Efficacy of the triple-combination SM-1 in a 5-h phase advance transient insomnia model

Abstract

Thomas Dahl, PO Box 404, Guilford, CT 06437 USA. Email: tadahl@outlook.com. The objectives of the study were to demonstrate the efficacy and safety of SM-1 in a circadian challenge model of transient insomnia. Randomized, double-blind, placebo-controlled cross-over study utilizing a 5-h phase advance model of transient insomnia. Subjects were 85 healthy adults reporting a history of transient insomnia, with an average age of 38.9 years. Both SM-1 and placebo were administered to all subjects in a randomly assigned sequence, with at least 1 week between treatments. The primary endpoint was total sleep time determined by polysomnography. Secondary endpoints included wakefulness after sleep onset, latency to persistent sleep, number of awakenings, subjective total sleep time and subjective sleep onset latency, total sleep time by quarters of the night, subjective number of awakenings, and sleep quality. Safety endpoints included adverse events, Karolinska Sleepiness Scale, Digit Symbol Substitution Test, and predischarge evaluation (tandem gait and Romberg tests). SM-1 provided an increase of 94.4 min in total sleep time over placebo (p < 0.0001). Wakefulness after sleep onset, subjective total sleep time, subjective sleep onset latency, and total sleep time in the first quarter of the night also showed significant improvement. SM-1 was well-tolerated with both type and frequency of adverse events being comparable to placebo, and no residual sleepiness upon awakening (i.e., after 8 h). SM-1 provided a robust and statistically significant increase in total sleep time compared to placebo in a circadian model of transient insomnia, without evidence of next-day impairment.

Introduction

Prescription hypnotic trials have focused on patients meeting diagnostic criteria for insomnia [1]. Trials with over-the-counter (OTC) sleep agents on the other hand have been carried out in insomniacs or in healthy volunteers undergoing an acute sleep disturbance [2, 3].

The NIH distinction between transient and chronic insomnia does not appear to be two parts of a continuum. Rather, epidemiological data on the frequency of disturbed sleep exhibits a bimodal distribution, with one peak at 1–4 weeks and the other at 51–52 weeks [4]. There are few individuals with a frequency of sleep disturbances between these two peaks. In essence the first peak represents transient insomnia, while the second peak consists of individuals with chronic insomnia.

Interestingly, data on the frequency of hypnotic use parallel the frequency of insomnia data [5]. The first peak in the frequency of hypnotic use is for 1–14 days, with a large majority of users, 74%, reporting this. The single largest segment, 1–2 days, represents the majority (54%) of use. The second peak is nightly use for a year or more, which was reported by only 11% of hypnotic users. One may infer that the first peak represents treatment for transient insomnia, while the second peak represents medication usage in chronic insomnia.

The present study evaluates a medication intended for the short-term treatment of individuals with transient insomnia. The investigational product referred to as SM-1, is a controlled-release combination of the lowest effective doses of established sedative medications intended to both induce and maintain sleep, while avoiding impairment of next-day functioning. SM-1 has previously been studied in a model of daytime sedation [6], as well as in a Phase 2 study of sleep disturbance induced using a 5-h phase advance model [7]. In this study, as in the Phase 2 study, sleep disturbance was induced using a 5-h phase advance model followed by an 8-h polysomnographic (PSG) recording as the primary efficacy assay. Five-hour phase advance, relative to shorter advances, impacts the ability to stay asleep (i.e., WASO), more strongly than sleep onset time (i.e., LPS).

SM-1 is a staged combination of currently available drugs (50 mg Diphenhydramine, 5 mg Delayed-Release Zolpidem, and 0.5 mg Delayed-Release Lorazepam). It is designed to both induce and maintain sleep without impairment of next day functioning. The three drugs were selected based on their documented evidence to induce and or to maintain sleep. Specifically, diphenhydramine and zolpidem were selected for their ability to initiate sleep and to improve sleep in the first part of the night. Both were selected as diphenhydramine has shown inconsistent results on sleep. Early studies used only subjective data with some studies showing positive results [2, 8], while others showed negative results [9]. More recent studies using objective polysomnographic measures demonstrated only modest effects of diphenhydramine [10]. In contrast, zolpidem has produced more positive results [11] especially at the 10-mg dose. The 5-mg dose has shown less consistent and more modest results [11]. To optimize safety and efficacy, we combined the low dose of zolpidem with diphenhydramine. The advantage is that while zolpidem works on the sleep system via GABA, diphenhydramine works on histamine, a wake-promoting neurotransmitter. Lorazepam was put in the combination as it has been shown to maintain sleep and to decrease wake time during the night [12]. In recognition that three medications are being used, the lowest dose showing hypnotic activity for each of these medications was utilized. Thus, while 10 mg of zolpidem is associated with a variety of side effects including rebound insomnia and amnesia, 5 mg is not. The same can be said for the doses chosen for lorazepam and for diphenhydramine. Diphenhydramine has anticholinergic as well as antihistaminic properties at higher doses; therefore, the 50-mg dose was selected to minimize the anticholinergic side effects while providing the antihistaminic-mediated hypnotic effects. This combination and release/availability profile is a unique approach to addressing the problem, and, if successful, would provide a unique therapeutic option, allowing for a significant contribution to the treatment of transient insomnia.

Methods

This was a randomized, double-blind, single-dose cross-over study to assess the efficacy of a single dose of SM-1 (50-mg diphenhydramine, 5-mg delayed-release zolpidem and 0.5-mg delayed-release lorazepam) versus placebo in a 5-h phase advance model of transient insomnia as determined by PSG. The study was conducted at 2 study centers in the United States.

Eighty-five male and female subjects 18 years of age and older who routinely spent between 6.5 and 9 h in bed per night, and reported previous bouts of transient insomnia (i.e., occasional difficulty falling asleep or staying asleep), but did not meet diagnostic criteria for chronic insomnia, were recruited. Subjects were excluded if they had a clinically significant history of medical, sleep, or psychiatric disorders.

Prospective subjects completed a screening visit at the study center (Visit 1) conducted no fewer than 8, and no more than 14 days prior to the first day of the first treatment period (Day 1). Eligible subjects were given a sleep diary and asked to record their bedtimes for a minimum of 7 consecutive nights. These data were used to calculate each subject’s median habitual bedtime prior to randomization at the first treatment visit (Visit 2).

The order in which study treatments were administered was randomized equally by a computer-generated randomization table.

For each of the treatment visits, subjects checked in to the study center 7 h earlier than their median habitual bedtime as calculated from their sleep diary data, and went to bed (“lights out”) 5 h ± 30 min earlier than their median habitual bedtime. The lights out time established prior to the first treatment was maintained for the subsequent treatment nights as well. At each visit, subjects were administered their assigned treatment 30 min prior to lights out. Doses of study drug were administered orally by study personnel with 250 mL of water. An oral cavity check was performed to assure compliance with treatment. Blinding was maintained by the identical appearance of the treatment capsules; neither the Investigator, the scorer of the PSG, nor the subjects knew the identity of the treatment being administered.

Polysomnographic (PSG) recording began at lights out and continued for 8 h. Approximately 30 min after the termination of PSG recording, subjects completed a post-sleep questionnaire, the Karolinska Sleepiness Scale (KSS), and the Digit Symbol Substitution Test (DSST).

Within 1 week after discharge, but after at least 72 h following administration of the final dose, subjects received an End-of-Study (EOS) follow-up telephone call from study personnel.

The primary endpoint was PSG-derived total sleep time (TST) for SM-1 vs placebo. Secondary endpoints included Wakefulness After Sleep Onset (WASO), Latency to Persistent Sleep (LPS), Number of Awakenings (NAW), subject self-report of TST (sTST) and sleep onset latency (sSOL), TST by quarters of the night, subjective number of awakenings (sNAW), and sleep quality for SM-1 vs. Placebo.

The study’s safety endpoints were: treatment-emergent adverse events (TEAE); Karolinska Sleepiness Scale (KSS) score; Digit Symbol Substitution Test (DSST), number correct in 90 s; and Pre-discharge evaluation (tandem gait test and the Romberg test).

The primary efficacy endpoint, TST, and all secondary endpoints were tested using repeated measures Analysis of Variance (ANOVA) and were used to analyze the overall treatment differences between SM-1 and placebo. Factors in the model included treatment, sequence, period, and sex. Analyses were conducted using SAS^® software (SAS Institute, Inc, Cary, North Carolina) Version 9.3.

Written informed consent was obtained from each subject prior to any study-related procedure. Before study onset, the protocol, informed consent, advertisements to be used for the recruitment of study subjects, and any other written information regarding this study to be provided to the subject were approved by the Shulman IRB (now Advarra IRB, Cincinnati, Ohio). All IRB approvals were signed by the IRB chairman or designee and identified the IRB name and address, the clinical protocol by title or protocol number or both, and the date that the approval or a favorable opinion was granted.

Results

Disposition

One hundred and fifty-seven subjects were screened. Eighty-five (85) subjects were enrolled and randomized equally to each treatment sequence. Of these, 84 subjects received SM-1 and 83 subjects received placebo. Three subjects discontinued the study due to: protocol violation (1), physician decision (1), and lost to follow-up (1). All 85 subjects were followed for safety and were included in the efficacy analyses.

Demographics: Subjects were representative of an otherwise healthy adult population with a history of transient insomnia, ranging from 19 to 69 years of age. The mean (SD) age was 38.9 (12.57), mean (SD) weight was 75.3 (14.22) kg and mean (SD) BMI was 25.7 (3.44). Racial composition was n (%): White 24 (28.2), African American 50 (58.8), Asian 5 (5.9), and Other 6 (7.1).

Safety

Adverse events

No serious adverse events (SAE) occurred and there were no AE-related discontinuations. The frequency of subjects who experienced at least 1 TEAE in the SM-1 condition was similar to the frequency in the placebo condition (6.0% and 3.6%, respectively).

Five subjects (6.0%) experienced at least 1 TEAE while on SM-1, and 3 (3.6%) while on placebo. In total, 9 TEAEs were reported, including 6 with SM-1 treatment, and 3 with placebo. Four of the TEAEs were considered possibly or probably related to the study drug, 3 with SM-1 and 1 with placebo.

Nervous system disorders was the only common system organ class (SOC) with TEAE related to study drug that was reported in more than 1 subject (2 subjects, 2.4%; 1 each with SM-1 and placebo), and somnolence was the only TEAE related to study drug reported in the category. There was no preponderance of events in either treatment condition. Other drug-related TEAE reported with SM-1 but not placebo were 1 instance each of nightmare and asthenia. All the TEAEs were classified as mild in severity and resolved by the end of the study. All AEs resolved without intervention. No deaths, other serious adverse events, other significant AEs, or AEs leading to discontinuation occurred during the study. Treatment with SM-1 resulted in a minor but statistically significant decrease in DSST compared to placebo (− 4.2, P = 0.0051). In terms of patient reports of sedation, the difference in KSS was not statistically significant. Importantly subjects showed no impairment on Tandem gait and Romberg tests at the predischarge evaluations. (Table 1).

Table 1.

Safety

	SM-1 (N = 84) n (%)	Placebo (N = 83) n (%)
Number of subjects with at least 1 TEAE	5 (6.0%)	3 (3.6%)
Number of subjects with at least 1 drug-related TEAE	3 (3.6%)	1 (1.2%)
Drug-related TEAE
Sleepiness (mild)	1	1
Asthenia	1	0
Nightmare	1	0
Serious adverse events	0	0
	Mean (SD)	Mean (SD)	P
KSS	4.2 (1.99)	3.9 (2.01)	0.252
DSST	56.0 (14.21)	60.4 (16.81)	0.005
Impairment on predischarge evaluationa	0	0

^aTandem gait and Romberg tests

Karolinska Sleepiness Scale

KSS results reported after waking (Table 1) showed no statistically significant differences between treatments (p = 0.252 for SM-1 vs Placebo).

Digit symbol substitution test

There was a minor albeit statistically significant difference (p = 0.005) in DSST number correct Mean (SD) with SM-1 at 56.0 (14.21) compared to placebo at 60.4 (16.81) correct responses.

Pre-discharge evaluation

No abnormal findings were reported at any of the evaluations for impairment (tandem gait and Romberg tests) performed in the morning prior to discharge from each study period (Table 1).

Vital signs

No clinically significant abnormal results or trends in vital signs were observed during the study.

Clinical laboratory test results and physical examination findings

There were no clinically significant abnormal results or trends in clinical laboratory tests or physical examination findings.

Efficacy results

Efficacy results with corresponding p values are summarized in Table 2.

Table 2.

Efficacy

Measure	SM-1 (N = 84)	Placebo (N = 83)	p value SM-1 vs placebo
Mean TST (SD)	401.82 (68.074)	307.38 (107.365)	< 0.0001
Mean WASO	64.11	140.06	< 0.0001
LPS	18.59	39.25	0.1033
NAW	11.0 (6.06)	8.6 (5.06)	< 0.0001
sTST	437.4	367.2	< 0.0001
Sleep quality	3.0	2.3	0.0001
sSOL	22.7	34.8	0.077
sNAW	2.0 (1.73)	2.8 (1.83)	0.0013
TST by quarter	98.5	83.7	0.0002
	104.8	71.1	< 0.0001
	102.3	71.8	< 0.0001
	96.3	80.8	0.0016

The study’s primary endpoint was PSG-determined TST. Mean TST following treatment with SM-1 was 401.8 (68.07) min, an increase of 94.4 min over placebo [307.4 (107.37) min; p < 0.0001].

Secondary and other sleep endpoints

SM-1 reduced WASO by 62.2 min compared with placebo (64.1 min with SM-1; 140.6 min for placebo; p < 0.0001). SM-1 also significantly reduced total number of awakenings (NAW) compared to placebo. Mean (standard deviation) NAW for SM-1 and Placebo were 11.0 (6.06) and 8.6 (5.06), respectively (p < 0.0001).

Mean LPS was reduced by 21 min (18.59 min with SM-1 compared to 39.25 min with placebo), although the reduction did not achieve statistical significance (p = 0.1033). When analyzed by PSG quarters of the night, SM-1 provided statistically significantly greater TST than placebo during each quarter (p = 0.0002 for Q1, p < 0.0001 for Q2 and Q3, p = 0.0016 for Q4).

Subjective reports

In the morning subjects reported, their total sleep time (sTST) was significantly increased by 70.2 min for SM-1 compared to placebo (437.4 vs 367.2 min; p < 0.0001). SM-1 significantly reduced total number of awakenings by subject report (sNAW) compared to placebo. Mean (standard deviation) sNAW for SM-1 and Placebo were 2.0 (1.73) and 2.8 (1.83), respectively (p = 0.0013). Subjects also reported statistically significantly improved sleep quality following treatment with SM-1 compared to Placebo (p < 0.0001). Mean sSOL was reduced by 12.1 min following treatment with SM-1 compared to placebo (22.7 min vs 34.8 min), although the effect did not achieve statistical significance (p = 0.077).

Discussion

Zolpidem and lorazepam both work on the sleep side of the Sleep Wake system (i.e., GABA_A). Diphenhydramine, an H1 antagonist, works on the wake system. Zolpidem has been shown to reduce sleep latency; however, its effect on total sleep time is minimal. Simply increasing the dose of a hypnotic may increase total sleep time, however, it may lead to next-day impairment and abnormal behaviors in sleep, as has been observed with the 10-mg dose of zolpidem as compared to the 5-mg dose [11]. By combining low doses of these agents in a modified release formulation, the goal is to provide greater total sleep time than is possible with the individual agents, and without the next-day effects or other side effects that accompany increasing doses beyond those producing sedative activity.

SM-1 was previously shown in a small Phase 2 study to be an effective sleep agent in a phase advance model of transient insomnia [7]. In that double-blind randomized trial, SM-1 was found to provide both induction and maintenance of sleep, with the resultant increase in total sleep time of a little over 2 h (126.7 min increase) compared to placebo. The increase in sleep duration by SM-1 was mediated by a robust effect on sleep maintenance demonstrated both objectively by a statistically significant reduction in WASO, as well as subjectively, in subjects’ self-report of perceived total sleep time (sTST). Although the decrease in LPS following SM-1 administration did not achieve statistical significance, the reduction in subjective sleep latency (sSL) was highly significant (p = 0.006).

In this large confirmatory double-blind randomized trial of SM-1, the key findings of the earlier smaller exploratory study were replicated. The primary efficacy assay, PSG TST was robustly and significantly increased (p < 0.0001) with SM-1 treatment compared to placebo. Similarly, key secondary measures of efficacy such as WASO, NAW, sTST, sleep quality, sNAW, and TST during each quarter of the night were highly statistically significant. Also, as in the previous study [7], an improvement in LPS was observed but did not achieve statistical significance.

In terms of safety, SM-1 was safe and well tolerated in this study population of healthy adult male and female subjects reporting occasional difficulty falling asleep or staying asleep, as the categories and frequency of AEs were similar to placebo. Importantly, upon arising there were no clinically significant residual drug effects as indicated by KSS or pre-discharge objective evaluations of impairment, and only a minor, albeit statistically significant effect on DSST. This result is entirely consistent with the earliest observation of pharmacodynamic measures of sedative activity, showing that effects had dissipated by 8 h after SM-1 administration [6].

Limitations: To more fully assess possible residual effects, future studies with other measures of impairment, such as driving, and evaluations later in the day need to be conducted. The single-dose intervention will need to be complemented with additional studies to support short-term treatment. While some data support it, we do not know whether phase advance generalizes to all causes of transient insomnia. And lastly, this controlled sleep lab study will need to be complemented with real-world patient-focused experience.

Conclusion

SM-1 provided a robust and statistically significant increase in TST compared to placebo in a circadian model of transient insomnia, without next-day impairment. This confirmatory study reproduces the key findings of safety and efficacy observed in earlier exploratory studies.

Author contributions

TR, TD, GZ, and LBC contributed to the study conception and design. MA and RR contributed patient care and data collection. TR, TD and GZ contributed to data analysis. All authors contributed to interpretation of results. The first draft of the manuscript was written by TD and TR, and all authors contributed to manuscript review and editing. All authors read and approved the final manuscript.

Declarations

Conflict of interest

This study was entirely funded by Sequential Medicine, Ltd. Lan Bo Chen is the founder of Sequential Medicine. Thomas Dahl and Thomas Roth are consultants to the company. The remaining authors have no financial interests in or conflict with the subject matter or materials discussed.

Ethical approval

The trial was performed in compliance with Good Clinical Practice (ICH GCP), applicable regulatory requirements, and the current Declaration of Helsinki. It was registered at http://www.clinicaltrials.gov (Registration Number NCT 03331042). Subjects provided written informed consent.