Cannabis and Sleep: A Self-Reinforcing Cycle That May Worsen Both

A new narrative review proposes that poor sleep drives cannabis use for self-medication, which in turn perpetuates sleep problems through tolerance and neurobiological disruption, but the bidirectional model remains conceptual and awaits the empirical testing its authors call for.

Why This Matters

Sleep disorders affect a substantial proportion of the U.S. population and carry significant health and economic costs. At the same time, cannabis is the most widely used federally illicit substance in the country, and public perception of it as a safe and effective sleep aid is growing rapidly. Up to one in four American adults may be using cannabis to help them sleep, yet the long-term consequences of this self-medication behavior remain poorly understood. A framework for organizing what is known, and identifying what is not, is both timely and clinically necessary.

Clinical Summary

Insomnia symptoms, short sleep duration, and evening chronotype during adolescence and young adulthood have been linked in longitudinal studies to earlier initiation and more frequent use of cannabis. Against this backdrop, Duong and colleagues published a narrative review in Current Sleep Medicine Reports (2025) proposing a bidirectional conceptual model: sleep problems drive cannabis self-medication, and cannabis use in turn maintains or worsens those sleep problems through predictable neurobiological mechanisms including tolerance development, suppression of slow-wave and REM sleep, and disruption of circadian signaling. The authors draw an explicit parallel to well-established alcohol-sleep self-medication models, embedding the cycle within a social determinants of health framework that accounts for sociodemographic disparities in both sleep health and cannabis access.

Prevalence data cited in the review suggest that up to 25% of U.S. adults use cannabis occasionally or regularly as a sleep aid, with figures rising to 79% in samples of frequent or hazardous users. The authors propose that as tolerance develops, users escalate dose or frequency to maintain perceived benefit, which deepens both sleep disruption and risk for cannabis use disorder. However, the review is a non-systematic narrative synthesis with no formal protocol, no PRISMA reporting, no quality appraisal of included studies, and no original data. The total number of studies reviewed and their individual rigor are not reported. The authors themselves state that the model requires prospective longitudinal research, experimental manipulation studies, and dismantling designs before any causal or clinical conclusions can be drawn.

Dr. Caplan’s Take

This review articulates something many of us see in practice: patients who started using cannabis for sleep and now cannot sleep without it, yet whose sleep quality has not meaningfully improved. The theoretical model the authors propose is intuitive and clinically resonant, drawing on solid parallels with the alcohol-sleep literature. But intuition and resonance are not evidence. Patients who ask whether cannabis is helping their sleep deserve an honest answer: we do not yet have the prospective, controlled data to confirm or quantify the cycle this paper describes, and individual responses to cannabis vary considerably by product, dose, and biology.

In practice, when I see patients using cannabis nightly for sleep, I treat it as a clinical conversation rather than an automatic recommendation to stop. I assess sleep hygiene, screen for underlying sleep disorders, and discuss cognitive behavioral therapy for insomnia as the first-line intervention with the strongest evidence base. If a patient is using cannabis and sleeping poorly, I frame the possibility that cannabis may be contributing to the problem, not just masking it, and I work with them on a monitored plan rather than issuing blanket prohibitions that erode trust.

Clinical Perspective

This review sits very early in the research arc. It does not confirm a causal cycle; it proposes one. The underlying observational studies linking sleep problems to subsequent cannabis use are suggestive but confounded by shared risk factors including mood disorders, chronic pain, and socioeconomic stress. The acute pharmacological evidence on cannabis and sleep architecture is real but drawn largely from short-duration experimental studies that do not capture the chronic use patterns most relevant to the self-medication hypothesis. Clinicians should not cite this model as established science when counseling patients, but they can reasonably use it as a framework for discussing why cannabis may not be delivering the sleep benefits patients expect over time.

From a pharmacological standpoint, THC’s effects on sleep are dose-dependent and biphasic, with low doses potentially reducing sleep onset latency and higher doses or chronic use associated with reduced slow-wave sleep and REM rebound upon cessation. CBD’s effects remain poorly characterized at clinically relevant doses. Clinicians should be aware that patients withdrawing from regular cannabis use commonly experience rebound insomnia lasting days to weeks, which reinforces the self-medication cycle. The single most actionable recommendation from this evidence base is to screen all patients presenting with sleep complaints for cannabis use, including frequency, timing, and product type, and to document it as a modifiable factor in sleep treatment planning.

Study at a Glance

Study Type

Narrative review with proposed conceptual model

Population

Lifespan, with emphasis on adolescents and emerging adults in the United States

Intervention

Non-synthetic cannabis products (THC/CBD) used for sleep self-medication

Comparator

Not applicable (no comparative analysis conducted)

Primary Outcomes

Proposed bidirectional model linking sleep problems and cannabis use; research agenda

Sample Size

Not applicable (no original data; total number of reviewed studies not reported)

Journal

Current Sleep Medicine Reports

Year

2025

DOI or PMID

DOI not provided in source material

Funding Source

Not reported

What Kind of Evidence Is This

This is a narrative review, which occupies a relatively low position in the evidence hierarchy compared to systematic reviews, meta-analyses, or original controlled studies. Narrative reviews synthesize existing literature without a predefined, reproducible search protocol or formal quality assessment, making them susceptible to selective inclusion and interpretive bias. The single most important inference constraint is that the bidirectional model proposed here is a hypothesis-generating framework, not an empirically validated causal claim.

How This Fits With the Broader Literature

The proposed cannabis-sleep cycle draws heavily on established models of alcohol and sleep self-medication, where bidirectional reinforcement between substance use and insomnia has stronger empirical support from prospective cohort studies and treatment dismantling designs. Within the cannabis-specific literature, the review is broadly consistent with findings from Babson and colleagues (2017), who documented associations between cannabis use and sleep disturbance, and with Winiger and colleagues (2021), whose twin studies suggested shared genetic liability between cannabis use and sleep problems. However, neither those studies nor any others cited in this review have confirmed the full feed-forward cycle the model proposes. The social determinants framing extends the conversation productively but also introduces variables that are largely untested as moderators in this specific context.

Common Misreadings

The most likely overinterpretation is treating this review as evidence that cannabis definitively worsens sleep, or that the self-reinforcing cycle described has been empirically demonstrated. It has not. The model is a theoretical synthesis of suggestive but heterogeneous observational and experimental findings. Individual studies cited within the review vary substantially in design, population, cannabis product examined, and methodological quality. Readers and clinicians who present this framework to patients as proven fact would be significantly exceeding what the evidence supports. Conversely, dismissing the model entirely because it is unproven would ignore the plausible biological mechanisms and consistent observational signals it organizes.

Bottom Line

This narrative review offers a coherent and clinically intuitive theoretical model proposing that sleep problems and cannabis self-medication reinforce each other over time. The model is plausible and useful for organizing future research, but it is not empirically validated. It should inform clinical questioning and treatment planning around cannabis and sleep, but it does not yet warrant changes to clinical recommendations or public health messaging beyond what existing evidence already supports.

References

1. Duong D, et al. A bidirectional model of cannabis use and sleep health across the lifespan. Current Sleep Medicine Reports. 2025.
2. Babson KA, Sottile J, Morabito D. Cannabis, cannabinoids, and sleep: a review of the literature. Current Psychiatry Reports. 2017;19(4):23. doi:10.1007/s11920-017-0775-9
3. Winiger EA, Hitchcock LN, Bryan AD, Bidwell LC. Cannabis use and sleep: expectations, outcomes, and the role of age. Addictive Behaviors. 2021;112:106642. doi:10.1016/j.addbeh.2020.106642