Efficacy and safety of Dimdazenil in adults with insomnia disorder: results from a multicenter, randomized, double-blind, placebo-controlled phase III trials

Abstract

Study Objectives

To evaluate the efficacy and safety of Dimdazenil, a novel partial positive allosteric modulator for GABA_A receptor in adults with insomnia disorder.

Methods

This was a 2-week, multicenter, randomized, double-blind, placebo-controlled, parallel-group phase III study of Dimdazenil. The primary efficacy outcome was total sleep time (TST) analyzed by polysomnography (PSG) on day 13/14. Latency to persistent sleep (LPS), sleep efficiency (SE), and wake after sleep onset (WASO) were analyzed in the same way by polysomnography (PSG). The other secondary outcomes including the average subjective sleep latency (sSL), subjective TST (sTST), subjective SE (sSE), subjective WASO (sWASO), and subjective number of awakenings (sNAW) were analyzed from sleep diary data, and the insomnia severity index (ISI) was also assessed. Treatment-emergent adverse events (TEAEs) were monitored throughout the study.

Results

A total of 546 participants with insomnia (age ≥18 years) were randomized (2:1), received treatment with an oral dose of Dimdazenil (2.5 mg) or placebo, and analyzed. Compared to baseline and placebo, Dimdazenil demonstrated significant improvements in PSG measures, increased TST (71.09, 31.68 minutes, respectively; both p < 0.001), increased SE (13.26%, 5.55%, respectively; both < 0.001), reduced WASO (49.67, 20.16 minutes, respectively; both p < 0.001), and reduced LPS (21.65 minutes, p < 0.001; 6.46 minutes, p = 0.023). Compared to placebo, Dimdazenil also improved key self-reported measures of sTST (18.33 minutes, p < 0.001), sWASO (14.60 minutes, p < 0.001), sSL (4.23 minutes, p < 0.001), sSE (2.97%, p < 0.001), and sNAW (0.29, p < 0.001). Participants treated with Dimdazenil reported a significant improvement in ISI. Dimdazenil was well tolerated. The majority of TEAEs were mild or moderate. There were no clinically relevant treatment-related serious AEs and no deaths.

Conclusions

Dimdazenil of 2.5 mg provided significant benefit on sleep maintenance and sleep onset in individuals with insomnia disorder versus placebo, with a favorable safety profile and was well tolerated.

Clinical Trial Information

A multicenter, randomized, double-blind phase III clinical study evaluating the efficacy and safety of EVT201 capsules compared to placebo in patients with insomnia disorders (http://www.chinadrugtrials.org), with the number of CTR20201068.

Graphical Abstract

Graphical Abstract.

Statement of Significance.

Insomnia is a highly prevalent sleep disorder, affecting approximately 10%–30% of the global adult population. While various treatment options are available, there are unmet needs in promoting sleep maintenance and minimizing next-day functional impairments. Dimdazenil is a novel partial positive allosteric modulator of the GABA_A receptor for the treatment of insomnia. In this randomized controlled trial, Dimdazenil could significantly improve objective and subjective measures of sleep maintenance and sleep onset in insomniacs. Compared to placebo, Dimdazenil increased total sleep time by 31.68 minutes and reduced sleep latency by 6.46 minutes as evidenced by polysomnography. These improvements were also supported by subjective assessments. Importantly, Dimdazenil exhibited a favorable safety profile, without any adverse impact on sleep architecture or daytime functioning. Given the chronic nature of insomnia, our results support Dimdazenil as a promising therapeutic option, with sustained efficacy and an absence of tolerance, withdrawal symptoms, or abuse potential. The findings warrant long-term studies and highlight Dimdazenil’s potential to reduce the public health impacts imposed by insomnia through improved sleep and next-day function.

Introduction

Insomnia is one of the most common sleep concerns, with the main symptoms of initial insomnia (sleep onset > 30 minutes), persistent sleep disorder (number of wakes overnight ≥ 2), early awakening, decreased sleep quality, and reduced total sleep time (TST; usually < 6.5 hours), accompanied by daytime dysfunction [1, 2]. The daytime dysfunction caused by insomnia primarily includes fatigue, depression or irritation, physical discomfort, and cognitive impairment.

In 2011, Roth et al. reported that the prevalence of insomnia in the general population in the United States was 22.1% (DSW-IV standard) [3]. According to the survey results published by the Chinese Sleep Research Society, the incidence of insomnia in China was 38.2% [4]. The treatment methods for insomnia mainly include drug therapy and non-drug therapy. Currently, the drugs for the clinical treatment of insomnia mainly include benzodiazepine receptor agonists (BZRAs), melatonin receptor agonists, orexin receptor antagonists, and antidepressants with hypnotic effects [1, 5]. Early benzodiazepines are generally full agonists with limitations such as next-day aftereffects, tolerance, harmful withdrawal signs and symptoms, and serious alcohol interactions. Current benzodiazepines have shorter hypnotic durations and generally show a good effect on inducing sleep, but a poor effect on maintaining sleep. An ideal drug for insomnia should have the following characteristics: inducing sleep, maintaining and improving sleep quality, without aftereffects on the next day [6, 7]. Therefore, improving sleep persistence remains an unmet need for patients with insomnia.

Dimdazenil is a novel partial positive allosteric modulator of the GABA_A receptor, which selectively acts on the α₁β₂γ₂ GABA_A receptors with high affinity and moderate levels of maximum receptor activation. Full GABA_A receptor agonists may induce desensitization of the target receptors, which can lead to tolerance and subsequent withdrawal symptoms, and the fact of associated side effects has led to interest being focused on partial GABA_A receptor agonists as potential therapeutics [8, 9]. Dimdazenil produces a low maximum potency strength against the GABA_A receptors (Supplementary Table S1). Hence, Dimdazenil exerts rapid inhibitory effects on the nervous system by selectively activating GABA_A receptors, thereby minimizing the potential for excessive receptor activation [10].

Dimdazenil had been initially developed by Evotec (as EVT 201) for the treatment of insomnia. The phase I and II clinical studies were conducted in adult and elderly healthy volunteers, and patients with insomnia. Dimdazenil is currently being developed in China by Zhejiang Jingxin Pharmaceutical Co., Ltd. with Evotec for the treatment of insomnia disorder [10] as new drug of category 1. This pivotal study was conducted to investigate the efficiency and safety of Dimdazenil compared with placebo in individuals with insomnia disorder.

Methods

Study design

This was a 2-week, randomized, multicenter, double-blind, placebo-controlled, parallel-group phase III study conducted at a total of 66 hospitals in China. The trials consisted of a screening period (14 days), a single-blind placebo run-in period (7 days), a double-blind treatment period (14 days), a single-blind placebo run-out period (3 days), followed by a safety follow-up period (7 days).

The study was approved by the appropriate health authority, ethics committee, or institutional review board for each participating site. All relevant protocol amendments involved in this study were approved by the appropriate authorities before implementation. The study adhered to good clinical practice guidelines, the Declaration of Helsinki, and local regulations. All study participants provided written informed consent before any screening procedures. This study was registered on the official Drug Clinical Trial Information Management Platform in China (http://www.chinadrugtrials.org), with the number CTR20201068. The study protocol remained unchanged following the initiation of the study.

Participants

Adult (aged ≥18 years) men and women recruited from hospital outpatient clinics with insomnia disorder meeting Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria were eligible for the study, an additional inclusion criterion was a self-reported history of disturbed sleep including: (1) duration of disease ≥3 months, and (2) >30 minutes to fall asleep or total subjective sleep ≤6 hours at least 3 nights per week within 1 month prior to screening.

Participants were required to meet the following inclusion criteria simultaneously: (1) over the last 3 months, they must have a normal bedtime between 9 pm and 1 am and spend at least 7 hours in bed; (2) at least 2 nights of polysomnography (PSG) evaluation conducted during the screening period, and mean TST ≥ 240 and ≤ 390 minutes. A full list of exclusion criteria is provided in Supplementary Table S2.

Randomization and masking

Participants were enrolled by researchers and randomly assigned (2:1) to receive Dimdazenil capsules or placebo capsules. Randomization was stratified by age (<65 and ≥65 years), conducted using a computer-generated scheme, which was reviewed and approved by an independent statistician. The centralized randomization process was facilitated through an interactive response system.

A randomization list was generated by Clinflash Healthcare Technology (Jiaxing, Zhejiang, China) and remained confidential until after database lock. Participants, investigators, site personnel, and sponsor-authorized personnel were unaware of treatment allocation. Only participants were masked to treatment during placebo run-in and run-out periods. Investigational treatment and matching placebo were indistinguishable and packaged in the same way. In the event of a medical emergency, investigators were permitted to initiate the unmasking process; no unmasking events occurred in either arm of the trial.

Procedures

The study treatment was divided into three phases: placebo run-in period, double-blind treatment period, and placebo run-out period. During the 7-day placebo run-in period, PSG assessments were conducted on 2 consecutive nights for eligible participants with completed sleep diary and questionnaire to define baseline values. During the double-blind treatment period, participants were administered orally with Dimdazenil capsules of 2.5 mg or placebo 15–30 minutes before going to bed for 14 days. Participants attended 2 consecutive nights of PSG recording and filled in a sleep diary and questionnaire on day 13/14. TST (defined as time of total sleep duration), sleep efficiency (SE; defined as the ratio of TST to bedtime), wake after sleep onset (WASO; defined as time spent awake between the onset of persistent sleep and lights on), and latency to persistent sleep (LPS; defined as time from lights off to persistent sleep) [11] were obtained through overnight PSG monitoring and averaged over 2 nights. A 3-day placebo run-out period followed the double-blind treatment period. The subjective measures of sTST (TST), sSE (SE), sSL (sleep latency), sWASO (WASO), and sNAW (number of awakenings) were obtained from the daily recorded sleep diaries. Insomnia severity index (ISI) was evaluated using baseline and post-treatment scales [12].

Outcomes

The primary efficacy measure was the TST monitored by PSG on 13/14 nights of the double-blind treatment period. Secondary efficacy outcome measures included (1) LPS, SE, and WASO monitored by PSG on 13/14 nights; (2) subjective sTST, sSL, and sWASO assessed from sleep diary and the patient-rated ISI assessed during the double-blind treatment period. Safety evaluation indicators included the frequency and severity of adverse events (AEs), vital signs, physical examination, intraocular pressure test, laboratory test indicators, and ECG.

Other evaluation indicators included sleep architecture assessed by PSG; withdrawal rebound evaluated through Benzodiazepine Withdrawal Symptom Questionnaire scores [13, 14] on day 18, drug residue effect analyzed of wakefulness, alertness, and attention recorded in the sleep diary, attention, and working memory function evaluated by The Digit Symbol Substitution Test (DSST) [15], daytime sleepiness evaluated by Epworth Sleepiness Scale [16], cognitive function evaluated by Trail Making Test A (TMT-A) [17], depression and anxiety evaluated by Hamilton Depression Scale (HAMD) [18] and Hamilton Anxiety Scale (HAMA) [19], and degree of fatigue evaluated through Flinders Fatigue Scale [20]. No changes to trial outcomes were made after the trial commenced.

Statistical analysis

Study sample size was calculated based on the primary efficacy measure TST. According to American Academy of Sleep Medicine Clinical Practice Guideline 2017 for chronic insomnia in adults [21], it was clinically significant to prolong TST by 20 minutes compared with placebo. In this study, a minimum of 107 participants were required for placebo and a minimum of 214 participants were required for Dimdazenil, calculated at a 2:1 ratio with α = 0.025 (one-sided) and power = 80%. Considering regulations on sample size and dropout rates, 360 participants are planned for Dimdazenil and 180 participants for placebo, with an enrollment total of 540 participants, including a minimum of 30% (162 cases) of elderly patients.

For the purposes of analysis, full analysis set is defined as all randomized participants who received at least one dose of the double-blind study medication and have at least one post-treatment efficacy assessment. Descriptive statistics were used for continuous variables and ANOVA was utilized for group comparisons. Categorical variables were summarized by frequency and Fisher’s exact test was used for inter-group comparison. The overall evaluation of enrolled patients was based on the judgment of efficacy, and subgroup analysis was performed for adult patients and elderly patients.

The primary efficacy endpoint was the average TST monitored by PSG on night 13/14 of the double-blind treatment period, based on an analysis of covariance model with TST as the response variable, baseline TST, and age as covariates and group as independent variables. The 95% confidence interval (CI) for the difference between the adjusted means of the two groups was estimated; if the lower limit of the 95% CI was greater than 0, the superiority was considered valid. The analysis methods for LPS, SE, and WASO monitored by PSG on the 13/14 night, subjective sTST, sSL, sWASO, sNAW, and ISI of the double-blind treatment period were the same as those used for the primary efficacy measure, irrespective of the central effect. Other indicators such as sleep architecture, withdrawal rebound, drug residual effects, DSST, Epworth, TMT-A, HAMD, HAMA, and fatigue were analyzed by Wilcoxon or ANOVA. Safety endpoints were analyzed in all participants who had received at least one dose of treatment. All statistical analyses were done using SAS software (version 9.4). No interim analysis or stopping guidelines were applied in this trial.

Results

Patient disposition and baseline demographics

The trial was halted when we reached our predetermined sample size. The first posted date was June 10, 2020 on Center for Drug Evaluation, National Medical Products Administration. Between August 7, 2020 and September 29, 2021, a total of 1098 participants were screened, of whom 546 were randomly assigned to receive Dimdazenil of 2.5 mg (n = 367), or placebo (n = 179). In all, 526 participants (96.3%) completed the double-blind treatment, 524 participants (96.0%) completed the whole study, while 22 participants (4.0%) withdrew (Figure 1). Most participants were women (366, 67.2% of the 546 participants). Elderly participants (≥65 years) were 86 (15.8%). Demographic and basic characteristics were comparable across treatment groups (Table 1). No events resulting in unmasking occurred throughout the treatment period.

Figure 1.

The study flow diagram.

Table 1.

Baseline Demographics and Baseline Characteristics

	Dimdazenil (n = 367)	Placebo (n = 178)
Age, years	47.3 (15.3)	45.9 (14.7)
Age group, years
≥65, n (%)	59 (16.08)	27 (15.17)
<65, n (%)	308 (83.92)	151 (84.83)
Sex
Male, n (%)	122 (33.24)	57 (32.02)
Female, n (%)	245 (66.76)	121 (67.98)
BMI, kg/m2	22.71 (2.91)	22.93 (2.91)
TST, min	335.85 (37.09)	335.00 (40.61)
LPS, min	58.81 (43.20)	57.49 (41.54)
SE, %	66.65 (8.27)	66.35 (9.05)
WASO, min	119.88 (52.98)	125.83 (56.76)
sSL, min	82.92 (48.55)	77.02 (49.18)
sTST, min	300.94 (72.02)	305.84 (66.53)
sSE, %	69.75 (16.32)	69.87 (14.27)
sWASO, min	70.89 (59.12)	74.72 (60.24)
sNAW	1.90 (0.97)	2.14 (1.095)
ISI, score	15.0 (4.7)	14.3 (4.9)

Variables expressed as mean (SD). BMI, body mass index; TST, total sleep time; LPS, latency to persistent sleep; SE, sleep efficiency; WASO, wake after sleep onset; sSL, subjective sleep latency; sNAW, subjective number of awakenings; ISI, insomnia severity index.

Efficacy outcomes

The clinical outcomes of the study are presented in Table 2. For the primary endpoint, the increase of TST monitored by PSG on 13/14 nights was larger and statistically significant for Dimdazenil compared with baseline (71.09 minutes [95% CI: 65.4368, 76.7418], p < 0.001) and placebo (LSM difference 31.68 minutes [95% CI: 22.2167, 41.1420], p < 0.001). Dimdazenil reduced LPS compared with baseline (21.65 minutes [95% CI: −25.8696, −17.4205], <0.001) and placebo (LSM difference −6.46 minutes [95% CI: −12.0424, −0.8783], p = 0.023). Dimdazenil increased SE compared with baseline (13.26% [95% CI: −12.1522%, 14.3655%], p < 0.001) and placebo (LSM difference 5.55% [95% CI: 3.7055, 7.3919], p < 0.001). Dimdazenil reduced WASO compared with baseline (49.67 minutes [95% CI: −54.9714, −44.3638], p < 0.001) and placebo (LSM difference −20.16 minutes [95% CI: −28.6071, −11.7131], p < 0.001). No significant differences were observed among participants for sNAW.

Table 2.

Primary and Secondary Outcomes

	Dimdazenil (n = 367)	Placebo (n = 178)	P-value
Primary endpoints
Day 13/14 TST, min	406.68 (55.84)	376.54 (58.85)	<0.001
LSM difference compared with baseline (95% CI)	32.05 (20.8333, 43.2682)	—	—
*p-value	<0.001	—	—
Secondary endpoints
Day 13/14 LPS, min	36.99 (32.62)	42.87 (39.11)	0.023
LSM difference compared with baseline (95% CI)	−6.46 (−12.0424, −0.8783)	—	—
*p-value	<0.001	—	—
Day 13/14 SE, %	79.91 (10.93)	74.22 (11.80)	<0.001
LSM difference compared with baseline (95% CI)	5.55 (3.71, 7.39)	—	—
*p-value	<0.001	—	—
Day 13/14 WASO, min	69.44 (52.56)	93.02 (57.44)	<0.001
LSM difference compared with baseline (95% CI)	−20.16 (−28.61, −11.71)	—	—
*p-value	<0.001	—	—
Day 13/14 sSL, min	52.90 (30.30)	54.66 (33.40)	0.049
LSM difference compared with baseline (95% CI)	−4.23 (−8.4318, −0.0238)	—	—
*p-value	<0.001	—	—
Day 13/14 sTST, min	362.91 (59.61)	347.28 (59.20)	<0.001
LSM difference compared with baseline (95% CI)	18.33 (10.09, 26.57)	—	—
*p-value	<0.001	—	—
Day 13/14 sWASO, min	41.11 (38.73)	57.66 (49.44)	<0.001
LSM difference compared with baseline (95% CI)	−14.60 (−19.95, −9.24)	—	—
*p-value	<0.001	—	—
Day 13/14 sNAW	1.35 (0.80)	1.79 (1.11)	<0.001
LSM difference compared with baseline (95% CI)	−0.29 (−0.4083, −0.1755)	—	—
*p-value	<0.001	—	—
Day 13/14 ISI, score	11.0 (5.09)	12.6 (5.20)	<0.001

Variables expressed as mean (SD). CI, confidence interval; LSM, least squares mean; TST, total sleep time; LPS, latency to persistent sleep; SE, sleep efficiency; WASO, wake after sleep onset; sSL, subjective sleep latency; ISI, insomnia severity index. *Two-sided p-value versus baseline.

Compared with placebo, Dimdazenil also improved key self-reported measures of sleep including sTST (LSM difference 18.33 minutes [95% CI: 10.0909, 26.5716], p < 0.001), sWASO (LSM difference −14.60 minutes [95% CI: −19.9533, −9.2438], p < 0.001), sSL (LSM difference −4.23 minutes [95% CI: −8.4318, −0.0238], p = 0.049), sSE (2.97% [95% CI: 1.4210, 4.5271], p < 0.001), and sNAW (LSM difference −0.29[95% CI: −0.4083, −0.1755], p < 0.001). Participants treated with Dimdazenil reported a significant improvement in ISI (p < 0.001).

The subgroup analysis of PSG TST is shown in Figure 2. Except for the subgroup with a sleep latency of ≤30 minutes, the lower limit CI of the 95% difference in TST for all other subgroups was consistently above zero. This indicates that the TST in the Dimdazenil group was significantly longer than in the placebo group, with statistical significance. The results of point estimation were all >20 minutes, which is a significant threshold for evaluating effectiveness in PSG TST versus placebo [21].

Figure 2.

Subgroup analysis of PSG TST. TST on each subgroup for Dimdazenil was longer than placebo (>20 minutes). The lower limit CI of the 95% difference in TST was below zero in sleep latency ≤ 30 minutes subgroup (highlighted). *p < 0.05; DSM, Diagnostic and Statistical Manual of Mental Disorders; PSG, polysomnography; CI, confidence interval.

Sleep architecture

The evaluation of sleep architecture is presented in Table 3. No statistically significant differences between Dimdazenil and placebo for any of the various indicators of sleep architecture at baseline. On day 13/14 of the double-blind treatment period, the duration of stage 2 sleep was significantly longer in Dimdazenil group compared with placebo group; whereas the duration of stage 3 sleep and rapid eye movement (REM) sleep were significantly shorter than placebo, but no change from baseline (Figure 3). The stage 2 sleep time, REM SL, and the percentage of stage 2 sleep time in TST were significantly prolonged in Dimdazenil group compared with the placebo group. The stage 3 sleep time, REM sleep time, and the percentage of stage 3 sleep time in TST and the percentage of REM sleep time in TST were significantly reduced in Dimdazenil group compared with the placebo group. In the sleep structure of normal adults, the percentage of stage 2 sleep time in TST is about 45%–55%, and the percentage of stage 3 sleep time in TST is about 10%–20%, of which stage 3 sleep time decreases with age [22].

Table 3.

Sleep Architecture Characteristics

Sleep time, min	Dimdazenil (n = 367)	Placebo (n = 178)	P-value
Baseline
Sleep latency of stage N1	47.05 (42.31)	42.07 (34.69)	0.173
Sleep latency of stage N2	51.58 (43.79)	48.57 (36.06)	0.426
Sleep latency of stage N2	85.51 (59.32)	83.06 (58.72)	0.652
Sleep latency of REM	114.86 (60.41)	114.00 (55.35)	0.873
Non-REM sleep time
Stage N1	41.31 (22.10)	38.93 (22.62)	0.243
Stage N2	184.57 (42.34)	182.62 (42.84)	0.616
Stage N3	50.65 (33.14)	54.21 (34.90)	0.248
REM sleep time	59.76 (21.06)	58.73 (21.38)	0.595
Percent of sleep time for stage N1, %	12.54 (6.63)	11.94 (7.14)	0.332
Percent of sleep time for stage N2, %	54.76 (11.10)	54.13 (10.73)	0.532
Percent of sleep time for stage N3, %	15.58 (10.76)	16.34 (10.47)	0.436
Percent of sleep time for REM, %	17.54 (6.17)	17.28 (5.94)	0.640
Day 13/14
Sleep latency of stage N1	30.73 (31.13)	33.74 (38.75)	0.335
Sleep latency of stage N2	34.40 (32.70)	38.04 (40.57)	0.267
Non-REM sleep time
Stage N1	39.76 (23.50)	38.86 (24.16)	0.678
Stage N2	252.30 (58.63)	205.26 (50.23)	<0.001
Stage N3	50.58 (38.61)	59.54 (38.15)	0.012
REM sleep time	63.34 (25.91)	72.82 (24.82)	<0.001
Percent of sleep time for stage N1, %	9.99 (6.12)	10.75 (7.11)	0.200
Percent of sleep time for stage N2, %	62.03 (11.23)	54.79 (14.07)	<0.001
Percent of sleep time for stage N3, %	12.52 (9.38)	16.05 (10.30)	<0.001
Percent of sleep time for REM, %	15.29 (5.69)	19.31 (6.52)	<0.001

Variables expressed as mean (SD). REM, rapid eye movement.

Figure 3.

Analysis of sleep architecture at day 13/14. *p < 0.05; REM, rapid eye movement.

Withdrawal symptoms and drug residues

No withdrawal symptoms were observed during the placebo run-out period, as assessed by AEs or the Benzodiazepine Withdrawal Symptom Questionnaire, except for one item “sensory static movement,” which was significantly lower in the Dimdazenil group than in the placebo group (p = 0.0081). No AEs suggested that drug misuse might have occurred (Supplementary Table S3 and S4). For drug residues, the Dimdazenil group had better wakefulness on days 2/3/5/7, alertness on days 3/6, and attention on days 2/5/6/13 than the placebo group; all other differences were not statistically significant. After 14 consecutive days of administering of Dimdazenil, no residual drug effects were observed. This may be attributed to the relatively short half-life of Dimdazenil, which is approximately 4 hours [23].

Evaluation of fatigue and other scales

The total fatigue scores of Dimdazenil and placebo groups were comparable at baseline, but during the double-blind treatment period, the total fatigue scores of the Dimdazenil group decreased compared to the placebo group, with a statistically significant difference (p = 0.0142). In addition, regarding the question “Is fatigue a problem that bothers you?” “Does fatigue also affect your daily functioning (e.g. work, family, social)?,” “Does fatigue make you feel miserable?,” “How often do you feel fatigued?,” and “How badly do you feel fatigued?” were all lower than the placebo group (higher scores suggesting more severe/frequent) and the differences were all statistically significant (all p < 0.05), suggesting that the fatigue level was reduced in Dimdazenil group compared to the placebo group (Supplementary Table S5). There were no differences between the two groups during the double-blind treatment period in attention and working memory function assessed by DSST, daytime sleepiness assessed by Epworth, cognitive function assessed by TMT-A (Supplementary Table S6–S8), and anxiety, depression assessed by HAMD, HAMA (Supplementary Table S9 and S10).

Safety

The safety analysis set comprised 545 participants (n = 367 for Dimdazenil group, n = 178 for placebo group). Study drug exposure was similar across treatment groups, with 100% of participants having 2 weeks of exposure for Dimdazenil and placebo, respectively. A similar incidence of treatment-emergent adverse events (TEAEs) was observed across the two treatment groups (Table 4), with the majority of TEAEs mild or moderate in severity. The most frequently reported TEAE was dizziness, with incidence of 12.53% in the Dimdazenil group and 5.06% in the placebo group (p = 0.006). Headache was also more common in the Dimdazenil group (4.36%) compared to the placebo group (2.25%), but the difference was not statistically significant (p = 0.331). Dimdazenil was well tolerated. No deaths occurred and no treatment-related serious adverse events (SAEs). There were only three participants who discontinued the study drug owing to a TEAE in Dimdazenil group (dizziness, vertigo, and skin rash) while one participant was in placebo group (electrode hypersensitivity). There were no clinically important mean changes from baseline for any hematology parameter, chemistry laboratory parameter, urinalysis parameter, vital sign parameter, electrocardiogram parameter, or weight at any study visit. The incidence of markedly abnormal laboratory values or vital signs was low and similar across groups. In addition, no suicidal ideation, suicidal behavior, or self-injurious behavior were observed during the study.

Table 4.

Adverse Events by Treatment Group

N (%)	Dimdazenil (n = 367)	Placebo (n = 178)	P-value
TEAE	174 (47.41)	71 (39.89)	0.100
Serious TEAE	2 (0.54)	2 (1.12)	0.600
Reduction of platelet count	1 (0.27)	0 (0.00)
Rib fracture	0 (0.00)	1 (0.56)
Diabetic foot	1 (0.27)	0 (0.00)
Spontaneous abortion	0 (0.00)	1 (0.56)
TEAE rate more than 2% in any group
Hyperuricemia	3 (0.82)	4 (2.25)	0.472
Vertigo	11 (3.00)	5 (2.81)	1.000
Headache	16 (4.36)	4 (2.25)	0.331
Dizzy	46 (12.53)	9 (5.06)	0.006
Ocular hypertension	10 (2.72)	7 (3.93)	0.441
TRAE	110 (29.97)	39 (21.91)	0.052
TRAE rate of more than 2% in any group
Vertigo	10 (2.72)	5 (2.81)	1.000
Dizzy	36 (9.81)	8 (4.49)	0.043
Ocular hypertension	9 (2.45)	7 (3.93)	0.417

TEAE, treatment emerged adverse event; TRAE, treatment-related adverse event.

Discussion

The improvements in sleep variables with Dimdazenil included both a reduction in sleep latency and an improvement in sleep maintenance. Amplitude of effect and dose was similar in adults (<65 years) and elderly (≥65 years). Referring to the clinical practice guidelines of the American Academy of Sleep Medicine on chronic insomnia in adults in 2017 [21], there is a clinical significance if TST is prolonged by 20 minutes compared with the placebo, in line with treatment goals for the management of chronic insomnia [24, 25]. This appropriate estimation of total sleep by the participants might reflect preservation of memory and, the proportion of time spent in different sleep stages is preserved, in contrast to findings reported with full benzodiazepine receptor agonists [26, 27]. The effects noted with partial benzodiazepine receptor agonist were achieved without excess sleepiness, and, in contrast, might be complemented by an improvement in daytime functioning.

Compared to men, women are at a higher risk of insomnia and exhibit poorer sleep quality [28]. In this study, a trend was observed where women showed a tendency towards better improvement in TST compared to men, although it did not reach statistical significance. In the phase I clinical trial of Dimdazenil, although no distinct trend was evident, women consistently exhibited slightly higher exposure levels, including C_max and AUC, when compared to healthy men [23]. This may potentially account for the observed improved therapeutic efficacy among women.

The study also analyzed the effect of sleep structure and found that on day 13/14 of the double-blind treatment period. The percentage of stage 3 sleep in the total sleep was 12.52% ± 9.38% in the test group and 16.05% ± 10.30% in the placebo group, respectively. Although the percentage in the Dimdazenil group was lower than that in the placebo group, it was basically within the normal range.Overall, administering Dimdazenil for 2 weeks prolonged stage 2 sleep in adult patients with insomnia disorder, without reducing stage 3 sleep and REM sleep. This indicates that Dimdazenil maintained basic sleep structure, unlike currently available BZDs which typically increase stage 2 sleep and decrease stage 3 sleep duration, along with reducing REM sleep duration during nocturnal sleep [22]. The increased stage 2 sleep is associated with a subjective improvement in sleep quality by reducing micro-awakenings [29], it not only presents a relevant number of micro-awakenings and is related to the restorative quality of sleep [30], but also may be related to an increased risk of cardiovascular issues [31]. The decrease in stage 3 sleep is usually associated with lesser “rest” for the brain, which leads to a lack of concentration, because considered as one stage characterized by a slow-wave sleep (SWS) [29]. In the present review, SWS reduction was observed in all studies using short-acting BZDs (1.5–4 hours). REM sleep contributes to the consolidation of declarative memories providing emotional facilitation [22]. Zolpidem has been reported to increase SWS but not affecting stage REM in patients with insomniac [32]. However, complex sleep behaviors such as sleepwalking, sleep terror, and abnormal nocturnal behavior typically arise from slow-wave (N3) NREM sleep [33]. Studies of orexin receptor antagonists in healthy participants and patients with insomnia suggest that they increase the time spent in both REM and NREM sleep [34]. Dual orexin receptor antagonists increase TST primarily by promoting REM sleep, without affecting, or even decreasing NREM sleep. Patients with major depressive disorder should avoid medications that prolong the REM sleep because it is characterized by trait-like increases in REM sleep [35]. The impact of different insomnia medications on sleep architecture remains uncertain and controversial in the scientific community, and further research is needed.

Dimdazenil, as a partial agonist, might lower the possibility of some adverse behavioral effects relative to full agonists (particularly at high doses) theoretically [36, 37]. In this study, the incidence of TEAEs was 47.41% (174/367) and 39.89% (71/178) in the Dimdazenil group and placebo group, respectively (P = 0.1); the incidence of drug-related TEAEs was 29.97% (110/367) and 21.91% (39/178), respectively (P = 0.052). There were no significant differences in safety events between Dimdazenil and placebo, which may be attributed to the novel selectivity and partial agonist specificity of Dimdazenil as a new BZD. Classic benzodiazepines modulate GABA_A neurotransmission by binding to a site that traverses an α and γ subunit of α1, α2, α3, or α5 containing GABA_A receptors [38]. The GABA_A receptor carrying the α1 subunit is believed to be the mediator of the sedative and amnesic effects of BZDs. The anxiolytic, myorelaxant, motor-impairing, and ethanol-potentiating effects are attributed to GABA_A receptor, carrying other subunits (α subunits 2, 3, and 5). Currently, available BZDs are nonselective for GABA_A receptors with different α subunits [39]. In contrast, Dimdazenil demonstrates higher selectivity for the α1 subunit, exhibiting approximately three to four times greater affinity compared to α2, α3, and α5. Additionally, Dimdazenil produces lower maximal efficacy. These selective and partial agonist properties provide the advantage of exerting sustained and stable hypnotic effects with fewer adverse reactions. Furthermore, they can help avoid excessive central nervous system suppression and the development of tolerance associated with long-term use, thereby reducing the potential for withdrawal symptoms [40, 41].

The incidence of SAEs was 0.54% (2/367) in the Dimdazenil group and 1.12% (2/178) in the placebo group. The whole SAEs were recovered/resolved, except that the outcome of decreased platelet count was returned to baseline. This patient exhibited undiagnosed and untreated low platelet counts upon enrollment. Effective treatment resulted in the successful restoration of platelet levels and led to a definitive diagnosis of secondary immune thrombocytopenia associated with thyroid disease. A key strength of this study was the assessment of most components of insomnia, as defined in DSM-5 [42]. In the trials, not only were polysomnography-based nighttime variables measured but also participant-reported subjective assessments of night and day symptoms were incorporated, with robust control for study-wise type I error for all primary and secondary endpoints.

Limitations

First, the whole participants who were randomly assigned were of Asian descent. Therefore, the randomly assigned participants might only partly reflect the racial and ethnic diversity in some regions of the world. Second, the 2-week duration does not allow for assessment of efficacy and safety of long-term use. Finally, this study did not enroll enough expected elderly participants (aged ≥65 years) of 30%, and the actual proportion was 15.7%. The effects of Dimdazenil on elderly participants will be investigated in a phase IV post-marketing study.

Conclusions

Dimdazenil of 2.5 mg provided significant benefit on sleep maintenance and sleep onset in individuals with insomnia disorder versus placebo. It demonstrated favorable safety and tolerability profiles, and did not impair daytime functioning.

Data Availability Statement

People can get a copy of trial protocol by emailing the corresponding author.

Funding

This study was funded by Zhejiang Jingxin Pharmaceutical Co., Ltd.

Author Contributions

Zhaoyang Huang, Shuqin Zhan, Chunyan Chen, and Yuping Wang were responsible for the conception and design of the study. Ruoxi Zhang contributed substantially to data acquisition. Yanling Zhou contributed to data interpretation. Jingjing He were responsible for statistical analysis. Chunyan Chen and Ruoxi Zhang drafted and revised the manuscript. Zhaocun Lin, Cungang Bao, Shuangpeng Zhu, Jianjun Zhao, and Shengan Zhang were responsible for the manufacture of the Dimdazenil. Yu Jiang and Yuping Wang supervised the manuscript. All authors contributed to the article and approved the submitted version.

Disclosure Statements

Financial disclosure: The authors declare no competing financial interests. Nonfinancial disclosure: Chunyan Chen, Ruoxi Zhang, Yanling Zhou, Jingjing He, Zhaocun Lin, Cungang Bao, Shuangpeng Zhu, Jianjun Zhao, Shengan Zhang, and Yu Jiang are full-time employees of ZHEJIANG JINGXIN PHARMACEUTICAL CO., LTD. The authors declare no competing nonfinancial interests.