Does insomnia treatment prevent depression?

Abstract

Rates of major depressive disorder (MDD) are increasing globally, in part due to the coronavirus disease 2019 pandemic, contributing to disease burden. It has long been known that insomnia is intricately connected with depression as indicated by greater depression severity and lower treatment response. Furthermore, insomnia is a significant risk factor for new-onset depression. Treatment of insomnia is thus a logical target for prevention of incidents and recurrent MDD. This systematic review sought to evaluate the current evidence for the preventive effects of insomnia treatment on depression onset. A database search yielded 186 studies, six of which met criteria for inclusion in this review. All of the studies utilized cognitive behavioral treatment for insomnia (CBT-I) as the target intervention and most delivered treatment via a digital platform. Four of the studies found significantly lower rates of MDD onset in those who received CBT-I compared to a control condition. The two remaining studies failed to confirm these effects in primary analyses but secondary analyses suggested evidence of a preventive effect. There was significant methodologic heterogeneity across studies in terms of sample selection, outcomes, and follow-up periods, limiting the ability to draw firm conclusions. The evidence overall is in the direction of insomnia treatment reducing the risk for onset of MDD, but further research is warranted.

Insomnia, Depression, Cognitive Behavioral Therapy

Statement of Significance

As depression rates continue to rise globally, the ability to identify modifiable factors that can prevent or reduce rates of disorder onset is critical. Treatment of insomnia has shown to be effective in reducing depression symptom severity, and recently research has begun to explore whether improving insomnia can prevent depression onset. This review discusses the burgeoning literature in this area and highlights the future research necessary to answer questions about the true efficacy of insomnia treatment as a depression prevention strategy.

Introduction

Major depressive disorder (MDD) is a leading cause of global disease burden, poor quality of life, functional disability, and suicide [1]. Depression rates have been consistently rising, particularly during the coronavirus disease 2019 (COVID-19) pandemic. The World Health Organization estimated the global prevalence of depression to be approximately 4.4% in 2015 [2]; however, a recent meta-analysis estimated a rate of 28% following the onset of the pandemic [3]. Now more than ever, there is an urgent need for early identification, treatment, and prevention efforts to address this significant global mental health issue.

Toward that end, there have been concerted efforts in the field to identify modifiable factors that influence both disorder and treatment trajectory. Sleep disturbances, particularly insomnia symptoms, influence the course of depression treatment [4, 5]. Clinical and epidemiological research has revealed that sleep disturbance often precedes the onset of an episode of depression [6], making the presence of clinically significant sleep problems a potent risk factor. Sleep disturbance has also been associated with a greater likelihood of attenuated treatment response [4], and greater time to remission and recovery [7]. Evidence is mixed regarding the influence of sleep disturbance on the relapse to depression, with some studies demonstrating greater rates of relapse among individuals with depression and insomnia symptoms [8–10], while others fail to demonstrate such associations [7, 11–13].

There is mounting evidence that the treatment of insomnia is associated with an improvement in depression severity. Recent meta-analyses have demonstrated that insomnia treatment is associated with moderate to large decreases in depression symptoms in individuals with both depression and insomnia symptoms [14, 15]. There is also evidence that reductions in insomnia symptoms mediate these associations [16]. Importantly, improvements in depression following insomnia treatment are often maintained during longitudinal follow-up [17–19].

Together, it is clear that sleep disturbance and depression are intimately connected; when sleep disturbance is significant, individuals take longer to recover from a depression episode and may be more likely to relapse or not respond at all. When individuals with depression receive treatment for their insomnia symptoms, they are likely to experience significant improvement in their depression symptoms. What remains unclear; however, is whether insomnia treatment can prevent depression onset, a question that has great importance to public health.

In this systematic review, we begin to answer this important question by examining studies that specifically tested the preventative effects of insomnia treatment on depression episode onset. We then discuss the state of the evidence thus far and identify research gaps to be filled in future studies.

Methods

Using PubMed and PsycINFO databases, we utilized the combined search terms “insomnia + depression + treatment + prevention” to identify clinical research studies that aimed to identify whether an improvement in sleep problems prevented depression onset. The first author (EB) reviewed all search results and identified abstracts that fit the following inclusion criteria: (1) utilized a single psychotherapeutic or pharmacologic intervention to improve sleep, (2) included participants with clinically significant insomnia symptoms but without a current categorical diagnosis of MDD (studies almost unilaterally allowed for the presence of subthreshold depression symptoms; however, a point upon which we will elaborate in the discussion), (3) examined rates of incident depression following insomnia intervention, and (4) was not a review or meta-analysis. Results were categorized based on the above criteria and confirmed by the senior author (PG). Reviews were conducted in July and August of 2022.

The database search yielded 186 entries. After duplicates were removed, we were left with 168 unique entries. Following review of these abstracts, six studies met our inclusion criteria, with one additional study identified through a review of selected studies but that did not appear in the database search. This manuscript was an update of one of the studies already included in the search and provided additional results from an 18-month follow-up. Studies were excluded for at least one of the following reasons, with some meeting multiple exclusion criteria: (1) not utilizing an insomnia treatment (n = 99), (2) not examining incident depression (n = 46), (3) describing a clinical trial protocol rather than results (n = 16), and (4) utilizing a combined insomnia and depression treatment from which the effects of sleep improvement could not be isolated (n = 1).

Results

Study methods and characteristics

A review of the articles revealed a high degree of variability in methods, sample characteristics, and length of follow-up (see Table 1). Although all studies aimed to detect incident depression, they did so via a range of self-report depression measures (e.g. Patient Health Quesitonnaire-9 (PHQ-9 [20];), Edinburgh Postnatal Depression Scale (EPDS [21];); Quick Inventory of Depressive Symptoms Self-Report (QIDS-SR [22];) or clinical interviews (e.g. Structured clinical interview for the DSM-5 (SCID-5 [23],); Composite International Diagnostic Interview Short Form, CIDI [24];). Studies also differed in terms of stated outcome: some [25] aimed to assess incidence of moderate to severe depression symptoms (as opposed to categorical MDD diagnosis) while others [26] aimed to assess incident or recurrent MDD via clinical diagnostic interview. All studies required a diagnosis of DSM-5 Insomnia Disorder with the exception of one study [27] which required either a diagnosis of Insomnia Disorder or an Insomnia Severity Index score >11. Samples included adults, older adults, and pregnant women, but no study examined prevention effects in adolescents or young adults specifically. Studies largely utilized Cognitive Behavioral Therapy for Insomnia (CBT-I) delivered in a digital format (dCBT-I); n = 5), although one included therapist guidance with the digital platform [28]. Length of follow-up also varied, from 6 to 36 months, with an additional study examining depression prevention rates during the early portion of the COVID-19 pandemic (April 2020) after receiving dCBT-I in 2016/2017. All studies utilized psychotherapy, with no study explicitly examining the depression prevention effects of pharmacotherapy for insomnia. Of note, we excluded one study that examined depression severity in women randomized to trazodone, diphenhydramine or placebo during pregnancy, and framed results in terms of prevention of postnatal depression. The authors reported lower depression severity in the diphenhydramine and trazodone groups compared to placebo [29]; however, given that they did not provide rates of MDD incidence, and given that trazodone can have antidepressant as well as hypnotic characteristics, we did not include this study in our overall discussion.

Table 1.

Studies Examining the Effects of Insomnia Treatment on Depression Prevention

Author/Year	Intervention	Study population	Sample size	Depression measurement	Inclusion criteria	Outcomes	Results
Christensen et al., 2016	dCBT-I vs. health education control	Adults with insomnia disorder and subclinical depression	N = 1149 randomized; 248 completed dCBT-I; 333 completed control intervention	PHQ-9 (self-report); MINI (clinician-administered)	PHQ-9 scores between 4 and 20; absence of 2-week diagnosis of MDD via MINI; Insomnia Disorder via MINI	Depression symptoms at 6 months (PHQ-9); Current MDD at 6 months (MINI)	No significant difference in incidence of MDD between groups; PHQ-9 symptoms significantly lower in dCBT-I group
Cheng et al., 2019	dCBT-I vs. online sleep education control	Adults with insomnia disorder	N = 1358 randomized; 358 completed dCBT-I, 300 completed control intervention	QIDS-SR (self-report)	DSM-5 Insomnia Disorder; <daily or near-daily self-reported depressed mood and anhedonia via QIDS-SR	Incidence of moderate to severe depression at 12 months (QIDS-SR ≥ 11)	Incidence of moderate-to-severe depression was 18.8% in control vs. 9.6% in dCBT-I.*
Cheng et al., 2021	dCBT-I vs. attentional control	Adults with insomnia disorder	N = 208 randomized (106 to dCBT-I, 102 to control).	QIDS-SR (self-report)	Participated in 2019 trial; Identical initial inclusion criteria	Odds of moderate-to-severe depression at beginning of COVID-19 pandemic (2 to 3 years post trial completion)	Odds of moderate-to-severe depression symptoms during COVID was 57% lower in dCBT-I.
Irwin et al., 2021	In-person CBT-I group vs. sleep education therapy	Older adults with insomnia disorder (60 and older)	N = 291 randomized; 140 completed CBT-I, 130 completed educational control.	SCID-5 (clinician-administered)	DSM-IV Insomnia Disorder; Absence of DSM-IV or DSM-5 MDD within last 12 months	Incident or recurrent MDD at 36 months	MDD was present in 12.2% in CBT-I group and 25.9% in the sleep education control at 36 months. Annual incidence of MDD was 4.1% in the CBT-I condition and 8.6% in SET.
Felder et al., 2022	dCBT-I vs. standard prenatal care with waitlist control	Pregnant women (up to 28 weeks gestation) with clinically significant insomnia symptoms	N = 208 randomized; 68 completed 6 sessions of CBT-I; 103 allocated to standard care	EPDS (self-report)	DSM-5 Insomnia Disorder or ISI ≥ 11; lack of probable MDD (EPDS < 15)	Incidence of probable MDD (EPDS ≥ 13) at 3 and 6 months	At 3 months, 3.9% had probable MDD vs. 18.4% in standard care. No significant effects at 6 months.†
Leerssen et al., 2022	Therapist-guided dCBT-I vs. CRT+dCBT-I vs. CRT vs. TAU	Adults with insomnia disorder and a subtype at high risk for depression	N = 132 randomized; 35 to TAU; 31/34 completed CRT; 29/31 completed CBT-I; 29/32 completed CRT+CBT-I	Depression severity: IDS-SR; MDD Incidence: CIDI-SF (clinician-administered; blinded interviews)	ITQ type 1, 2 or 3 (high risk for depression); DSM-5 Insomnia Disorder; ISI ≥ 10; no current MDD via CIDI-SF	Change in depression severity over 12 months (Reliable Change Index using IDS-SR); Incidence of MDD at 12 months (CIDI-SF)	Therapist-guided dCBT-I + CRT intervention significant reduced 1-year incidence of clinically meaningful worsening of depression symptoms compared to control; no group differences in MDD cases at 12 months

^*Analyses conducted on subset of participants who reported minimal to no depression at baseline (QIDS-SR ≤ 11; n = 339).

^†Analyses conducted on subset of participants who reported minimal to no depression at baseline (EPDS < 10; n = 143).

CBT-I, Cognitive Behavioral Therapy for Insomnia; dCBT-I, digital Cognitive Behavioral Therapy for Insomnia; CIDI-SF, Composite International Diagnostic Interview – Short Form; CRT, Circadian Rhythm Treatment; EPDS, Edinburgh Postnatal Depression Scale; IDS-SR, Inventory of Depression Symptoms Self-Report; ISI, Insomnia Severity Index; ITQ, Insomnia Type Questionnaire; MINI, Mini International Neuropsychiatric Interview; PHQ-9, Patient Health Questionnaire-9; QIDS-SR, Quick Inventory for Depression Symptoms – Self-Report; SCID-5, Structured Clinical Interview for the DSM-5; TAU, Treatment as Usual

Effects of insomnia treatment on prevention of incident depression

Of the articles reviewed, only one study examined the effects of CBT-I delivered in its original, in-person, therapist-guided format. In this study, 291 older adults with insomnia but with no concomitant MDD diagnosis, were randomized to receive either eight sessions of group CBT-I delivered by a therapist or eight sessions of sleep education therapy delivered by a public health educator over the course of 2 months [26]. Incident depression cases were then documented by diagnostic interview at 6-month intervals over the course of 36 months of follow-up. Incident depression rates were roughly doubled in the education control condition over 36 months of follow-up (12.2% of participants in the CBT-I condition vs. 25.9% in the control condition). Importantly, the authors found that remission of insomnia, not merely exposure to targeted insomnia treatment, was an important predictor of depression prevention. Hazard ratios for incident depression were more than triple those among participants who received CBT-I but did not achieve remission (HR = 0.59) compared to those who received CBT-I and experienced a remission of their insomnia (HR = 0.17).

Comparable findings were observed in a separate study that utilized digital CBT-I (dCBT-I), where the core modules of CBT-I (e.g. stimulus control strategies, cognitive restructuring of negative beliefs about sleep, sleep restriction therapy) were delivered via automated platforms that do not require the presence of a trained therapist [25]. This study analyzed the data of 658 individuals with a DSM-5 diagnosis of insomnia disorder. Those who reported daily or near-daily symptoms of depressed mood or anhedonia that would meet criteria for MDD were excluded from participation; however, subthreshold symptoms of depression were permitted in this study. Participants were randomized to receive either digital CBT-I through the internet platform Sleepio (www.sleepio.com, Big Health Ltd); (n = 358) where participants received 6 core sessions of CBT-I over the course of 12 weeks of access, or to an online sleep education control condition, where participants (n = 300) received six weekly emails that provided psychoeducation about sleep and sleep hygiene guidelines. The authors then examined the incidence of moderate-to-severe depression at 1-year follow-up among the 339 participants who reported absent or minimal depression symptoms at baseline. At follow-up, the incidence of moderate-to-severe depression was 9.6% in the dCBT-I condition, and nearly double that in the control condition (18.8%). In other words, receiving digital CBT-I nearly halved the risk of developing moderate-to-severe depression compared to the educational control, findings that are consistent with the in-person intervention study. Similarly, the authors reported that the response to digital CBT-I was significantly linked to the odds of developing depression at follow-up. Greater risk of depression at 1-year follow-up was observed among participants whose response to digital CBT-I was not clinically significant or who did not experience remission of their insomnia.

Cheng and colleagues extended their analysis of the preventative effects of CBT-I on depression by examining how previous treatment may have conferred resilience to depression onset during the early months of the COVID-19 pandemic [30]. During April and May of 2020, the investigators contacted 102 participants who had received the dCBT-I intervention, and 101 participants who received the control intervention and measured their depression symptoms using the Quick Inventory for Depression Symptomatology. Using a cutoff of >10 on the Quick Inventory For Depression Symptomatology to indicate the presence of moderate-to-severe depression, the authors found that the odds of more severe depression were 57% lower among participants who had received dCBT-I than in the education control condition.

There is also some evidence of a preventative effect of dCBT-I on depression during the postpartum period, a time of particularly high risk for depression onset in women [31]. Felder et al. [27] reported on the effects of dCBT-I on prevention of postnatal depression in 208 pregnant women with insomnia symptoms (either a diagnosis of Insomnia Disorder or ISI scores >11) but without a diagnosis of probable major depression (EPDS score of ≤15). Patients randomized to the active condition received 6 weekly sessions of Sleepio (n = 105) while those randomized to the waitlist control condition received usual care for prenatal insomnia (n = 103), which included no constraints on seeking medications or therapy outside of the study protocol. Waitlist control participants received access to the dCBT-I treatment after study completion. At 3 months postpartum, probable MDD was reported in approximately 4% of the dCBT-I group and 18% in the usual care group. When examining only participants who had minimal depression symptoms during the prenatal period (i.e. those whose EPDS scores were <10 at baseline) 18% in the usual care group developed postpartum depression compared to 0% in the dCBT-I group, demonstrating a particular robust difference among those with minimal depression at study outset.

Findings from other studies are inconsistent with the data reviewed thus far. In the first and largest study to date to examine the impact of insomnia treatment on the prevention of incident MDD, 1149 adults with insomnia and PHQ-9 scores between 4 and 20 were randomized to receive either 6 weeks of dCBT-I through the SHUTi platform, or an attention-matched placebo condition called HealthWatch, which although digitally delivered did not provide any sleep or mental health content [32]. Using the Mini International Neuropsychiatric Interview (MINI [33];) to detect cases of MDD, the authors found no difference in the rates of depression diagnoses at 6-month follow-up. When examining PHQ-9 scores rather than MINI-defined diagnoses; however, the authors found fewer moderate or moderate-to-severe depression cases among patients randomized to the dCBT-I condition than in the control condition. The discrepant findings between this and the studies of dCBT-I described above could be related to several factors including differences in the digital platform itself and the fact that the authors reported that they observed far fewer cases of depression than they had expected during the follow-up period. This study also included individuals who did not meet diagnostic criteria for MDD but who had moderate to severe symptoms at baseline (approximately 32% of the sample) which also may have influenced the discordant findings. This study also employed a shorter follow-up period (6 months in this study vs. 12 and 36-month follow-up periods, respectively); however results of 18-month follow-up from this same sample were subsequently published, and rates of moderate-to-severe depression (e.g. PHQ-9 scores ≥ 10) continued to be significantly higher in the control group than the dCBT-I condition (19% vs. 2%, respectively [34]). No rates of MINI-defined MDD episodes were reported for the 18-month follow-up.

The most recent study of the preventative effects of insomnia treatment on depression selected participants based on whether they reported a subtype of insomnia considered to be at greater risk of depression (e.g. types 1–3 on the Insomnia Type Questionnaire [35] which consist of high pre-sleep arousal, high negative affect, and reduced subjective happiness) without reporting symptoms consistent with a current diagnosis of MDD [28]. One hundred thirty-two participants were randomized to receive either dCBT-I (iSleep [36];), a chronobiological intervention that focused on increasing bright light exposure and physical activity and body warming interventions (iCycle [37]; considered an active control condition), a combination of CBT-I and chronobiological intervention, or a treatment as usual condition, where participants did not receive CBT-I or chronobiological treatment but were allowed to obtain help from outside physicians or specialists (use of outside treatments was monitored by study staff). At 1-year follow-up, MDD was diagnosed in 6.5%, 11.4%, 12.1%, and 17.2% of the combined, treatment as usual, chronobiological and CBT-I conditions, respectively, though these groups did not differ statistically. The authors did find; however, that the combined chronobiological and CBT-I intervention was superior to the no-treatment and chronobiological interventions alone in preventing clinically significant worsening of depression symptoms over 1 year of follow-up. Conversely, both CBT-I and the combined CBT-I/chronobiological intervention were associated with reductions in depression symptoms greater than both the active and control conditions [28].

Discussion

The research to date demonstrates that individuals without categorical diagnoses of MDD who receive CBT-I for their insomnia symptoms tend to have lower subsequent rates of moderate-to-severe depression than those who receive no treatment or placebo conditions. To answer the question of whether insomnia treatment can prevent depression, however, it may be helpful to discuss the current literature in terms of primary prevention (e.g. applying insomnia interventions prior to the development of any symptoms of depression) and secondary prevention (e.g. intervening to reduce the symptomatic worsening and clinical impact of depression). In the present analysis, we evaluated studies that explicitly excluded participants who had categorical diagnoses of MDD. However, studies almost invariably included participants with subthreshold symptoms of depression, some in the moderate-to-severe range [32]. Given the symptom overlap in depression and insomnia, it is understandable that at least some depression symptoms may be present (e.g. fatigue, poor concentration) even in absence of the cardinal symptoms of depressed mood and/or anhedonia. However, there was generally wide variability in the range of subthreshold symptom severity allowable in each study and how these symptoms were measured. Thus, when taken as a whole, the literature to date makes it difficult to determine whether insomnia treatment leads to primary prevention of depression, given that only one study required euthymic mood at study entry [26], albeit with significant preventative effects for CBT-I.

What we may be observing more robustly in the literature; however, is evidence of significant secondary prevention of depression following insomnia intervention. The studies that evaluated depression prevention among only those participants with minimal depression at baseline demonstrated that subsequent rates of depression in the moderate-to-severe range were significantly lower in the active condition, suggesting that CBT-I may have halted the progression of mild symptoms into more severe clinical presentations. Indeed, whether insomnia treatment can prevent clinical worsening was the aim of Leerssen et al. [28], which also reported significant secondary prevention effects. As more research is conducted in this area, recruitment of samples that include both euthymic and mild depression symptoms followed by a stratified analysis can help answer focused questions concerning primary or secondary prevention of depression. Reducing and/or managing heterogeneity of depression symptoms will also better enable future meta-analyses to answer these questions quantitatively.

It is also possible that limitations inherent to digital interventions as well as the presence of subthreshold symptoms in these studies may have impacted the observed effects of insomnia treatment on depression prevention. As mentioned previously, the studies reviewed here almost universally utilized digital psychotherapy for insomnia. Digital interventions have a number of advantages including improved access and scalability, options for personalizing the timing of intervention (e.g. at odd hours when in-person therapy may not be offered), and some evidence of cost-effectiveness [38]. However, disadvantages include high attrition rates [39], lack of therapist alliance, which is a known factor connected to positive outcomes [40], and some evidence that dCBT-I in particular may not be equivalent to clinician-administered CBT-I in efficacy [41]. Indeed, large dropout rates were observed in the reviewed studies that utilized dCBT-I. Research from in-person administration of CBT-I has indicated that depression may bias an individual toward dropout [42]; however, data are not readily available across studies to discern whether dropout rates were biased toward those individuals with greater depression severity, even if that severity is within the subclinical range. Thus, it is possible we may not have observed the full magnitude of the impact of insomnia treatment on depression prevention given some of the limitations of dCBT-I as well as the presence of mild, and sometimes moderate-to-severe, levels of depression symptoms at baseline. The use of therapist guidance with the dCBT-I intervention in Leerssen et al. [28] appears to have mitigated dropout in that study, and maybe a promising approach, particularly if subthreshold depression symptoms are present. Research is ongoing to evaluate the superiority of a stepped-care approach in insomnia treatment, i.e. beginning with dCBT-I and then switching to a therapist-delivered format in the event of unsatisfactory progress [43]. Results of this work may be particularly applicable to the treatment of individuals with insomnia and depression symptoms.

Another key question that remains is for whom insomnia treatment is most likely to prevent the development of depression. Not all individuals with insomnia are equally likely to develop depression, an issue that was central to Leerssen et al.’s [28] use of the Insomnia Type Questionnaire to distinguish high- from low-risk insomnia types. To truly test whether an insomnia intervention can prevent depression, and to ensure that interventions are offered to those who need them most, high-risk designs are critically needed. Thus, the Leerssen et al. [28] study represents an advance in the field, as it attempted to evaluate whether insomnia treatment can prevent depression in those who, by virtue of their insomnia symptom presentation, may be more likely than others to develop subsequent depression. Participants with insomnia who were considered to be at high risk for depression with this questionnaire were characterized by moderate to high distress, pre-sleep arousal, and both reward sensitivity and insensitivity [35]. Replication of findings with this questionnaire or some similar symptom sampling approach could enable the development of treatment-relevant phenotypes that could guide the clinical application of insomnia treatment as a depression prevention strategy, although other strategies would also be informative. For example, only a few studies reported on the proportion of the sample with a history of MDD [26, 28], and it is thus unclear how well insomnia treatment may prevent new episodes among those with more chronic, highly recurrent histories, and for how long. Establishing whether insomnia treatment would be a more potent depression prevention strategy among individuals without a lifetime history of MDD as compared to those with a history of recurrent episodes would be a meaningful clinical contribution. Additionally, we did not review any studies aimed at assessing depression prevention in children, adolescents, and/or young adults with insomnia. Addressing insomnia during these developmental windows, particularly among those with a family history of depression, and investigating depression incidence over several years of follow-up would further establish a depression prevention effect.

Perhaps most critical to understanding the preventative effects of insomnia treatment on depression onset is knowledge of the mechanism of the effect. Mechanistic pathways were only explored in one study, although others generated testable hypotheses. Cheng et al. [30] noted in their original trial that while their study did not explicitly examine mechanisms of action, they observed that participants who completed dCBT-I reported less stress, cognitive intrusion, and hyperarousal symptoms specific to COVID-19, domains they acknowledge are specifically targeted by CBT-I. Indeed, in a subsequent analysis, the authors found that reductions in rumination mediated reductions in depression severity as well as prevention of incident depression in those assigned to dCBT-I [44]. Rumination may thus be an important factor for future research on the relationship of insomnia to subsequent depression. Irwin et al. [26] reported plans for secondary analyses investigating inflammation as a potential preventative pathway, citing research demonstrating the influence of inflammation on depression and reductions in the levels of inflammatory markers following insomnia treatment. As mentioned previously, reward sensitivity/insensitivity were markers of a subtype of insomnia considered to be particularly vulnerable to depression. Reward processing has been increasingly studied as a mechanistic link between insomnia and depression [45–49] and may thus be a promising target for assessing how the treatment of insomnia may prevent depression onset or worsening. Regulation of circadian rhythms, known to be disturbed in depression [50], may also be a worthwhile focus of exploration given the findings of Leerssen et al [28]. where only the combination treatment that included CBT-I and a chronotherapeutic intervention significantly prevented clinical worsening of depression. The stabilization of circadian rhythms is a meaningful target that should be examined in future studies. Finally, while not directly tested, improvement in sleep quality via remission from insomnia may itself be a mechanism through which depression onset is prevented. Because sleep disturbances are a frequent and significant symptom in most presentations of MDD, addressing insomnia symptoms could lessen overall depressive symptoms before they reach the clinical threshold. Indeed, there is some evidence that the sleep restriction component of CBT-I administered as a stand-alone intervention can reduce depression symptoms [51, 52] which may support this hypothesis given that sleep restriction alone has been associated with improved sleep quality [53], however further investigation is needed.

In conclusion, the limited evidence to date is promising that insomnia interventions may contribute to secondary prevention of MDD, however critical gaps in knowledge remain about its effects as a primary prevention tool as well as to the mechanisms of action. Although heterogeneity in baseline depression symptomatology and operationalization of outcomes may have influenced findings across studies, it is important to note that significant effects were observed in these studies among diverse samples (e.g. adults and older adults, pregnant women), across diverse modalities (clinician-administered CBT-I, dCBT-I), and with diverse depression measures. In many ways, these variations contribute to the robustness of this early preventative signal even though strong quantitative conclusions are yet to be drawn. Indeed, because the current evidence suggests that CBT-I and dCBT-I are promising interventions for secondary prevention of MDD, we encourage all future studies in this domain to ensure that this standard of care treatment (i.e. CBT-I) is eventually provided to all participants in studies evaluating insomnia either with or without mental health disorders. Additionally, although there are limitations to digital interventions, they are increasingly utilized to help meet the overwhelming need for sleep and mental health services globally. It is important that research aimed at assessing the preventative efficacy of insomnia treatment on depression utilize treatment modalities that will generalize to the current and future states of healthcare delivery. As research on the relationship between insomnia and depression continues, and as dCBT-I interventions continue to evolve and improve, it is likely these platforms will play a significant role in depression prevention efforts in the future. Research utilizing multiple forms of treatment delivery is encouraged to aid in treatment matching, investigate mechanisms, improve adherence to dCBT-I, and optimize outcomes for individuals at risk for depression.

Funding

Dr. Boland’s time was supported by Clinical Science Research and Development, Department of Veterans Affairs, IK2-CX001501, and by the National Institute of Mental Health, R21MH130642. Dr. Goldschmied’s time was supported by the National Institute of Mental Health, K23MH118580-04. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the U.S. Government.

Disclosure Statements

Dr. Gehrman has research funding from Merck, Inc. and is a consultant for Idorsia and Eight Sleep. No other authors have financial disclosures. No authors report nonfinancial conflicts of interest.