to the higher dose (5mg). Table 3 depicts the efficacy of PedPRM MANUSCRIPT by dose. In general, participants in the PedPRM group who responded to the 2-mg dose maintained that response for the entire 13 weeks double-blind period. The subgroup of participants who did not respond enough to the 2-mg dose level showed marked improvements in sleep variables after the dose increase. Of participants who completed the 13 weeks treatment on 5mg PedPRM dose, 50% (15 of 30) improved by 60 minutes or more in mean TST and/or SL. The improvement in TST and SL in these participants was similar in the 2mg and 5mg PedPRM doses. The rest would have justified a further dose escalation. Actigraphy Despite major efforts to ensure adherence, actigraphy monitoring was challenging in this population, and a majority of participants (75% in the PedPRM and 77% in the placebo group) refusedACCEPTED to wear the device during one or both periods and/or took it off some time during the night. Only 12 participants in the PedPRM and 13 in the placebo group had data for both baseline and 13 weeks of treatment, and even in those it was not possible to ascertain that they wore the device throughout the night. Based on the limited dataset available, by the end of the 13-week double-blind treatment, TST increased with PedPRM by mean (SD) 7.23 (96.5) min and decreased with placebo by -23.94 (61.86) min. SL decreased with PedPRM by mean (SD) -25.11 (38.37) min and decreased with placebo by -12.50 (24.74) min.

8 8

ACCEPTED MANUSCRIPT CSDI (Composite Sleep Disturbance Index) The PedPRM-treated group showed a greater improvement (decrease) in total CSDI score compared to placebo-treated group (see Table S4, available online). The estimated mean (SE) treatment difference for the change from baseline in total CSDI score was -0.95 (0.436) (p=.032) at week 3 and -0.92 (0.511) units ( p= .074) at week 13. In the PedPRM group, there was a significant improvement in CSDI-assessed parents' satisfaction in child sleep pattern ( p= .005) and tendencies to benefit in “problem settling” (p= .054) and “co-sleeping” (p=.067) compared to placebo (see Table S4, available online). There were no statistically significant differences between the treatment groups for any of the remaining CSDI subscores. Safety Evaluation During the double-blind phase, treatment-emergent AEs (TEAEs) were reported by 51 (85.0%) participants in the PedPRM-treated (202 events) and 50 (76.9%) participants in the placebo-treated group (159 events) (see Table S5, available online). Most of these TEAEs were similar between groups and known symptoms in children with ASD (e.g., agitation, mood swings) or generally in children (e.g., upper respiratory tract infection, cough, dyspnea, vomiting). Nervous system disorders were more common in the PedPRM-treated group (41.7% vs 21.5%); the difference was mainly driven by somnolence and headache, which were more common in the PedPRM-treated group. TEAEs judged by the clinician as treatment-related occurred in 12 (20.0%) participan MANUSCRIPTts in the PedPRM and 11 (16.9%) in the placebo group (28 and 20 events respectively) (see Table S6, available online). Somnolence, an expected AE of PRM, was more commonly reported in the PedPRM than placebo group.18 During the double-blind phase, severe events were reported by 13 (21.7%) participants in the PedPRM and 13 (20.0%) participants in the placebo group, with similar patterns (see Table S7, available online). One patient in the PedPRM group discontinued due to non-serious AEs (fatigue, agitation, and stereotypy). One patient in the placebo group temporarily discontinued due to 2 SAEs (pneumonia and respiratory tract infection viral) and one non-serious AE (tachypnea). There were no notable differences between PedPRM and placebo for mean changes in blood pressure, pulse rate, respiratory rate, or temperature. Overall, there were no clinically significant differencesACCEPTED between PedPRM and placebo for weight, height, or BMI. At baseline, the Tanner assessments of pubertal development showed a greater proportion of participants in the placebo group were preadolescent compared to the PedPRM group, reflecting the slightly lower age of participants in the placebo group (mean age 8.4 versus 9.0 years). In the 13-week double-blind phase, there was no apparent difference between PedPRM and placebo in change from baseline for SD scores of pubic hair, breast, and genitalia development. Of 16 participants with a history of epilepsy in the study, only one had experienced a

9 9

ACCEPTED MANUSCRIPT seizure (non-serious absence seizure) and this was during placebo treatment. DISCUSSION This randomized placebo-controlled study showed that in children with ASD, a pediatric-appropriate PRM was more effective than placebo in increasing TST, the primary endpoint of this study, and reducing SL. Both these beneficial actions of PedPRM on sleep were established by 3 weeks and maintained throughout the 13 weeks of treatment. Importantly, the drug was similarly effective in children with or without ADHD comorbidity. No unexpected safety issues were reported. Adverse effects were few and mild, with only headaches and daytime somnolence increased in the treatment group relative to placebo group, and there was no increase in or new onset of seizures. In contrast to the usual difficulties with tablet formulations experienced by children with ASD, compliance was excellent without the need to crush or dissolve PedPRM (which would negate the prolonged- release properties). The effect sizes of PedPRM on TST and SL (0.43, 0.52) and NNTs (4.6, 3.2) indicate efficacy that would be noticed in clinical practice. Altogether, 68.9% of children had benefited from PedPRM 2/5 mg, achieving longer sleep duration by 45 minutes or more and/or falling asleep quicker by 15 minutes or more compared to 39.3% with placebo. Thus 2 of 3 children with ASD are expected to have a clinically meaningful response to a 2/5 mg dose. The improvements culminated in increased parental satisfaction in child's sleep, bed resistance, and co-sleeping patterns with PedPRM. MANUSCRIPT Interestingly, 41% of participants who responded initially to 2 mg PedPRM remained responsive throughout the study period without requiring dose escalation. Of the participants who required a dose increase to 5 mg PedPRM, half responded and the rest were eligible for further dose escalation. (Further dose escalation to 10 mg is permitted in the open-label phase of this study.) This study corroborates the potential benefits of exogenous melatonin on sleep of children with ASD and other neurodevelopmental disorders. Mean increases of 32 minutes in TST and reduction of 25 minutes in SL over placebo are essentially similar to values from two other similarly sized controlled trials in comparable populations using IR and CR melatonin formulations.6,15 In theACCEPTED study of IR-melatonin, a proportion of children on melatonin fell asleep quicker but started to wake earlier, and as a result, the mean improvement in TST in that study was shorter than the improvement in SL.6 In our study, children on melatonin fell asleep quicker but did not wake up earlier, and as a result, the mean improvement in TST was longer than the improvement in SL, compatible with the prolonged-release nature of this formulation. The earlier onset and wakeup of children treated with IR-melatonin suggested an unwarranted advance of the sleep-wake cycle phase. This doesn’t occur with PedPRM, which

10 10

ACCEPTED MANUSCRIPT provides a prolonged melatonin action from the earlier through the later hours of the night. Another difficulty with some melatonin formulations has been the gradual loss of effect of exogenous melatonin ascribed in part to poor CYP1A2 , resulting in sustained levels of melatonin in these participants.30 We did not see any attrition of effect during the initial 3 months of this study, and this will be an important factor to study during the open-label extension phase. This study included well-accepted and widely used subjective and objective sleep measurement tools. Dose escalation was systematic, and when the open-label results are later reported, this will be the longest robust study of melatonin in this population to date. This study chose to explore an important explanatory, often neglected, sleep outcome LSE and showed that participant increase in LSE by 77.9 minutes in the PedPRM on average compared to 25.4 minutes in the placebo group was a key contributor to the increase in TST. Despite our efforts to maximize compliance with actigraphy, one of our secondary outcomes, we encountered implementation difficulties, as reported in other studies with this population. In fact, data from only one quarter of the children studied were informative. Future trials should consider non-contact forms of objective sleep monitoring in these populations. Although the study had to span two distinct geographical regions (US and Europe), and participants with and without comorbid ADHD, the investigational drug PedPRM was found equally effective regardless of location or AMANUSCRIPTDHD. The age group covered in this study was wide, but we found similar benefits in younger and older patient subgroups (2–5-year-olds and children over 8 years of age). These data indicate that PedPRM (once-daily 2 mg or 5 mg dose) was effective and safe for 13 weeks of treatment in children with ASD and other neurodevelopmental disorders who suffer from sleep disorders refractory to parental sleep behavioral interventions. There was no sign of tolerance development throughout the study period. Drug compliance was very good in this study, supporting the use of the mini-tablets in this population. Efficacy was demonstrated in terms of increased total sleep time, reduced SL, and improved longest continuous sleep period. Adverse events on PedPRM treatment included somnolence, headache, and fatigue. No unexpected safety issues were reported. ACCEPTED

11 11

ACCEPTED MANUSCRIPT REFERENCES 1. Kotagal S, Broomall E. Sleep in children with autism spectrum disorder. Pediatr Neurol . 2012;47:242-251. 2. Elrod MG, Hood BS. Sleep differences among children with autism spectrum disorders and typically developing peers: a meta-analysis. J Dev Behav Pediatr . 2015;36:166- 177. 3. APA. Diagnostic and statistical manual of mental disorders DSM-5. 5th ed. Washington, DC: American Psychiatric Publishing; 2013. 4. Doo S, Wing YK. Sleep problems of children with pervasive developmental disorders: correlation with parental stress. Dev Med Child Neurol. 2006;48:650-655. 5. Malow BA, Byars K, Johnson K, et al. A practice pathway for the identification, evaluation, and management of insomnia in children and adolescents with autism spectrum disorders. Pediatrics . 2012;130 Suppl 2:S106-124. 6. Gringras P, Gamble C, Jones AP, et al. Melatonin for sleep problems in children with neurodevelopmental disorders: randomised double masked placebo controlled trial. BMJ . 2012;345:e6664. 7. Mindell JA, Emslie G, Blumer J, et al. Pharmacologic management of insomnia in children and adolescents: consensus statement. Pediatrics . 2006;117: e1223-1232. 8. Hollway JA, Aman MG. Pharmacological treatment of sleep disturbance in developmental disabilities: a review of the literature. Res Dev Disabil. 2011;32:939-962. 9. Erland LA, Saxena PK. Melatonin Natural Health Products and Supplements: Presence of and Significant Variability of Melatonin Content. J Clin Sleep Med . 2017;13:275-281. 10. Arendt J. Melatonin. Clin Endocrinol (Oxf) . 1988;29:205-229. 11. Tordjman S, Anderson GM, Pichard N, Charbuy H, Touitou Y. Nocturnal excretion of 6-sulphatoxymelatonin in children and adolescents with autistic disorder. Biol Psychiatry . 2005;57:134-138. 12. Rossignol DA, Frye RE. Melatonin in autism spectrum disorders: a systematic review and meta-analysis. Dev Med Child Neurol. 2011;53:783-792.MANUSCRIPT 13. De Leersnyder H. Inverted rhythm of melatonin secretion in Smith-Magenis syndrome: from symptoms to treatment. Trends Endocrinol Metab . 2006;17:291-298. 14. Cuomo BM, Vaz S, Lee EAL, Thompson C, Rogerson JM, Falkmer T. Effectiveness of Sleep-Based Interventions for Children with Autism Spectrum Disorder: A Meta- Synthesis. Pharmacotherapy. 2017;37:555-578. 15. Cortesi F, Giannotti F, Sebastiani T, Panunzi S, Valente D. Controlled-release melatonin, singly and combined with cognitive behavioural therapy, for persistent insomnia in children with autism spectrum disorders: a randomized placebo-controlled trial. J Sleep Res . 2012;21:700-709. 16. Malow B, Adkins KW, McGrew SG, et al. Melatonin for sleep in children with autism: a controlled trial examining dose, , and outcomes. J Autism Dev Disord . 2012;42:1729-1737; author reply 38. 17. De Leersnyder H, Zisapel N, Laudon M. Prolonged-release melatonin for children with neurodevelopmental disorders. Pediatr Neurol . 2011;45:23-26. 18. EMA. Circadin (melatonin). 2010. http://www.ema.europa.eu/ema/index.jsp?curl=pages/medicines/human/medicines/000695/hu man_med_000701.jsp&mid=WC0b01ac058001d124ACCEPTED Accessed August 17, 2017. 19. Chua HM, Hauet Richer N, Swedrowska M, Ingham S, Tomlin S, Forbes B. Dissolution of Intact, Divided and Crushed Circadin Tablets: Prolonged vs. Immediate Release of Melatonin. Pharmaceutics . 2016;8(1). doi: 10.3390/pharmaceutics8010002. 20. Thomson SA, Tuleu C, Wong IC, Keady S, Pitt KG, Sutcliffe AG. Minitablets: new modality to deliver medicines to preschool-aged children. Pediatrics . 2009;123:e235-238. 21. Appleton RE, Jones AP, Gamble C, et al. The use of MElatonin in children with neurodevelopmental disorders and impaired sleep: a randomised, double-blind, placebo- controlled, parallel study (MENDS). Health Technol Assess . 2012;16:i-239.

12 12

ACCEPTED MANUSCRIPT 22. Montgomery P, Stores G, Wiggs L. The relative efficacy of two brief treatments for sleep problems in young learning disabled (mentally retarded) children: a randomised controlled trial. Arch Dis Child . 2004;89:125-130. 23. Carney CE, Buysse DJ, Ancoli-Israel S, et al. The consensus sleep diary: standardizing prospective sleep self-monitoring. Sleep . 2012;35:287-302. 24. Wade AG, Crawford G, Ford I, et al. Prolonged release melatonin in the treatment of primary insomnia: evaluation of the age cut-off for short- and long-term response. Curr Med Res Opin . 2011;27:87-98. 25. Laje G, Morse R, Richter W, Ball J, Pao M, Smith AC. Autism spectrum features in Smith-Magenis syndrome. Am J Med Genet C Semin Med Genet . 2010; 154C:456-462. 26. Cohen J. A power primer. Psychol Bull . 1992;112:155-159. 27. FDA. Medical Review Application number 21-782. 2005. https://www.accessdata.fda.gov/drugsatfda_docs/nda/2005/021782s000_Rozerem_medr.pdf Accessed August 17, 2017. 28. FDA. Highlights of prescribing information Hetlioz R (). 2014. https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/205677s000lbl.pdf Accessed August 17, 2017. 29. Hajak G, Clarenbach P, Fischer W, Haase W, Ruther E. improves sleep quality and daytime well-being in insomniac patients: comparison with , and placebo. Int Clin Psychopharmacol . 1994;9:251-261. 30. Braam W, van Geijlswijk I, Keijzer H, Smits MG, Didden R, Curfs LM. Loss of response to melatonin treatment is associated with slow melatonin metabolism. J Intellect Disabil Res . 2010;54:547-555.

FIGURE 1. Consolidated standards of reporting trials (CONSORT) diagram of the randomization and follow-up of the study participants. Note: PedPRM = pediatric-appropriate prolonged-release melatonin minitablets. MANUSCRIPT FIGURE 2. A) Caregivers’ sleep and nap diary (cSND) reported change from baseline in mean total sleep time (minutes) during the double-blind period . B) Change from baseline in mean sleep latency (minutes) during the double-blind period. Note: PedPRM = pediatric- appropriate prolonged-release melatonin minitablets; TST = total sleep time.

ACCEPTED

13 13

ACCEPTED MANUSCRIPT Table 1. Demographic Characteristic at Screening (Full Analysis Set Population)

PedPRM Placebo Overall (n=60) (n=65) (N=125)

Age, years Mean ± SD 9.0 ± 4.08 8.4 ± 4.24 8.7 ± 4.15 Range 2, 17 2, 17 2, 17 Sex, n (%) Male 45 (75.0%) 47 (72.3%) 92 (73.6%) Female 15 (25.0%) 18 (27.7%) 33 (26.4%) Ethnicity, n (%) Not Hispanic or Latino 40 (66.7%) 49 (75.4%) 89 (71.2%) Hispanic or Latino 12 (20.0%) 7 (10.8%) 19 (15.2%) Other 8 (13.3%) 8 (12.3%) 16 (12.8%) Unknown 0 1 (1.5%) 1 (0.8%) Race, n (%) White 57 (95.0%) 55 (84.6%) 112 (89.6%) Black or African American 1 (1.7%)MANUSCRIPT 8 (12.3%) 9 (7.2%) Other 3 (5.0%) 3 (4.6%) 6 (4.8%) Asian 0 2 (3.1%) 2 (1.6%) ASD 58 (96.7%) 63 (96.9%) 121 (96.8%) NGD (SMS) 2 (3.3%) 2 (3.1%) 4 (3.2%) Height, cm Mean ± SD 133.4 ± 24.17 130.4 ± 27.20 131.8 ± 25.73 Range 89, 180 79, 197 79, 197 Weight, kg Mean ± SD 37.86 ± 21.495 35.22 ± 23.249 36.49 ± 22.374 Range ACCEPTED 11.7, 90.1 9.8, 129.9 9.8, 129.9 BMI, kg/m2 Mean ± SD 19.50 ± 4.899 18.79 ± 4.901 19.13 ± 4.893 Range 12.7, 32.8 12.3, 35.3 12.3, 35.3

14 14

ACCEPTED MANUSCRIPT Note: ASD = autism spectrum disorder; BMI = body mass index; NGD = neurogenetic disorders; PedPRM = pediatric-appropriate prolonged-release melatonin minitablets; SMS = Smith-Magenis syndrome.

Table 2. Sleep and Nap Diary Variables After 3 and 13 Weeks of Double-Blind Treatment (Full Analysis Set Population)

Variable Group n Adjusted Treatment 95% CI p Value ES a Treatment Difference (SE) Mean (SE)

3 weeks

TST (min) PedPRM 58 35.00 (9.18) 35.49 (12.73) 10.26, 60.72 .006 Placebo 61 -0.49 (8.82)

SL (min) PedPRM 58 -30.46 (6.62) -30.61 (9.19) -48.82,-12.39 .001 Placebo 61 0.15 (6.38) MANUSCRIPT Duration of PedPRM 58 -4.06 (2.49) 1.11 (3.48) -5.80, 8.01 .751 wake time Placebo 61 -5.16 (2.4290) (min)

Number of PedPRM 58 -0.14 (0.08) -0.08 (0.108) -0.29, 0.14 .476 awakenings Placebo 61 -0.06 (0.08)

LSE (min) PedPRM 58 35.89 (12.43) 37.38 (17.44) 2.79, 71.97 .034 Placebo 61 -1.49 (12.23)

Total time PedPRMACCEPTED 58 4.01 (7.87) 1.69 (10.93) -19.95,23.33 .877 in bed (hrs) Placebo 61 2.32 (7.57) Time to bed b PedPRM 58 0.05(0.072) 0.00 (0.100) -0.20, 0.20 .987 (hrs) Placebo 61 0.04(0.069) Time awoke c PedPRM 58 -0.15(0.121) 0.18 (0.168) -0.15,0.51 .288

(hrs) Placebo 61 -0.33(0.116)

15 15

ACCEPTED MANUSCRIPT 13 weeks TST (min) PedPRM 52 51.16 (10.46) 32.43 (15.11) 2.48, 62.38 .034 0.43 Placebo 48 18.73 (10.82)

SL (min) PedPRM 52 -37.88 (6.82) -25.3 (9.79) -44.7, -5.9 .011 0.52 Placebo 48 -12.58 (7.01)

Duration of PedPRM 44 -10.38 (2.41) -0.08 (3.49) -7.02, 6.86 .981 - wake time Placebo 41 -10.30 (2.49) (min)

Number of PedPRM 52 -0.30 (0.09) -0.09 (0.129) -0.35, 0.16 .474 0.14 awakenings Placebo 48 -0.2 (0.09)

LSE (min) PedPRM 58 72.18 (14.77) 42.16 (21.44) -0.42, 84.73 .052 0.41 Placebo 61 30.02 (15.5)

Total time in PedPRM 58 13.33 (8.98) 4.75 (12.89) -20.8, 30.29 .713 - bed (hrs) Placebo 61 8.58 (9.23) MANUSCRIPT Time to bed b PedPRM 58 0.06(0.093) 0.21 (0.134) -0.05, 0.48 .118 (hrs) Placebo 61 -0.15(0.096) Time awoke c PedPRM 58 -0.04(0.149) 0.09 (0.215) -0.34, 0.51 .687 (hrs) Placebo 61 -0.13(0.154)

Note: LSE = longest sleep episode; PedPRM = pediatric-appropriate prolonged-release melatonin minitablets; SE = standard error; SL = sleep latency; TST = total sleep time. a Cohen's D effect size. b Number of hours before midnight. c Number of hoursACCEPTED after midnight.

16 16

ACCEPTED MANUSCRIPT Table 3. Sleep and Nap Diary Variables (Minutes) After 3 and 13 Weeks of Double- Blind Treatment by Dose (Full Analysis Set Population)

Variable Group N Mean Change From N Mean Change From Baseline (SE) Baseline (SE) 3 weeks 13 weeks 2 mg PedPRM - No further dose escalation

TST PedPRM 23 82.65(23.1) 22 81.40 (23.2) Placebo 12 35.76(21.8) 10 35.17 (30.3) SL PedPRM 23 -49.72(11.4) 22 -46.67 (12.8) Placebo 12 -31.93(16.5) 10 -15.40 (19.7) 5 mg PedPRM population

TST PedPRM 30 39.98 (19.1) Placebo 38 2.29 (12.3) SL PedPRM 30 -34.37 (11.2) Placebo 38 -11.76 (7.5) 5 mg PedPRM subpopulation --No further dose escalation needed TST PedPRM MANUSCRIPT 15/30 117.34 (18.29) SL PedPRM 15/30 -66.15 (15.88) 5 mg PedPRM subpopulation - Dose escalation to be considered

TST PedPRM 15/30 -37.38( 17.9 ) SL PedPRM 15/30 -2.56( 10.9 ) Note: PedPRM = pediatric-appropriate prolonged-release melatonin minitablets; SL = sleep latency; TST = total sleep time.

ACCEPTED

17 17

ACCEPTED MANUSCRIPT

Enrolled: 267

Not randomized: 142 Protocol non-compliance: 51 Screening failure: 44 Withdrawal of parent consent: 21 Lost to follow-up: 8 Adverse event: 5 Investigator decision: 4 Other: 9

Randomized: 125

Double-blind phase

PedPRM: 60 Placebo: 65

Withdrawal of parent consent: 2 Withdrawal of parent consent: 4

Treated: 58 Treated: 61

Prematurely discontinued: 7 Prematurely discontinued: 17 Protocol non-compliance: 2 MANUSCRIPTWithdrawal of parent consent: 8 Lost to follow-up: 2 Lost to follow-up: 5 Adverse event: 1 Investigator decision: 1 Investigator decision: 1 Withdrawal of assent: 1 Withdrawal of assent: 1 Other: 2

Completed: 51 Completed: 44

ACCEPTED ACCEPTED MANUSCRIPT A

80 p=.034 70

60 PedPRM p=.006 50 Placebo 40 TST (min) 30

20 Mean Change MeanFrom Baseline 10

0

3 Weeks 13 Weeks Treatment Duration

B MANUSCRIPT Treatment Duration

3 Weeks 13 Weeks 0

-10

-20 PedPRM -30 Placebo

-40 ACCEPTED =.001 Sleep Latency (min) SleepLatency p

Mean Change MeanFrom Baseline -50 p=.011 ACCEPTED MANUSCRIPT

Supplementary Data

Efficacy and Safety of Pediatric Prolonged-Release Melatonin for Insomnia in Children with Autism Spectrum Disorder

Short Title: Pediatric PR Melatonin for Sleep in ASD

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

Table S1 Prohibited Medication List

Active compound Trade name Class Reason for Exclusion Valdoxan Melatonin agonist Major depressive disorder Melitor Thymanax Rozerem Melatonin agonist Anxiolytic,myorelaxant, and amnesic Zopiclone Imovane Z-Drugs Zopiclone Nocturno Nocturno S.L. Sonata Z-Drugs Sedative/hypnotic Starnoc Zodorm Z-Drugs Hypnotic Stilnox Lunesta Z-Drugs Sedative/hypnotic Bondormin Hypnotic Brotizolam * Clonex Benzodiazepine Hypnotic Rivotril Klonopin MANUSCRIPT Diazepam Assival Benzodiazepine Hypnotic Diaz Valium Stesolid Flunitrazepam Hypnodorm Benzodiazepine Hypnotic Lorivan Benzodiazepine Hypnotic Ativan Loramet Benzodiazepine Sedative/hypnotic Noctamid Numbon Benzodiazepine Sedative Oxazepam Vaben Benzodiazepine Sedative/hypnotic Alepam Medopam Alopam Sobril Oxascand Dormicum Benzodiazepine Hypnotic ACCEPTEDHypnovel Versed Trialzolam Trialzolam Benzodiazepine Hypnotic Restoril Benzodiazepine Hypnotic Normison ACCEPTED MANUSCRIPT

Loprazolam Dormonoct Benzodiazepine Sedative ProSomHavlane Sonin Benzodiazepinederivative Sedative Eurodin derivative Triprolidine Anti-histamine Sedative/hypnotic hydrochloride / Anti-histamine Sedative/hypnotic hydrochloride Chlorpheniramine Anti-histamine Sedative/hypnotic maleate Dexchlorpheniramine Anti-histamine Sedative/hypnotic maleate Phenergan Anti-histamine Sedative Promethegan Romergan Farganesse Lergigan Unisom Anti-histamine Sedative Niaprazine Nopron class Sedative/hypnotic

Clomethiazole Thiamine- vitamin B1 like Sedative/hypnotic Phenobarbital MANUSCRIPT Sedative Sedative/hypnotic Sedative Alimemazine Nedeltran, Panectyl, Antipruritic Sedative/hypnotic Repeltin, Therafene, Theraligene, Theralen, Theralene, Vallergan, Vanectyl, and Temaril Herbal Fluvoxamine Luvox Antidepressant Interact with endogenous melatonin and increases melatonin levels by inhibiting its metabolism ACCEPTED

ACCEPTED MANUSCRIPT

Table S2. Multiple Imputation (MI) and Baseline Observation Carried Forward (BOCF) Analyses of Sleep and Nap Diary Variables

MANUSCRIPT

ACCEPTED ACCEPTED MANUSCRIPT

Variable MMRM Multiple BOCF Analysis (FAS) Imputation TST (minutes) Treatment difference 32.43 (15.107) 33.11 (14.856) 32.57 (SE) (15.563)

95% CI 2.48, 62.38 3.99, 62.23 1.68, 63.46 p-value .034 .026 .039

Sleep latency Treatment difference -25.30 (9.788) -25.31 -23.09 (minutes) (SE) (10.022) (10.080) 95% CI -44.71, -5.90 -44.95, -5.67 -43.10, -3.09

p-value .011 .012 .024

Duration of Treatment difference -0.08 (3.489) -0.19 (3.156) -0.19 (3.499) wake time (SE) (minutes) 95% CI -7.02, 6.86 -6.38, 5.99 -6.77, 7.15 p-value .981 MANUSCRIPT .952 .957

Number of Treatment difference -0.09 (0.129) -0.08 (0.134) -0.14 (0.131) awakenings (SE) 95% CI -0.35, 0.16 -0.35, 0.18 -0.40, 0.12 p-value .474 .531 .286

Longest sleep Treatment difference 42.16 (21.440) 43.20 (20.878) 44.72 duration (SE) (21.856) (minutes) ACCEPTED95% CI -0.42, 84.73 2.28, 84.12 1.29, 88.15 p-value .052 .039 .044 Note: FAS = full analysis set; MMRM = mixed-effects model for repeated-measures; SE = standard error; TST = total sleep time. ACCEPTED MANUSCRIPT

Table S3 Responder Analyses (Logistic Regression)

Week Responder PedPRM Placebo Treatment Circadin Versus Placebo

(n=58) (n=61) Odds ratio 95% CI p-value

TST responders 5 Yes 19 (32.8%) 11 (18.0%) 2.56 0.99, 6.58 .052 No 39 (67.2%) 50 (82.0%) 15 Yes 22 (37.9%) 10 (16.4%) 5.75 1.80, 18.36 .003 No 36 (62.1%) 51 (83.6%) Sleep latency responders 5 Yes 29 (50.0%) 15 (24.6%) 4.26 1.73, 10.53 .002 No 29 (50.0%) 46 (75.4%) 15 Yes 37 (63.8%) 20 (32.8%) 3.83 1.77, 8.30 .001 No 21 (36.2%) 41 (67.2%)

Note: Analyzed using logistic regression with factors for treatment and mean baseline value. Total sleep time (TST) responder: a patient is defined as a TST responder if the change from baseline in mean TST is 45 minutes or greater (increase in TST) over the 14 days prior to each scheduled visit.MANUSCRIPT Sleep latency responder: a patient is defined as a sleep latency responder if the change from baseline in mean sleep latency is 15 minutes or greater (reduction in sleep latency) over the 14 days prior to each scheduled visit. PedPRM = pediatric-appropriate prolonged-release melatonin minitablets.

ACCEPTED ACCEPTED MANUSCRIPT

Table S4. Composite Sleep Disturbance Index (CSDI) Sleep Disturbance Change From Baseline at 3 and 13 Weeks Double Blind Period (Mixed-Effects Model for Repeated Measures Analysis Full Analysis Set [FAS])

FAS Adjusted treatment Treatment 95% CI p-value means (SE) difference (SE) (N, N) PedPRM Placebo CSDI total score (56, 60) Week 3 -1.74(0.312) -0.79(0.303) -0.95(0.436) (-1.81,-0.08) .032 (55, 48) Week 13 -2.44(0.352) -1.52(0.370) -0.92(0.511) (-1.93,0.09) .074 CSDI subscore 1: How often does your child have problems settling at bed time? (55, 48) Week 13 -0.56(0.097) -0.28(0.103) -0.28(0.142) (-0.56,0.01) .054 CSDI subscore 2: How long does it take them to settle to sleep? (55, 48) Week 13 -0.48(0.084) -0.29(0.089) -0.19(0.122) (-0.44,0.05) .115 CSDI subscore 3: How often does your child wake in the night? (55, 48 ) Week 13 -0.51(0.111) -0.42(0.117) -0.09(0.163) (-0.41,0.24) .593 CSDI subscore 4: How long does it usually take to resettle him/her? (55, 48) Week 13 -0.32(0.107) -0.20(0.112) -0.12(0.155) (-0.43,0.19) .445 CSDI subscore 5: How often does your child wake before 5am in the morning and remain awake? (55, 48) Week 13 -0.19(0.083) -0.23(0.089) 0.04(0.122) (-0.20,0.28) .74 CSDI subscore 6: How often does your child insist on sleeping with someone else for most/all of the night? (55, 48) Week 13 -0.39(0.103) -0.11(0.108) MANUSCRIPT-0.28(0.149) (-0.57,0.02) .067 CSDI subscore 7: How satisfied are you with your child's current sleep pattern? (55, 48) Week 13 1.43(0.175) 0.71(0.184) 0.72(0.254) (0.22,1.23) .005

Note: PedPRM = pediatric-appropriate prolonged-release melatonin minitablets.

ACCEPTED ACCEPTED MANUSCRIPT

SAFETY EVALUATION

Table S5 Most Commonly Reported Treatment-Emergent Adverse Events (TEAEs)

Double-Blind Phase

PedPRM Placebo

Patients Events Patients Events

(n=60) (n=65)

Number of patients with at 51 (85.0%) 50 (76.9%) least one TEAE Total number of AEs 202 159 Preferred term a

Somnolence 17 (28.3%) 18 7 (10.8%) 7

Fatigue 14 (23.3%) 18 12 (18.5%) 13 MANUSCRIPT Upper respiratory tract 9 (15.0%) 9 8 (12.3%) 9 infection

Mood swings 9 (15.0%) 9 11 (16.9%) 12

Vomiting 8 (13.3%) 10 9 (13.8%) 9

Agitation 8 (13.3%) 9 7 (10.8%) 8

Headache 8 (13.3%) 8 4 (6.2%) 4

Cough 7 (11.7%) 7 5 (7.7%) 5

Dyspnoea 6 (10.0%) 6 4 (6.2%) 4

Note: PedPRM =ACCEPTED pediatric-appropriate prolonged-release melatonin minitablets. aList includes TEAEs reported by ≥10% patients in any group.

ACCEPTED MANUSCRIPT

Table S6 Most Commonly Reported Treatment-Related Adverse Events (AEs)

Double-Blind Phase

PedPRM Placebo

Patients Events Patients Events

(n=60) (n=65)

Number of patients with at least 12 (20.0%) 11 (16.9%) one treatment-related AE

Total number of AEs 28 20

Most common AEs a

Somnolence 7 (11.7%) 7 2 (3.1%) 2

Mood swings 1 (1.7%) 1 4 (6.2%) 4

Nightmare 0 0 4 (6.2%) 4 Note: PedPRM = pediatric-appropriate prolonged-rele MANUSCRIPTase melatonin minitablets. aAEs reported as treatment-related by ≥5% patients in any group.

ACCEPTED ACCEPTED MANUSCRIPT

Table S7 Most Commonly Reported Severe Adverse Events (AEs)

Double-Blind Phase

PedPRM Placebo

Patients Patients

(n=60) (n=65)

Number of patients with at least 13 (21.7%) 13 (20.0%) one severe AE

Most Common AEs

Agitation 5 (8.3%) 3 (4.6%)

Fatigue 4 (6.7%) 2 (3.1%)

Mood swings 3 (5.0%) 5 (7.7%)

Note: This table includes severe AEs reported by ≥5% patients in any group. PedPRM = pediatric-appropriate prolonged-release melatonin minitablets. MANUSCRIPT

ACCEPTED