Cannabis and Sleep: A Proposed Vicious Cycle That Urgently Needs Testing

A 2025 narrative review outlines how poor sleep may drive cannabis self-medication, which then worsens sleep architecture and escalates use toward disorder, but the integrated bidirectional model remains entirely theoretical and has not been empirically validated as a complete pathway.

Why This Matters

Sleep problems are among the most common reasons adults cite for using cannabis, with roughly one in four U.S. adults reporting cannabis as a sleep aid. As legalization expands and consumer perceptions of safety grow, clinicians increasingly encounter patients who have already begun self-medicating with cannabis for insomnia. At the same time, evidence is accumulating that chronic cannabis use may degrade the very sleep architecture users are trying to improve. A clear conceptual framework connecting these observations is overdue, even if the framework itself still requires rigorous empirical testing.

Clinical Summary

Two largely separate research literatures have developed around cannabis use and sleep disturbance, rarely brought into direct conversation with each other. A 2025 narrative review by Watts and colleagues, published in Cannabis and Cannabinoid Research, attempts to bridge this gap by proposing a bidirectional, feed-forward model. The authors synthesize evidence from PubMed and PsychINFO searches to argue that sleep problems, particularly insomnia symptoms, short sleep duration, and evening chronotype, prospectively predict cannabis initiation and escalation, especially during adolescence. The mechanistic rationale centers on the endocannabinoid system’s known role in sleep-wake regulation: exogenous cannabinoids acutely modulate sleep onset and architecture, but chronic exposure likely disrupts endocannabinoid signaling, leading to tolerance and rebound sleep disturbance.

The review’s central claim is that this creates a self-reinforcing cycle: poor sleep motivates cannabis use, which temporarily improves subjective sleep perception but objectively maintains or worsens sleep quality, prompting dose escalation and increasing the risk of cannabis use disorder. The authors additionally identify social determinants of health as understudied modulators that may explain disparities in both sleep problems and cannabis risk. However, no new data are presented, the literature search lacked systematic methodology or risk-of-bias appraisal, and the full feed-forward cycle has never been empirically tested as an integrated model. The authors themselves acknowledge that longitudinal and experimental studies are needed before the model can inform clinical guidance.

Dr. Caplan’s Take

This review does something genuinely useful by naming a pattern that many of us see in practice: patients who start using cannabis for sleep, find it works for a while, and then find themselves using more and sleeping worse. The biological reasoning is sound, and the developmental framing from adolescence through adulthood adds real clinical texture. But the gap between a plausible conceptual model and validated clinical evidence is substantial. When a patient tells me cannabis is the only thing that helps them sleep, I need to be honest that the science supporting or refuting that claim is still immature and that this particular model, however intuitive, has not been directly tested.

In practice, I take sleep complaints in cannabis-using patients seriously as both a potential driver and a consequence of their use. I screen for sleep problems early, discuss the tolerance and rebound dynamics that this review highlights, and avoid blanket statements in either direction. For patients interested in tapering cannabis, I work on sleep hygiene and evidence-based insomnia interventions first, because addressing the sleep problem reduces one of the strongest motivations to continue or escalate use. That sequencing is practical, defensible, and does not require the full model to be validated to be clinically useful.

Clinical Perspective

This review sits at the earliest stage of the research arc: model proposal and agenda-setting. It synthesizes real findings from observational and neurobiological studies, but the integration into a unified feed-forward cycle is the authors’ theoretical contribution, not an empirically demonstrated pathway. Individual links in the chain have supporting evidence of varying strength. Prospective studies do suggest that adolescent sleep problems predict later cannabis use, and laboratory data indicate chronic cannabis use alters sleep architecture. However, these findings come from heterogeneous study designs with limited control for confounders, and the temporal dynamics and dose-response relationships within the proposed cycle remain entirely uncharacterized. Clinicians should not present this model to patients as established science.

From a pharmacological standpoint, the review’s discussion of endocannabinoid system disruption is relevant but incomplete. THC and CBD have distinct and sometimes opposing effects on sleep, and most consumer products contain variable ratios of both along with other cannabinoids whose sleep effects are poorly understood. Drug-interaction considerations are essentially unaddressed, which matters for patients taking sedative-hypnotics, antidepressants, or other CNS-active medications alongside cannabis. The most actionable recommendation for clinicians right now is to routinely screen for sleep disturbances in patients who report cannabis use and to offer cognitive behavioral therapy for insomnia as a first-line intervention, which has a robust evidence base independent of cannabis status.

Study at a Glance

Study Type

Narrative review with conceptual model

Population

U.S. lifespan populations; non-clinical samples emphasized, with attention to adolescents and young adults

Intervention

Not applicable (no intervention tested)

Comparator

Not applicable

Primary Outcomes

Proposed bidirectional relationship between sleep disturbance and cannabis use escalation

Sample Size

Not applicable (no original data collected)

Journal

Cannabis and Cannabinoid Research

Year

2025

DOI or PMID

Published January 3, 2025 (specific DOI not provided in source)

Funding Source

Not specified in available data

What Kind of Evidence Is This

This is a narrative review, which occupies a position near the base of the evidence hierarchy for clinical decision-making. Unlike systematic reviews, narrative reviews do not employ pre-specified search strategies, formal inclusion and exclusion criteria, or risk-of-bias assessment of cited studies. The single most important inference constraint this imposes is that the completeness and representativeness of the literature surveyed cannot be independently verified, making the review susceptible to selection bias and potentially overstating the coherence of its proposed model.

How This Fits With the Broader Literature

The individual components of this model are consistent with prior work. Longitudinal studies such as those by Babson and colleagues have documented prospective associations between sleep problems and subsequent cannabis use in adolescent samples, and polysomnographic research has shown that chronic cannabis users exhibit reduced slow-wave sleep and altered REM architecture. The self-medication hypothesis for substance use more broadly has a long history in the addiction literature. What this review adds is an explicit integration of these threads into a single developmental framework, which is a conceptual rather than empirical advance.

Notably, the review does not engage deeply with the growing literature on CBD-specific effects on sleep, which in some studies appear distinct from and potentially opposite to THC effects. This omission limits the model’s applicability to the diverse cannabis product landscape patients actually encounter.

Common Misreadings

The most likely overinterpretation is treating the proposed feed-forward cycle as an established causal pathway. The review presents a model, not a tested mechanism. While individual associations between sleep problems and cannabis use have empirical support, no study has tracked the complete cycle from sleep disturbance through cannabis initiation, tolerance development, sleep degradation, dose escalation, and cannabis use disorder in a single longitudinal or experimental design. Citing this review as evidence that cannabis worsens sleep in a clinically meaningful, dose-dependent manner for all users would exceed what the evidence supports, particularly given the heterogeneity of cannabis products and individual neurobiological responses.

Bottom Line

This narrative review offers a clinically intuitive and biologically plausible model connecting sleep disturbance and cannabis use in a self-reinforcing cycle, but it remains a hypothesis, not a demonstrated finding. No new data are presented and the integrated pathway has not been empirically tested. For now, the most defensible clinical action is to screen cannabis users for sleep problems and to offer evidence-based insomnia treatments as a parallel or alternative intervention.

References

1. Watts AK, et al. Cannabis and sleep: a proposed bidirectional self-medication model. Cannabis and Cannabinoid Research. Published online January 3, 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. Gates PJ, Albertella L, Copeland J. The effects of cannabinoid administration on sleep: a systematic review of human studies. Sleep Medicine Reviews. 2014;18(6):477-487. doi:10.1016/j.smrv.2014.02.005
4. Kesner AJ, Lovinger DM. Cannabinoids, endocannabinoids and sleep. Frontiers in Molecular Neuroscience. 2020;13:125. doi:10.3389/fnmol.2020.00125