Understanding the unmet needs in insomnia treatment: a systematic literature review of real-world evidence

Abstract

Study objective

The objective of this study was to define and characterize the unmet needs in the pharmacological management of insomnia.

Methods

A systematic literature review was conducted to identify relevant literature reporting real-world evidence in insomnia, published from January 2009 to April 2020. Pharmacological treatments – both prescription (benzodiazepines, ‘Z-drugs’ and suvorexant) and off-label (antidepressants, antipsychotics, and antihistamines) – were considered.

Results

Overall, 108 publications describing the humanistic (n = 59) and economic burden (n = 20) of insomnia, off-label treatment patterns (n = 28) and factors influencing treatment adherence or persistence (n = 8) were identified. A high prevalence of comorbid conditions was reported in patients with insomnia resulting in significantly lower health-related QoL compared to those with insomnia or a comorbidity alone. Current treatment options were associated with adverse events, including reduced sleep quality and next-day somnolence. An increased risk of accidents/injuries was also associated with insomnia and its treatment. Furthermore, safety concerns and perceived lack of efficacy for approved treatments have led to frequent off-label prescribing, despite a lack of clinical evidence of risk/benefit ratios. Safety concerns associated with benzodiazepines include risk of dependence, leading to prolonged treatment persistence and exacerbated adverse events, making them unsuitable for use in patients with chronic insomnia. Finally, the substantial economic burden of insomnia was evident, with reduced work productivity demonstrated in patients with insomnia compared to the general population.

Conclusions

This review highlights a clear unmet need for insomnia therapies that improve sleep quality without resulting in next-day impairment and/or dependence.

Keywords:

• Insomnia burden
• treatment patterns
• unmet needs
• real-world evidence

Introduction

Insomnia is a common sleep disorder characterized by difficulty falling asleep, maintaining sleep, early morning awakenings or a combination of both sleep onset and maintenance difficulties [1]. It is associated with fatigue, mood disturbances, difficulty concentrating and reduced performance at work or at school [1].

In patients who suffer from psychiatric or behavioral disorders (i.e. ADHD), cardiovascular disease, and sleep disorders (namely obstructive sleep apnea (OSA)), the prevalence of insomnia is significantly higher than in the general population [2,3]. Moreover, health-related quality of life (HRQoL) is significantly impaired in patients with insomnia and comorbid conditions have a significant additive effect on HRQoL [4]. Conversely, the presence of insomnia can have an additive effect on reduced HRQoL in patients with comorbidities [4].

Insomnia is also related to a high economic burden, in terms of work absences and reduced work productivity [5]. Workplace accidents associated with insomnia, and adverse events associated with some insomnia treatments, contribute to lost work productivity and the subsequent high indirect costs for patients with the disorder [6,7].

In addition to increasing the risk of accidents/injuries, prescription drugs available for the treatment of insomnia, including the sedative hypnotic drugs benzodiazepines (such as triazolam and temazepam) and non-benzodiazepines or ‘Z-drugs’ (zaleplon, zolpidem and eszopiclone), are associated with tolerability issues [8–11] and are not recommended for prolonged use due to adverse psychomotor effects, cognitive symptoms, sleep symptoms that carry forward to the next day and potential for addiction [11]. Despite these well-documented adverse events, a high number of patients continue to use benzodiazepines on a long-term basis, typically as a result of developing tolerance and subsequent dependence on these drugs [12].

Concerns over the side effect profile of currently approved treatments for insomnia has led to frequent off-label prescribing of antidepressants, antipsychotics and some antihistamines for patients with insomnia, despite a lack of evidence regarding their safety and efficacy and evidence of serious adverse events [13]. For example, antidepressants are associated with anticholinergic effects and orthostatic hypotension, while antipsychotics can lead to metabolic syndrome and extrapyramidal effects, including involuntary movements, tremors or muscle contractions [14]. There is therefore an unmet need for new pharmacological therapies for patients with insomnia, which improve sleep quality without the safety concerns of current treatment options.

The objective of this study was to conduct a systematic literature review (SLR) to define and characterize the unmet needs in the pharmacological management of insomnia for new therapies in clinical practice.

Materials and methods

Data sources and collection process

A literature search was conducted to capture all relevant publications reporting on the real-world evidence of insomnia, following the Cochrane Handbook of Systematic Reviews of Interventions (CHSRI) guidelines [15]. Using the OVID platform, the following databases were searched: EMBASE, Ovid MEDLINE®, EBM Reviews, PsychINFO and EconLit. The search was also extended to review the bibliographies of selected articles to identify additional publications. Details of the OVID search strategy are provided in Table S1.

Table 1. Summarized results from common direct costs outcomes reported in studies by Bolge et al. [52,64,65].^a

Population	Sample size (n)	Emergency room visits, mean (SD)	Days hospitalized, mean (SD)	Traditional medical provider visits, mean (SD)
General population split into CINA sufferers and non-insomnia sufferers	25,629	CINA: 0.4 (0.9) No insomnia: 0.1 (0.7) -p < 0.001	CINA: 0.8 (4.2) No insomnia: 0.3 (2.4) -p < 0.001	CINA: 8.0 (11.9) No insomnia: 3.3 (4.6) -p < 0.001
Patients diagnosed with depression or anxiety split into CINA suffers and those without insomnia	4,510	Depression with CINA: 0.5 (1.1) Depression without CINA: 0.3 (1.1) -p < 0.001 Anxiety with CINA: 0.5 (1.2) Anxiety without CINA: 0.3 (1.0) -p < 0.001	Depression with CINA: 0.8 (3.8) Depression without CINA: 0.6 (4.8) -p = 0.410 Anxiety with CINA: 1.0 (3.4) Anxiety without CINA: 0.5 (3.6) -p = 0.025	Depression with CINA: 9.7 (10.6) Depression without insomnia: 6.0 (7.1) -p < 0.001 Anxiety with CINA: 10.6 (11.4) Anxiety without insomnia: 6.4 (7.2) -p < 0.001
Females experiencing symptoms of menopause split into CINA suffers and those without insomnia	1,446	Experiencing CINA: 0.3 (0.6) No insomnia: 0.1 (0.5) -p < 0.001	Experiencing CINA: 0.6 (2.7) No insomnia: 0.2 (1.1) -p = 0.001	Experiencing CINA: 5.5 (5.6) No insomnia: 3.8 (4.8) -p < 0.001

CINA: chronic sleep maintenance insomnia characterized by night-time awakenings; SD: standard deviation.

^aUnadjusted results are reported in the table. Results adjusted by linear regression models are also reported in the original publications.

Proceedings of relevant conferences published from January 2017 to April 2020 were also reviewed.

Study selection

Studies of real-world evidence in insomnia published from January 2009 to April 2020 were included based on the PICOS (Population, Intervention, Comparator, Outcome, Study design) criteria for inclusion and exclusion (Table S2) to address the key research questions to understand the humanistic and economic burden, real-world treatment patterns and factors influencing treatment adherence and persistence in insomnia. Publications reporting on non-pharmacological interventions (i.e. cognitive behavioral therapy) for the treatment of insomnia were not included in this review.

Table 2. Summarized results from common indirect cost outcomes reported in studies by Bolge et al. [52,64,65].^a

Population	Sample size (n)	Absenteeism, mean (SD)	Presenteeism, mean (SD)	Overall work productivity, mean (SD)	Activity impairment, mean (SD)
General population split into CINA sufferers and non-insomnia sufferers	25,629	CINA: 8.5 (21.2) No insomnia: 2.8 (12.3) -p < 0.001	CINA: 34.4 (25.4) No insomnia: 10.4 (19.2) -p < 0.001	CINA: 30 (30.8) No insomnia: 9.0 (20.3) -p < 0.001	CINA: 52.0 (29.1) No insomnia: 16.8 (24.9) -p < 0.001
Patients diagnosed with depression or anxiety split into CINA suffers and those without insomnia	4,510	Depression with CINA: 10.4 (22.0) Depression without CINA: 4.2 (14.0) -p < 0.001 Anxiety with CINA: 12.2 (25.2) Anxiety without CINA: 5.3 (16.4) -p < 0.001	Depression with CINA: 40.9 (24.0) Depression without CINA: 20.5 (23.7) -p < 0.001 Anxiety with CINA: 42.7 (23.8) Anxiety without CINA: 20.0 (24.2) -p < 0.001	Depression with CINA: 33.6 (31.4) Depression without CINA: 17.0 (25.1) -p < 0.001 Anxiety with CINA: 37.4 (32.5) Anxiety without CINA: 17.4 (26.6) -p < 0.001	Depression with CINA: 59.6 (26.3) Depression without CINA: 35.5 (30.3) -p < 0.001 Anxiety with CINA: 61.5 (26.1) Anxiety without CINA: 32.5 (30.5) -p < 0.001
Females experiencing symptoms of menopause split into CINA suffers and those without insomnia	1,446	CINA: 4.3 (13.2) No insomnia: 1.7 (9.1) -p = 0.089	CINA: 31.7 (26.1) No insomnia: 9.2 (17.5) -p < 0.001	CINA: 29.0 (23.3) No insomnia: 8.6 (16.2) -p < 0.001	CINA: 47.8 (29.4) No insomnia: 16.9 (24.9) -p < 0.001

CINA: chronic sleep maintenance insomnia characterized by night-time awakenings; n: number; SD: standard deviation.

^aUnadjusted results are reported in the table. Results adjusted by linear regression models are also reported in the original publications.

Titles and abstracts of retrieved publications were reviewed by two researchers independently and abstracts that met the PICOS criteria were carried forward to full-text publication screening and data extraction.

Study bias and quality control assessments

All full-text publications were assessed for bias using the Risk of Bias in Non-Randomized Studies of Interventions (ROBINS-I) questionnaire [16]. Abstract-only publications were not assessed for risk of bias due to insufficient information. Within-study risk of bias was assessed qualitatively (i.e. either low, moderate, serious, critical or unclear risk of bias among key domains, presented visually with color-coding) [15].

The quality of each full-text publication was assessed using one of the following four National Institute of Health and Care Excellence (NICE) checklists for: quantitative intervention studies, quantitative correlation and association studies, qualitative studies and economic evaluations. The checklist used to assess each study was determined based on the key outcomes reported.

Results

Overview of search results

The OVID search identified 3,725 abstracts, of which 118 relevant full-text publications were screened and 76 publications were included for data extraction (Figure 1). A further 1,517 conference abstracts were identified, of which 31 were included for data extraction. One publication was identified following review of the bibliographies of selected articles (Figure 1).

Figure 1. PRISMA diagram of the systematic literature review.

Study characteristics and objectives

Of the 108 publications extracted, 59 reported on the humanistic burden and 20 on the economic burden of insomnia, 28 reported on real-world off-label treatment patterns of insomnia, while eight publications discussed the factors influencing treatment adherence and persistence. The majority of the publications identified were from North America (n = 48) and the Asia-Pacific region (n = 28). An overview of the heterogeneity in definitions used for insomnia across the publications and the outcomes reported in each study identified by region is provided in Table S3.

Risk of bias and quality control assessments

Fifty-eight full-text publications were assessed using the ROBINS-I questionnaire. Overall, the studies presented low or unclear risk of bias. The area where the highest level of bias risk was identified across the publications was where confounding factors were not sufficiently controlled.

Fifteen publications were assessed using the NICE Quantitative Intervention Study quality assessment, with no publications judged to show a risk of bias and most findings being generalizable to the source population [13,17–30]. The NICE Quantitative Correlation and Association study checklist was used to assess 31 publications [6,7,31–59], with only one publication judged not to have findings generalizable to the source population [40]. Two publications were assessed using the NICE Qualitative Study appraisal checklist and both were deemed to have an appropriate theoretical approach, study design, data collection and analysis method and were judged to be trustworthy [60,61]. Finally, 10 publications identified were assessed with the NICE Economic Evaluation checklist and were mostly judged to have minor limitations [62–71].

Overall, the studies were deemed to demonstrate low or unclear risk of bias and no major quality issues were identified in the full-text publications assessed for this review.

The humanistic burden of insomnia

Insomnia significantly impairs QoL compared to patients without the disorder, as demonstrated in a Japanese cohort study (n = 2,822), in which sleep medications (class of medication not reported) were associated with a further reduction in patient QoL, measured by both physical component summary (PCS) and mental component summary (MCS) (Figure 2) [58].

Figure 2. QoL (mean MCS and PCS scores) of patients with insomnia compared to good sleepers and the impact of using sleep medication.

Along with adverse events of insomnia treatments, the symptoms associated with insomnia itself and the high prevalence of comorbid conditions in patients with insomnia, or vice versa, may contribute to the deterioration in QoL observed in patients with insomnia, as described in the sections below.

The patient impact of insomnia and its comorbidities

Psychiatric or behavioral disorders

Overall, 19 publications reported an increased prevalence or severity of psychiatric or behavioral disorders (i.e. ADHD) in patients with insomnia, or vice versa, and a decrease in patient QoL (Table S4) [34,39,46,50–54,64,72–81].

Patients with insomnia demonstrated an increased likelihood of developing a psychiatric condition, such as depression and anxiety, and suffered from exacerbated symptoms of the comorbidity, compared to those without insomnia, across 12 studies identified (Table S4) [34,39,46,51–53,72–74,77–79]. Notably, a cross-sectional study of 6,899 adolescents in Portugal demonstrated that individuals with insomnia had significantly more severe depression, compared to those with normal sleep (18.53 ± 12.34 versus 7.43 ± 8.41, p < 0.001), and subsequently had significantly reduced QoL scores (mean HRQoL = 85.0 [SD 19.0], p < 0.001), measured using the SF-36 [51]. Furthermore, insomnia was significantly associated with suicidal ideation in a US study of 113,185 student athletes (OR = 1.95, 95% CI [1.86, 2.06]; p < 0.0001) and in a study of 432 adults in South Korea (mean [SD] = 4.61 [5.28] in insomnia subgroup versus non-insomnia subgroup: 2.53 [3.74]; p < 0.05), highlighting the significant impact of insomnia on a patient’s mental health [75,76].

In addition to negatively impacting overall QoL, four studies demonstrated a significant reduction in cognitive performance in patients with psychiatric or behavioral (i.e. ADHD) disorders and insomnia, which was further exacerbated in those with more severe insomnia (Table S4) [50,54,80,81]. A cohort study of 28,485 individuals in Canada reported significantly higher rates of anxiety (24.8% versus 7.9%, p < 0.001) and depression (38.3% versus 15.3%, p < 0.001) in patients with insomnia compared to those without and these patients had poorer processing speed, executive function and memory [54]. Similarly, a study of 70 university students in Germany demonstrated a significant association between ADHD and insomnia (p < 0.001) and that students with both insomnia and ADHD had significant rates of impaired attention compared to those without ADHD (p = 0.045) [81].

Diabetes and cardiovascular diseases

Five publications reported a significant association between insomnia, the prevalence of diabetes, cardiovascular comorbidities and reduced patient QoL (Table S5). Similar to psychiatric conditions, the prevalence of cardiovascular conditions has been found to be significantly higher in patients with chronic insomnia compared to those without insomnia (congestive heart failure: 5.4% versus 1.3%, p < 0.001; peripheral vascular disease: 1.2% versus 0.4%, p < 0.001), as demonstrated in a nationwide US survey with 25,629 respondents [65]. From this survey, SF-8 physical and mental component scores were 6.2 and 6.8 points lower than those without insomnia, respectively (p < 0.001) [65]. Conversely, a high prevalence of insomnia was reported in patients with type 2 diabetes and cardiovascular conditions, such as hypertension, with these populations experiencing significantly reduced QoL scores across several domains (Table S5; Figure 3) [56,82].

Figure 3. QoL of patients with comorbid insomnia and type 2 diabetes mellitus compared to those without insomnia.

Cognitive function

Moreover, a cohort study using data from the Penn State Adult Cohort (N = 1,741) in the US demonstrated that chronic insomnia with objective short sleep duration (<6 h) was significantly associated with an increased risk of cognitive impairment compared to the reference group (OR: 2.65 [95% CI: 1.28–5.51]) [83]. This risk was further increased in those with vascular cognitive impairment, defined as cognitive impairment with cardiometabolic risk factors or cerebrovascular disease (OR: 3.15 [95% CI: 1.42–7.00]) [83].

One publication reported on QoL outcomes for patients with comorbid obstructive sleep apnea (OSA) and insomnia [84]. From 61 patients with OSA in Australia, 18 had concomitant insomnia and these patients had significant cognitive impairment compared to patients with OSA alone, as demonstrated by more lapses in a choice reaction time test (p = 0.001) and a psychomotor vigilance test (p = 0.01) [84].

The evidence above demonstrates the need for an effective therapy to treat insomnia symptoms and subsequently reduce the severity of comorbid conditions to improve QoL.

Relationships between comorbidities and increased insomnia severity or persistent insomnia

Across nine publications, substance abuse, psychiatric and behavioral disorders were associated with an increase in insomnia severity compared to those without these comorbidities (Table S6) [17,46,49,60,73,81,85–87].

Insomnia symptoms were more severe in patients with substance abuse disorders compared to those without, with an increased severity of insomnia symptoms reported in patients with opioid use compared to alcohol use [49,73]. Similarly, patients with psychiatric or behavioral (i.e. ADHD) disorders suffered from more severe insomnia, as indicated by increased scores on the Athens Insomnia Scale (AIS) and Insomnia Severity Index (ISI), compared to patients with insomnia alone [81,85,86]. Furthermore, two studies indicated the presence and severity of psychiatric comorbidities were associated with insomnia persistence (Table S6) [46,60]. In one study, mental disorders were associated with higher rates of persistent insomnia symptoms compared to those without mental disorders (42.0% versus 31.0%) [60]. In the other, individuals with persistent insomnia symptoms had higher depression symptom severity and lower physical and mental health QoL (i.e. mean (SD) Beck Depression Inventory values were 13.94 (9.58) for the persistent insomnia group versus 8.05 (6.78) for the group without persistent insomnia (p < 0.0001) [46].

Association between current insomnia treatments and adverse symptoms that can negatively affect patients’ QoL

A summary of the adverse events reported across the literature identified, for each treatment type, is provided in Table S7 [17,27,29,31,33,35,36,61,92–95]. Benzodiazepines were reported to be associated with daytime hangover effects and impaired motor coordination (flurazepam and quazepam) and psychological dependence and rebound insomnia (triazolam), while non-benzodiazepines were associated with balance and memory impairments (zolpidem), and somnolence and dizziness (suvorexant) [17,27,61]. These adverse events can increase the risk of accidents and injuries in patients with insomnia, as described in the subsection below.

Numerous therapies have been associated with adverse events that can negatively impact sleep quality, despite an indication to treat insomnia [17,27,61]. For example, suvorexant has been associated with insomnia itself, triazolam with rebound insomnia and zolpidem with unusual night-time behaviors [17,27,61]. Furthermore, a cohort study in Taiwan (n = 3,235) demonstrated an increased risk of sleep-related breathing disorders in patients using hypnotics, which was significantly increased in those receiving high doses compared to low doses (OR: 1.678 [95% CI: 1.051–2.680], p = 0.030) [29].

The risks associated with high doses of hypnotics were further highlighted in two retrospective cohort studies (n = 5,593 and n = 7,862) which reported an increased risk of dementia in patients receiving hypnotics, such as benzodiazepines, with the greatest increase in those receiving higher doses compared to lower doses (HR: 2.10 [95% CI: 1.63–2.72] and 1.53 [95% CI: 1.15–2.05], p < 0.001), respectively) [31,35]. This study emphasizes that the use of high doses of hypnotics (i.e. benzodiazepines) should be discouraged [31,35].

Association between insomnia and its treatment with risk of accidents

Both insomnia and use of certain insomnia therapies (benzodiazepines and Z-drugs) have been associated with an increase in the risk of accidents (e.g. falls and motor vehicle accidents) [7,28,55,61,92–95] The risk of accidents was reported to be higher in patients with more severe insomnia symptoms [92]. A Norwegian cohort study reported that unintentional fatal injuries were significantly more common in patients who experienced more frequent symptoms of insomnia (i.e. often or almost every night, compared to never or occasionally) or a greater number of symptoms, after controlling for the use of sleep medications (p < 0.001 and p = 0.001, respectively) [92].

A number of studies reported an increased risk and higher incidence of both motor vehicle accidents and accidents at home in patients with insomnia [7,55,94,95]. From 2,946 car accidents reported, in a cross-sectional study across ten countries, 2,365 (80.17%) involved patients who had been diagnosed with insomnia, based on the DSM-IV criteria, and who had not received treatment in the previous six months [7]. In the same study, the incidence of accidents at home was higher in patients with insomnia, with 4,009 (76.88%) of a total 5,207 accidents at home reported occurring in patients with insomnia compared to 1,207 accidents in patients without (p = 0.01) [7].

Insomnia treatments have been associated with serious adverse events (as described above), which significantly increase the risk of falls resulting in fractures and motor vehicle accidents [61]. In a retrospective study of 4,603 patients living with dementia, the non-benzodiazepine Z-drugs (e.g. zolpidem, zoplicone and zaleplon) were associated with an increased risk of fractures (HR: 1.47 [95% CI: 1.10–1.97]) and this risk was further increased by higher treatment doses (p = 0.008) [93]. Furthermore, a cohort study in the US reported a significantly increased risk of falls in hospital in 1,595 patients administered zolpidem compared to 1,942 patients not administered zolpidem (OR: 4.37 [95% CI: 3.34–5.76], p < 0.001) [28]. This risk remained significant after adjusting for variables such as other medication use, age and sex (OR: 6.39 [95% CI: 3.07–14.49], p < 0.001) [28]. Overall, zolpidem has been reported to increase the risk of fractures and major injuries leading to hospitalization [28,61,93].

Workplace accidents

Two publications demonstrated an increased risk of workplace accidents in patients with insomnia compared to the general population, contributing to significantly reduced work productivity as detailed in the economic burden section below.

A significantly higher incidence of workplace accidents was reported in patients with insomnia (80.2%; n = 2,529), who had not received treatment in the previous six months, compared to those without (p = 0.001), in a cross-sectional study of 3,302 reported accidents at work [7].

Furthermore, a Swedish cohort study of women aged ≥20 years reported a significant increase in the risk of occupational accidents in those with persistent insomnia (n = 541) (OR: 1.5 [95% CI: 1.2–2.0], p = 0.002) compared to those with insomnia recorded only at baseline (n = 473) (OR: 0.9 [95% CI: 0.6–1.2], p = 0.446) or follow-up (n = 564) (OR: 1.2 [95% CI: 0.9–1.6], p = 0.180) [6]. No information on hypnotic use was collected for this study, however, the authors recognized that treatment with hypnotics is a potential cause of accidents in patients with insomnia [6].

Real-world treatment patterns in insomnia

Twenty-eight publications were identified that reported on frequent off-label treatment use for insomnia, despite a lack of evidence for their efficacy [18–23,25,30,38,40,42,61,63,91,96–108]. Several publications highlighted concerns surrounding off-label prescribing, such as the lack of clinical evidence of risk/benefit ratio and evidence of serious adverse events [18,21,30,106]. Although publications highlighted these concerns, no studies explored the side-effect profile, efficacy or the potential risks associated with off-label prescribing for insomnia.

The studies identified reporting off-label prescribing in insomnia covered multiple geographical regions, as illustrated in Figure 4, with almost half of the publications identified (n = 15) published in North America [21,22,38,63,91,97–104]. Globally, antidepressants (including trazodone, mirtazapine and citalopram) were the most commonly studied off-label treatments for insomnia (Figure 4).

Figure 4. Geographic distribution of publications reporting data on off-label treatment of insomnia stratified by drug-class and region.

Thirteen publications reported on the use of antidepressants in patients with insomnia (Table S8) [13,19,20,22,23,38,61,63,91,100,103,104,107]. Trazodone was the most frequently prescribed anti-depressant in a longitudinal study using data from the 2006 US National Ambulatory Medical Care Survey [13]. From 30.43 million outpatient visits analyzed, 45.2% of patients were prescribed off-label antidepressants, with 17.9% of patients prescribed trazodone (see Table S8) compared to 43.2% prescribed non-benzodiazepines/Z-drugs and only 11.7% prescribed benzodiazepines [13]. Furthermore, a cross-sectional study in the US analyzing 8,000 patients per year between 2011 and 2016 found the annual percentage of patients who were dispensed low-dose trazodone increased, while the percentage of patients who were dispensed approved treatments (i.e. zolpidem or eszopiclone) decreased (Table S8) [104]. This demonstrates the preference of physicians to prescribe off-label treatments for insomnia despite guidelines recommending treatments, such as Z-drugs [104].

Ten publications examined the off-label use of antipsychotics for the treatment of insomnia (Table S8), with nine of these reporting off-label prescribing of quetiapine, despite the lack of evidence for its use in insomnia and the well-documented metabolic adverse events [18,21,30,42,97–99,105,106]. In a retrospective study conducted in Canada, 13% of patients diagnosed with insomnia and no other psychiatric comorbidities (n = 517) received antipsychotics, 85% of which were prescribed quetiapine [97]. The evidence highlights that quetiapine is frequently prescribed off-label for insomnia, even in patients without psychiatric comorbidities.

Although Z-drugs and benzodiazepines are approved for the treatment of insomnia, these treatments are often used outside their product licenses, with chronic use or prescription of higher than recommended doses considered off-label use [109]. Chronic use of hypnotic/sedative drugs for insomnia was demonstrated in a cross-sectional study in Italy, which reported that 52 (64.2%) of a total 81 patients had been taking hypnotics/sedatives for more than 3 years (Figure 5) [40]. Overall, seven publications examined the off-label use of hypnotic/sedative drugs for insomnia and are summarized in Table S8.

Figure 5. Duration of benzodiazepine use for patients with insomnia across six regions in Italy.

Off-label use of antihistamines was reported in four studies (Table S8), covering prescription and over the counter (OTC) off-label treatments [23,25,91,108]. A retrospective study in Saudi Arabia reported 44% of patients with insomnia (n = 307) were treated with antihistamines or an antihistamine–analgesic combination product [23]. The authors speculated that this demonstrates a lack of awareness of treatment guidelines by physicians or a lack of concern regarding side effects [23].

Factors that may influence adherence, persistence and discontinuation of insomnia treatments

Overall, eight publications reported on the adherence (defined as patient behavior corresponding with recommendations of a health care professional [110]), persistence (defined as continuing the treatment for the prescribed duration [111]) and discontinuation of insomnia treatment (Table S9) [17,24,26,27,87,103,112,113].

In a retrospective study in Japan, suvorexant discontinuation at one month was measured and compared between patients who switched from benzodiazepine receptor agonists to suvorexant (switching group: n = 119) to those who were taking suvorexant at the same time as benzodiazepine receptor agonists (add-on group: n = 109) [24]. A significantly higher proportion of patients discontinued in the add-on group (45.0%) compared to the switching group (24.4%) (p < 0.001) [24]. “Intolerability” was the most common reason for discontinuation in the add-on group, with 22% of patients discontinuing for this reason compared to 7.6% in the switching group (p < 0.002) [24]. Somnolence on the following day was the main reported reason for a lack of tolerance [24]. In contrast, the most common reason for discontinuation in the switching group was inefficacy (15.1%) [24].

Both the EMA and FDA recommend that benzodiazepines should be used occasionally for a short-term period to avoid dependence, highlighting the need for a treatment that is safe to use over a prolonged period for patients with chronic insomnia [40,114,115]. In an Italian observational study, 32 (39.5%) of 81 patients treated with benzodiazepines reported wanting to discontinue their treatment and 17 (21.0%] had previously tried to stop taking benzodiazepines, but experienced withdrawal symptoms [40]. Subsequently, all 32 patients continued benzodiazepine treatment for longer than three years [40]. These findings highlight the importance of short-term use of benzodiazepines to prevent adverse symptoms, such as dependence and subsequent withdrawal, occurring once a patient discontinues their treatment.

The severity of symptoms in patients with insomnia is one explanation for frequent inappropriate/chronic use of medication, as patients develop a tolerance to the treatment over time leading to a need for dose escalation to experience a similar effect [37]. A longitudinal cohort study in the Netherlands (n = 401) demonstrated that insomnia severity was significantly associated with more problematic use of benzodiazepines (β = 0.145, p = 0.0004), as measured by the benzodiazepines dependence subscale [37].

In a cohort study of 9,049 patients in Canada, patients prescribed antidepressants for insomnia as a new therapy were more likely to be adherent over 90-days compared to those with a prescription for depression (OR: 0.73 [95% CI: 0.55–0.98]) [103]. However, when antidepressants were prescribed as an ongoing therapy, patients with insomnia were less likely to be adherent compared to those with depression (OR: 1.58 [95% CI: 1.25–1.99]) [103]. The authors speculated the lower adherence to antidepressants as an ongoing therapy may be due to ineffectiveness [103].

The evidence highlights the unmet need for a treatment with reduced adverse events to prevent early discontinuation as a result of tolerability issues, without the risk of dependence to avoid long-term persistence.

The economic burden of insomnia

A total of 20 publications reported on the economic burden of insomnia [43–45,47,48,52,62–65,67–71,87,116–119]. An overview of the studies identified and the economic burden outcomes reported is presented in Table S10.

The detrimental impact of chronic sleep maintenance insomnia characterized by nighttime awakenings (CINA) on healthcare resource utilization has been highlighted in three cross-sectional sub-analyses of the economic burden of CINA within different population subsets (general population, depression/anxiety sufferers, women experiencing menopause) [52,64,65]. Table 1 summarizes the direct costs reported across the studies, with significantly higher costs for all parameters in patients with CINA compared to those without (p ≤ 0.001), except for mean days hospitalized in patients with comorbid anxiety and depression (p > 0.05) [52,64,65].

Insomnia has a substantial detrimental impact on work productivity as a result of both absenteeism and lost work performance. Absenteeism, presenteeism, overall work productivity and activity impairment were investigated as measures of indirect costs in three different sub-populations with CINA (summarized in Table 2) [52,64,65]. As with the direct costs described above, each indirect cost measure was significantly higher in all CINA groups versus those without CINA (p < 0.001), except for absenteeism in women experiencing menopause (p > 0.05) [52,64,65].

Two retrospective studies reported on the risk of sick leave associated with insomnia [70,71]. In a 2009 study, participants with insomnia (n = 791) were found to have more than twice the risk of having sick days during follow-up, compared to good sleepers (crude OR: 2.20 [95% CI: 1.77–2.74]; fully adjusted OR: 1.51 [95% CI: 1.19–1.94]) [71]. Similar findings were reported in a 2013 study, where risk of sick leave was determined in a population suffering from insomnia, or comorbid insomnia and OSA, compared to good sleepers [70]. Both disorders were found to more than double the risk of having four or eight weeks of sick leave (full adjusted ORs: 2.07–2.42) [70].

A national US-based study investigated the association between insomnia and work productivity using data from The American Insomnia Survey (n = 7,428) [67]. In a sub-analyses of the study in patients with comorbidities (n = 4,991), human capital values of individual decrements in work performance was reported as $3,274 (SE: 66) before adjusting for comorbidities and $2,280 (SE: 48) after adjustment [67]. At a population level, annual number of days lost associated with insomnia were projected as 367 million days (SE: 4.2) and 252.7 days (SE: 3.0) before and after adjustment, respectively [67]. This was equivalent to $91.7 billion (SE: 8,967) and $63.2 billion (SE: 10,001) in annual human capital losses associated with insomnia, before and after adjustment, respectively [67].

Discussion

This comprehensive literature review identified substantial real-world evidence demonstrating the considerable humanistic and economic burden of insomnia and its associated treatments. Overall, most of the studies in this review generated data from North America (n = 48), Asia Pacific (n = 28) and Europe (n = 18), with limited data from the Middle East (n = 3) and Latin America (n = 2). Therefore, conducting further observational research to characterize and quantify burden of insomnia in countries outside of North America would be useful to support the unmet need for new therapies globally. For example, a survey panel to collate real-world experiences from patients with insomnia, starting with key markets (such as the United Kingdom (UK) and US), and followed by additional regions, such as Brazil, Mexico, United Arab Emirates and South Africa, should ensure a global representative sample with directly comparable outcomes.

The evidence highlights the negative impact of insomnia on patient QoL and the high prevalence of comorbid conditions (psychiatric and behavioral disorders (ADHD), cardiovascular conditions, substance abuse and OSA) in patients with insomnia leading to further deterioration in QoL [34,39,46,50–54,64,72–78,80,81,118]. An increase in the severity of insomnia symptoms was reported in patients with psychiatric comorbidities and substance abuse disorders, notably opioid addiction, highlighting a potential mechanism by which comorbidities lead to poor QoL in patients with insomnia [17,46,49,60,73,81,85–87]. This demonstrates the importance of an effective treatment in patients with insomnia and comorbid conditions to improve patient QoL. However, the direction of the relationship between insomnia and comorbidities was often unclear or was not explored by the study authors (i.e. whether the comorbidity was a result of insomnia or vice versa).

Furthermore, few publications reported whether patients with insomnia and a comorbid disorder were being treated for their insomnia or which treatments were used. Notably, studies reporting antidepressant use for the treatment of insomnia were included in this review. However, it is recognized that antidepressants may be considered as treatment for comorbidity management. It is therefore difficult to fully characterize the off-label use of antidepressants in patients with insomnia and psychiatric comorbidities.

Additional research into the impact of adverse events associated with current treatments on QoL and comorbidity control in patients with insomnia, and the potential effect of comorbid conditions on treatment response, could help understand the unmet needs and support the positioning of therapies in these subpopulations. Moreover, the ongoing opioid crisis highlights the need for further studies to investigate the efficacy of treatments for patients with insomnia associated with substance withdrawal [120]. Future study design should incorporate additional outcomes to determine treatment response in comorbid patients versus non-comorbid patients. Additional domains in QoL questionnaires and patient-reported outcome (PRO) measures to investigate the impact of treatment-related adverse events and comorbidities in patients with insomnia should also be incorporated into study design. Qualitative interview-based research could be conducted to further characterize the unmet needs relating to insomnia treatment in patients with comorbidities.

Despite having an indication to treat insomnia, benzodiazepines, Z-drugs and the orexin receptor antagonist, suvorexant, were associated with reduced sleep quality and somnolence that carries forward to the next day [40,42]. This review highlights that both insomnia itself and treatment tolerability issues are impacting on work productivity and lead to an increased risk of accidents, including falls, injuries at work and motor vehicle accidents [40,42,59,61]. Therefore, an effective treatment is required to improve sleep quality without residual adverse effects carrying forward to the next day to reduce the burden of accidents in patients with insomnia.

Concerns over the side-effect profile and perceived lack of efficacy associated with currently licensed insomnia treatments (including benzodiazepines and Z-drugs) result in substantial off-label use of other treatments, despite a lack of clinical evidence of risk/benefit ratios and evidence of serious adverse events [18–23,25,30,38,40,42,61,63,91,96–108]. Specifically, antidepressants such as trazodone were often prescribed for treating insomnia, despite a lack of evidence for their efficacy and evidence they can lead to anticholinergic events and orthostatic hypotension [14]. Moreover, concerns over the highly addictive properties of benzodiazepines prevent health care professionals prescribing this approved treatment for insomnia [11]. Patients with insomnia who are prescribed benzodiazepines, or Z-drugs, often develop a tolerance leading to the use of higher than approved doses and a longer treatment duration to achieve the desired effect. This can result in dependence and subsequent exacerbation of adverse events (e.g. impaired motor coordination). Therefore, effective and approved long-term treatments for insomnia are needed, which are considered by health care professionals to have an acceptable safety profile compared to off-label treatment options. Specifically, the introduction of a therapy without the risk of dependence would allow its prolonged use to ameliorate symptoms in patients with chronic insomnia.

In addition to a substantial humanistic burden, insomnia is associated with significantly higher direct costs, especially through increased health care utilization, and indirect costs, due to significantly reduced work productivity, compared to the general population [17]. Notably, the increased risk of accidents suggests that both direct and indirect costs may be much higher than those reported in the current evidence base. Further research is required to quantify the cost of insomnia and to determine the key outcomes that should be assessed when investigating the economic impact associated with insomnia. Consistency in the reporting of economic outcomes across regions and populations would be beneficial for economic evaluations when considering potential new therapies for the treatment of insomnia. A cost-of-illness study to quantify the economic burden of insomnia and the associated treatment-related adverse events would be beneficial to quantify the cost of insomnia in a consistent manner.

This review had a number of strengths including the application of an up-to-date, rigorous systematic approach to identify published studies reporting on the real-world evidence of insomnia and comorbidities of interest. The limitations of this research include the selective use of published data, inclusion of publications reported in English, the robustness of some studies identified and small patient populations in some studies. Additionally, there was significant heterogeneity in the definitions used to classify insomnia across the publications identified making it difficult to draw comparative conclusions across the data in this review.

Conclusion

This SLR examined the real-world evidence for the burden and current treatment patterns for insomnia, providing an overview of the landscape in terms of humanistic and economic burden, off-label prescribing and real-world adherence and persistence outcomes. Although causality is unclear, the reporting of a high prevalence of comorbidities (including psychiatric/behavioral disorders (i.e. ADHD), cardiovascular conditions and substance abuse) and impact on QoL highlights the importance of considering these conditions in the management of insomnia. Notably, the evidence demonstrates both insomnia and treatment tolerability issues are impacting on work productivity and lead to an increase risk of accidents. Moreover, concerns over the side effect profile and perceived lack of efficacy associated with currently licensed insomnia treatments result in substantial off-label use of other treatments despite a lack of clinical evidence of risk/benefit ratios and evidence of serious adverse events. Overall, the findings of this research highlight the necessity for new therapies that improve sleep quality without resulting in next-day impairment and/or dependence.

Acknowledgments

There are no acknowledgements to be made.

Disclosure statement

This work was funded by Eisai Inc., and Eisai was involved with all stages of the study conduct and analysis.

Funding

Eisai Inc. was the funding source and was involved with all stages of the systematic literature review.

Data availability statement

No new data were generated or analyzed in support of this research.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.