CBD for sleep and stress: what Shannon 2019, Bergamaschi 2011, and the mechanism evidence actually show

The primary evidence: Shannon 2019 and Bergamaschi 2011

Shannon 2019: the best available human data

Shannon 2019 is a retrospective case series of 72 adults presenting to a psychiatric clinic with primary complaints of anxiety and/or sleep disruption. Patients received CBD 25mg/day (anxiety cases) or 25–175mg/day (sleep cases, titrated by clinical assessment) added to their existing medication regimens.

Over three months, sleep scores improved in 66.7% of patients within the first month. Anxiety scores improved in 79.2% of patients. The anxiety signal is stronger — which is the key clinical finding. It supports the mechanistic hypothesis that CBD's sleep benefit is substantially mediated through anxiety reduction, not through a direct sleep-promoting mechanism.

Important limitations to hold clearly: this is a case series, not a placebo-controlled randomized trial. Patients were also taking other psychiatric medications. The lack of a control arm means we cannot rule out placebo effect, natural disease course, or concomitant medication effects. Shannon 2019 is the best available human evidence — but "best available" and "high-quality RCT" are not the same thing.

Sleep scores also fluctuated across the observation period. Month one improvement was not uniformly sustained at month two or three. This volatility is consistent with a non-sedative mechanism: CBD doesn't produce the sustained blunting that GABA-A-acting substances do, but it also doesn't produce rebound insomnia on cessation.

Bergamaschi 2011: the strongest anxiety evidence

Bergamaschi 2011 is a randomized, placebo-controlled study in treatment-naive social anxiety disorder patients. Participants received a single 600mg dose of CBD or placebo before a simulated public-speaking task. CBD significantly reduced subjective anxiety, cognitive impairment, and discomfort during speech — compared to placebo. This is the highest-quality evidence we have for CBD as an anxiolytic.

Two caveats matter here. First, 600mg is a high single dose — far above the 25mg daily dose used in Shannon 2019 and well above what most people take for ongoing use. The dose-response context is important. Second, this was an acute social-anxiety model (public speaking) — not a model of chronic generalized anxiety disorder or the daily, low-grade anxiety that disrupts sleep. Extrapolating from acute social anxiety to chronic GAD requires caution.

Linares 2019: dose-response and sleep architecture

Linares 2019 tested 150mg, 300mg, and 600mg CBD against placebo in healthy volunteers, measuring anxiety and sleep architecture. Two findings matter here:

• Inverted-U dose-response for anxiety: 300mg outperformed both 150mg and 600mg on anxiety reduction. The implication for ongoing use: higher dose is not better, and may be worse. This is consistent with what many users report — that doses above their individual threshold cause alertness rather than calm.
• No sleep architecture disruption: At 300mg, CBD did not significantly alter sleep architecture in healthy volunteers — no REM suppression, no distortion of sleep stages. This matters for people concerned about quality of sleep, not just duration. Unlike alcohol and benzodiazepines, CBD does not appear to suppress restorative sleep stages at typical doses.

Chagas 2014: the RBD signal

Chagas 2014 documented CBD reducing REM sleep behavior disorder (RBD) episodes in Parkinson's disease patients — a specific condition where the normal motor inhibition during REM sleep fails, causing patients to physically act out dreams. This is the most specific and mechanistically coherent sleep signal we have for CBD: a discrete disorder with a defined mechanism, not a general "CBD helps sleep" claim.

The n=4 case series size limits generalizability, and the Parkinson's context means this should not be extrapolated to general insomnia. But the signal is real and biologically coherent.

Sleep continuity vs initiation: where the evidence applies

This distinction is clinically important and often collapsed in wellness marketing.

Sleep initiation is the ability to fall asleep. This is impaired in classic sleep-onset insomnia. CBD has weak and inconsistent evidence here. It is not a hypnotic, does not have sedative pharmacology, and has no established effect on sleep-onset latency in controlled studies. If falling asleep is your primary problem and there is no clear anxiety or pain driver, CBD is unlikely to help more than a placebo would.

Sleep continuity is the ability to stay asleep through the night. The 3am wake-up pattern — awakening after 4–5 hours of sleep and lying awake for 1–2 hours — is the most common sleep complaint in the 35–65 cohort and is heavily driven by anxiety, cortisol dysregulation, and pain. This is where CBD's mechanistic profile is most relevant. Reducing nocturnal hyperarousal (via 5-HT1A agonism and amygdala modulation) and pain signaling can allow return to sleep after waking — not by inducing sedation, but by quieting the arousal systems that prevent it.

The Shannon 2019 patient reports are most consistent with the continuity pattern. Patients described better overall sleep quality and less frequent waking, rather than faster sleep onset. This should calibrate expectations: if your problem is taking an hour to fall asleep in a quiet room with no pain or anxiety, CBD's evidence is not directly applicable.

The anxiety-sleep connection

Anxiety and sleep disruption are not independent problems for most people in the target population. Nocturnal rumination — the cognitive hyperarousal of lying awake reviewing today's failures and tomorrow's threats — is a primary mechanism of sleep-maintenance insomnia. Cortisol elevation (common in chronic stress states and in the age-related HPA axis dysregulation seen in the 45–65 cohort) causes arousal, suppresses slow-wave sleep, and produces the characteristic early-morning awakening pattern.

This is why Shannon 2019's finding — that the anxiety signal is stronger than the sleep signal in the same population — is more than a footnote. It suggests that CBD's most reliable path to improved sleep runs through its anxiolytic mechanism. Treating anxiety and hyperarousal is the target; sleep improvement follows as a downstream effect. This has practical implications: if you take CBD and notice anxiety reduction but no immediate sleep change, the sleep benefit may follow over subsequent weeks as the anxiolytic effect compounds.

Crippa 2004 showed CBD reduced anxiety in an fMRI paradigm, with corresponding decreases in amygdala activity and altered parahippocampal gyrus responses. The amygdala is central to threat detection and fear consolidation — overactivation at night drives hyperarousal and prevents sleep. Modulating amygdala responsiveness is a biologically plausible mechanism for the nocturnal anxiety-hyperarousal-waking pattern.

Mechanism: what CBD actually does (and does not do)

Cortisol and the HPA axis: the stress-sleep link

Zuardi 1993 (PMID 8481739) showed that CBD blunted the cortisol response to stress in healthy volunteers — an early, pharmacological signal that CBD can modulate hypothalamic-pituitary-adrenal (HPA) axis activity. This was healthy-volunteer data with single-dose administration, not a clinical study of cortisol dysregulation in anxious or sleep-disrupted patients. The finding is suggestive, not definitive.

The relevance to the 35–65 cohort is real: age-related HPA axis dysregulation produces elevated evening cortisol, blunted cortisol awakening response, and flattened diurnal rhythms. This pattern correlates with sleep fragmentation, early-morning awakening, and the subjective experience of "wired but tired." If CBD modulates HPA reactivity (as Zuardi 1993 suggests), there is a plausible mechanism connecting CBD use to improved sleep continuity in this cohort — but the direct evidence for this specific mechanism in aging adults with sleep disruption does not yet exist.

Stress, sleep, and the inflammaging connection

Chronic psychological stress is not merely uncomfortable — it is pro-inflammatory at a molecular level. The pathway is well-characterised: sustained cortisol elevation (via HPA axis activation) paradoxically promotes NF-κB activation in immune cells, increasing transcription of IL-6, TNF-α, and other pro-inflammatory cytokines. This is the stress-to-inflammation axis that connects the sleep-stress pillar to the broader inflammaging framework.

Sleep deprivation independently elevates CRP and IL-6 (Irwin 2016 — contextual reference without PMID, as the data are from multiple Irwin-group publications over multiple years). Even a single night of partial sleep loss measurably elevates inflammatory markers the following day. Chronic sleep disruption, accumulated over months and years, is an independent driver of inflammaging — distinct from (but compounding with) other drivers such as senescent cell SASP and gut dysbiosis.

This creates a biologically plausible case for CBD's indirect anti-inflammatory benefit via the sleep and anxiety pathways: if CBD reduces nocturnal anxiety and improves sleep continuity, it removes a driver of elevated IL-6 and CRP. Whether this secondary effect is clinically meaningful has not been tested. It is mechanistic inference, not direct evidence. We present it as such.

For full detail on the inflammaging framework, NF-κB signaling, and CBD's direct anti-inflammatory mechanisms, see the inflammaging page and the CBD and inflammation page.

What CBD is not: honest limits for E-E-A-T

Dosing: context matters enormously

Two studies dominate the dose discussion — and they used very different doses for very different purposes. Shannon 2019 used 25mg/day as the starting dose for ongoing anxiety and sleep management. Bergamaschi 2011 used 600mg as a single acute dose for an experimental social anxiety model. These are not interchangeable data points.

For ongoing use in the sleep-stress context, 25mg/day is the evidence-based starting point. A more conservative entry for the drug-interaction-aware 45+ cohort is 15mg, titrating up over 4–6 weeks. Take with food — Taylor 2018 showed food increases oral CBD bioavailability 4–5-fold. Evening timing, 1–2 hours before bed, is the common patient-reported approach; CBD's Tmax (time to peak plasma concentration) after oral dosing with food is approximately 3–4 hours.

Remember the Linares 2019 inverted-U: 300mg outperformed 600mg. The equivalent principle applies at lower doses — if you find CBD alerting at 50mg, try 25mg. More is not better, and may be worse.

For the full titration protocol — including how to track response, when to adjust, and how to interpret variable effects — see the CBD titration protocol.