A 2026 Phase II randomized controlled trial from Thailand found that a traditional cannabis-based multi-herbal medicine performed as well as lorazepam 0.5 mg for chronic insomnia over four weeks, with a comparable safety profile. The finding is encouraging but carries important limits: the intervention is a multi-herb formulation, the sample is small, and the study population is specific to Thailand. This is early, directionally interesting evidence, not a prescription recommendation.
Cannabis-Based Medicine Matches Lorazepam for Chronic Insomnia: What a Phase II Trial Reveals
Insomnia is one of the most common reasons patients ask about cannabis, and one of the conditions where the evidence is most actively developing. A newly published clinical trial offers the strongest head-to-head comparison yet between a cannabis-based preparation and a benzodiazepine, and the results challenge the assumption that a pharmaceutical sleeping pill is the default first option.
This is a peer-reviewed Phase II RCT with a clinically meaningful active comparator. Sleep is one of the most prevalent concerns in the CED patient population, and the benzodiazepine comparison addresses a real-world decision clinicians face regularly. The multi-herb formulation and Thai patient population introduce meaningful generalizability limits.
- – What the Thai Phase II insomnia trial actually tested, and what it found
- – Why comparing cannabis to a benzodiazepine is clinically significant
- – The specific limits of this study that every reader should understand
- – How this fits into the broader clinical picture of cannabis and sleep
- – What the evidence does and does not support when patients ask about cannabis for sleep
- A Phase II RCT published in February 2026 found that a traditional Thai cannabis-based herbal medicine performed non-inferiorly to lorazepam 0.5 mg for chronic insomnia over four weeks.
- Both groups saw meaningful improvements in sleep quality scores, quality of life, and stress, with comparable safety outcomes.
- The intervention is a multi-herb formulation, not isolated cannabis, and the study is small and conducted in Thailand, limiting direct application to US clinical practice.
- This is promising directional evidence that warrants replication in larger, better-characterized populations before clinical guidance can shift.
- Multi-herb formulation: The active treatment is not isolated cannabis. The Anti-Pom-Leung Fever medicine contains several herbal ingredients alongside cannabis. Effects observed cannot be attributed to cannabis alone.
- Small sample: 82 participants completed the trial (41 per arm). This is underpowered for subgroup analyses or detection of rare adverse events.
- Single-center, Thai population: Results were obtained in a hospital setting in Thailand. Dietary patterns, baseline medication use, and cannabis-naive status of Thai patients may differ substantially from US populations.
- Short duration: Four weeks does not address long-term efficacy, tolerance development, or safety over time, which are critical concerns with any sleep medication including lorazepam.
- Dose of comparator: Lorazepam 0.5 mg is a low dose. It is clinically appropriate for a trial, but the comparison group does not represent the full range of benzodiazepine prescribing in practice.
| Study Type | Phase II Randomized, Double-Blind, Active-Controlled Non-Inferiority Trial |
| Sample Size | 100 enrolled; 82 completed (41 per group) |
| Population | Adults with chronic insomnia; Prapokklao Hospital, Thailand |
| Intervention | Anti-Pom-Leung Fever medicine (traditional Thai cannabis-based multi-herbal formulation) vs. lorazepam 0.5 mg; 4 weeks |
| Primary Outcome | Pittsburgh Sleep Quality Index (PSQI) at Week 4; non-inferiority margin of 2.1 points |
| Key Results | PSQI 3.44 (herbal) vs. 4.78 (lorazepam); mean difference -1.34 (95% CI: -2.99 to 0.31). Non-inferiority demonstrated. QoL and stress improved in both groups. No clinically significant adverse effects in either arm. |
| Funding | Not reported in available source material |
| Conflicts of Interest | Not reported in available source material |
| Publication | Journal of Cannabis Research, Vol. 8, Article 46 (February 24, 2026) — Open Access |
| Key Limitations | Small sample; single center; Thai population; multi-herb formulation (effects not attributable to cannabis alone); 4-week duration; low benzodiazepine comparator dose |
Why This Matters
Insomnia affects roughly one in three adults at some point in their lives. In clinical practice, it is among the most common reasons patients seek out cannabis, and among the conditions where the demand for non-pharmaceutical options is strongest. Yet clinical guidance remains cautious precisely because rigorous head-to-head trials have been scarce.
Benzodiazepines like lorazepam are effective for short-term insomnia but carry well-documented risks: tolerance develops quickly, physical dependence is common, and discontinuation can trigger rebound insomnia that is sometimes worse than the original complaint. The existence of a trial that formally tests a cannabis-based alternative against this class of medication, using a validated sleep outcome measure, is meaningful even when the study is small.
What the Study Found
Researchers at Prapokklao Hospital in Thailand enrolled 100 adults with confirmed chronic insomnia and randomized them to receive either the traditional Thai cannabis-based formulation known as Anti-Pom-Leung Fever medicine (delivered as herbal capsules twice daily) or lorazepam 0.5 mg at bedtime for four weeks. The study was double-blind and designed as a non-inferiority trial, meaning the pre-specified question was whether the herbal preparation performed within an acceptable margin of the drug, not whether it outperformed it. Eighty-two participants completed the study, 41 in each arm.
By week four, Pittsburgh Sleep Quality Index scores had improved dramatically in both groups. The herbal group achieved a mean PSQI of 3.44, compared to 4.78 in the lorazepam group, a difference of -1.34 points in favor of the herbal arm (95% CI: -2.99 to 0.31). Since the upper bound of the confidence interval did not exceed the predefined non-inferiority margin of 2.1, the trial confirmed non-inferiority. Both groups also showed improvements in quality of life and perceived stress. Safety was comparable, with no clinically significant abnormalities detected in liver, kidney, or cardiac function in either arm. The absolute improvement in PSQI was approximately 9 points in the herbal group and 8.3 points in the lorazepam group, both large changes from a clinical standpoint.
What This Paper Does Not Show
This trial does not establish that cannabis as a substance, in any product form available to US patients, is equivalent to lorazepam for insomnia. The intervention is a traditional multi-herb preparation with a complex botanical composition that includes cannabis alongside other plant medicines. The proportion of cannabinoids, the THC-to-CBD ratio, and the contribution of individual plant components to the outcome are not reported. Isolating cannabis as the effective ingredient is not possible from this data.
The study also does not address what happens after four weeks. Lorazepam’s most problematic effects, tolerance and physical dependence, typically emerge with extended use. The study design cannot compare long-term trajectories. And a single-center trial of 82 patients in Thailand, however well-conducted, cannot override the need for replication across diverse populations and settings before any clinical practice change is warranted.
How This Fits Into the Broader Picture
The Thai trial is not the first to suggest that cannabis-based preparations may be useful for sleep. Several observational studies, registry analyses, and small randomized trials have examined the question, and the pattern that emerges is consistent: cannabis and cannabinoids tend to reduce the time it takes to fall asleep and are associated with improvements in self-reported sleep quality, particularly in people whose sleep disruption is tied to pain, anxiety, or other underlying conditions. A 2024 systematic review found moderate evidence for cannabinoids improving insomnia symptoms, though the quality of included trials varied significantly.
What this trial adds is the active-comparator design. Most cannabis and sleep studies have compared cannabis to placebo. Comparing to a real-world clinical alternative, specifically a short-acting benzodiazepine, is a methodological step forward. The trial raises a question that matters in the exam room: if cannabis-based treatment produces comparable improvements in sleep quality with a different safety profile, is there a population of patients for whom it represents a reasonable first-line or alternative option? That question cannot be answered yet, but it is now legitimate to ask it in a research context.
In Massachusetts, as in many US states, patients regularly bring their cannabis use to appointments and ask about its role in managing sleep. Clinicians need a way to discuss the evidence honestly, which means acknowledging both what the data suggest and where its limits lie.
“Sleep is one of the most urgent conversations in my clinic. Patients come in after years of benzodiazepine prescriptions they never wanted to continue, asking whether there is something else. What I find in this trial is not a simple answer, but something nearly as valuable: a properly designed study that actually asks the head-to-head question. The finding that a cannabis-based preparation performs comparably to lorazepam, even in a small Thai sample, gives clinicians a scientific foothold to take this question seriously rather than dismissing it. That matters. The problem is not that the cannabis group performed well; it is that we still don’t know which patients benefit most, what product profile matters, or how the outcomes hold past the four-week mark. That’s not a reason to ignore the trial. It’s a reason to fund the next one.”
“In my practice in Boston, the insomnia patients who do best with cannabis tend to be those whose sleep disruption connects to something the cannabis is addressing: pain that keeps them awake, anxiety that prevents them from settling, or a hyperaroused nervous system that hasn’t quieted down. Pure maintenance insomnia with no underlying driver responds differently. This study doesn’t differentiate by insomnia subtype, which is a real gap. Until that granularity exists in the evidence, matching patients to treatment still requires clinical judgment. The trial is a signal worth following. It is not yet a guide for practice.”
What a Careful Reader Should Take Away
A well-designed Phase II trial found that a traditional Thai cannabis-based formulation performed non-inferiorly to lorazepam 0.5 mg for chronic insomnia over four weeks, with a similar safety profile. Both groups improved substantially. The upper bound of the confidence interval for the difference was 0.31, which fell within the predefined margin, confirming the non-inferiority claim. This is real, peer-reviewed evidence from a randomized controlled trial.
But a careful reader also holds the limits clearly: 82 patients, one hospital, four weeks, a multi-herb compound with no cannabinoid profile reported, and a population whose background cannabis familiarity and metabolism may differ from US patients. The study is a beginning, not a conclusion. It warrants replication. It does not warrant changing prescribing habits or self-treating with cannabis as a benzodiazepine substitute based on this trial alone.
If you are currently using a benzodiazepine for sleep or considering cannabis for insomnia, the right path is a conversation with a clinician who knows both the evidence and your history. That is exactly the kind of conversation the CED Clinic is built for.
Read This Paper Through Eight Different Lenses
A single study can mean different things depending on who is reading it. This card separates the patient takeaway, clinical meaning, skepticism, study critique, prior research context, practical implications, future directions, and likely public misreadings.
How to use this: Choose a lens above to see how the same paper reads from a different evidence, clinical, or practical angle.
Patient Takeaway
If you have chronic insomnia and you are wondering whether cannabis could be an option instead of a sleeping pill, this trial offers something genuinely useful: it is now documented in a peer-reviewed journal that a cannabis-based preparation performed comparably to a low dose of lorazepam in a formal clinical trial. That matters. It means the question you are asking is legitimate and worth raising with your clinician, not a fringe idea.
What it does not mean is that you should swap your medication for a cannabis product based on this single study. The preparation tested in Thailand contains multiple herbs, not just cannabis, and the study was conducted in 82 people over four weeks. We don’t know if the same results would hold in a US population, or with the cannabis products available in a Massachusetts dispensary. Sleep is also highly individual. What helps one person’s sleep disruption can worsen another’s. The right takeaway is not “cannabis replaces my sleeping pill.” The right takeaway is “this study suggests that cannabis-based medicine is worth taking seriously for sleep, and I should talk to someone who understands both the evidence and my specific sleep problem.”
Clinician’s POV
For a clinician managing patients with chronic insomnia, particularly those who express concern about benzodiazepine dependence or who are already using cannabis self-directed for sleep, this trial provides useful framing. It is now reasonable to discuss cannabis-based approaches in clinical conversation without dismissing them as unproven in a controlled setting. The non-inferiority finding against lorazepam 0.5 mg in 82 patients is a starting point, not a standard of care recommendation.
The more important clinical nuance is patient selection. The trial did not report insomnia subtypes, concurrent diagnoses, or cannabinoid profiles of the preparation. In practice, cannabis may be more promising for patients whose insomnia is driven by pain, anxiety, or hyperarousal than for those with primary sleep maintenance insomnia with no comorbidity. Caution is appropriate around any patient with a history of cannabis use disorder, psychosis, or significant respiratory compromise. The 4-week duration also means this study cannot inform conversations about long-term management, which is where benzodiazepine risk accumulates. Shared decision-making, rather than protocol-driven prescribing, is the appropriate frame here.
A Skeptical Read
Several things warrant scrutiny before taking this trial at face value. First, the intervention is not cannabis; it is a multi-herb formulation with an unknown cannabinoid profile. The trial’s results belong to the Anti-Pom-Leung Fever medicine as a composite botanical product. Attributing the outcome to cannabis is an inference, not a finding. Second, the non-inferiority margin of 2.1 PSQI points was defined by the authors. Non-inferiority margins are always pre-specified choices, and the clinical meaningfulness of that threshold in this population is not established by independent standards.
Third, the baseline PSQI scores and the dramatic improvement in both arms deserve attention. The lorazepam group started with similar baseline scores and ended at 4.78, which is near the clinical threshold for good sleep (PSQI under 5). This raises a question about whether the improvements reflect the treatment or reflect being enrolled in a structured trial, receiving follow-up care, and having a sleep problem formally acknowledged. A placebo arm would have clarified this. The absence of placebo means neither group’s improvement can be cleanly attributed to the pharmacological intervention.
Study Critic
The trial’s headline finding, non-inferiority, is methodologically appropriate for the research question. But the statistical underpinnings carry constraints worth naming clearly. With 41 completers per arm, the confidence interval for the PSQI difference runs from -2.99 to 0.31. The range spans nearly 3.3 PSQI points. That is a wide interval for a scale with a total range of 21 points and a clinical threshold at 5. The upper bound of 0.31 just clears the non-inferiority margin of 2.1, which is the correct result for the predefined test but leaves open a scenario where the true difference could be near the margin or slightly above it in a larger trial.
The study also reports no statistically significant time-by-treatment interaction, meaning both groups improved at similar rates over time. This supports the non-inferiority claim but does not reveal whether the trajectories diverged at any specific week. Single-timepoint evaluation at week 4 is appropriate for the primary analysis, but additional time points would have revealed whether improvements were stable, accelerating, or beginning to plateau. A longer follow-up would have captured the post-treatment rebound period that is particularly relevant for benzodiazepine comparisons.
Compared to Past Research
Prior research on cannabis and sleep has been conducted almost entirely against placebo. The existing body of work, including several randomized crossover trials and observational registry studies, generally shows that cannabinoids, particularly CBD and in some cases THC or CBN, reduce sleep latency and improve self-reported sleep quality compared to placebo in adults with insomnia or insomnia secondary to other conditions. A 2024 systematic review and meta-analysis in Sleep Medicine Reviews, which included over a dozen RCTs, found moderate-quality evidence for cannabinoids improving sleep outcomes, with the strongest effects seen in populations with comorbid pain or anxiety.
What this Thai trial adds is the active-comparator design, which has not been a feature of prior cannabis and sleep trials. Head-to-head comparisons against drugs with known clinical benchmarks are essential for situating the magnitude of cannabis-based effects within a clinical framework. However, because the intervention is a multi-herb preparation rather than a characterized cannabinoid product, this trial cannot be directly concatenated with prior cannabinoid studies. It belongs to a distinct but adjacent category of evidence. Comparison with prior research should be drawn carefully, with this distinction preserved.
Practical Considerations
Even if this trial’s findings are real and replicate in US populations, practical translation is not straightforward. The intervention is a Thai traditional medicine with a specific botanical composition that is not commercially available in the United States in its current form. US cannabis products vary enormously in cannabinoid content, terpene profile, and formulation. A dispensary product labeled for sleep is not the same compound that was tested in Thailand, and no direct equivalency can be drawn.
Dosing remains a practical challenge. The herbal capsule regimen (twice daily) differs from how most US patients use cannabis for sleep, which is typically a single evening dose. Timing, route of administration, and the presence or absence of tolerance all influence outcomes in clinical practice. Cost and insurance coverage are also relevant: lorazepam is inexpensive and commonly covered; cannabis remains federally unscheduled and out-of-pocket in most US states, including Massachusetts. For patients weighing options, these real-world factors shape what “non-inferior in a trial” actually translates to in daily life.
Future Directions (Expected)
The most important next step is replication in a larger, more diverse sample with a well-characterized cannabinoid intervention rather than a composite herbal preparation. A trial enrolling 200 to 400 patients, using a standardized CBD or THC:CBD formulation with a documented dosing protocol and a longer follow-up period, would add substantially more to the field than additional small trials in similar populations.
The field also needs trials that stratify by insomnia subtype. Sleep onset insomnia, sleep maintenance insomnia, and early morning awakening have distinct mechanisms and may respond differently to cannabis. Trials that enroll undifferentiated insomnia samples obscure subgroup effects. Biomarker-enriched designs, perhaps using inflammatory markers given prior evidence that baseline inflammation moderates cannabis response, could identify who is most likely to benefit. Withdrawal and rebound insomnia periods following cannabis discontinuation also deserve formal study, since this is a known concern with THC-heavy products used for sleep that has not been adequately characterized in RCT settings.
Misreadings & Bad-Faith Takes
Distortion: “Science proves cannabis is better than Ativan for sleep.” This misreads the trial in two ways. First, the study found non-inferiority, meaning comparable performance, not superiority. Second, the intervention was a multi-herb formulation, not cannabis alone. Attributing the finding to “cannabis” is inaccurate.
Distortion: “This study is fake news because cannabis causes sleep problems.” This ignores the actual data. Both groups improved substantially. The trial was double-blind, active-controlled, and published in a peer-reviewed journal. Dismissing the finding without engaging the methodology is not a scientific critique.
Distortion: “Thai herbal medicine beats Big Pharma’s sleeping pills.” This sensationalizes a small, four-week, single-center non-inferiority trial into a culturally charged narrative it does not support. The study is promising and deserves replication, not celebration as a decisive victory for traditional medicine over pharmaceutical treatment.
Correct framing: A carefully designed clinical trial found that a cannabis-based traditional preparation performed comparably to a low dose of lorazepam over four weeks in 82 Thai patients with chronic insomnia. The result is directionally encouraging and methodologically legitimate. It is also small, limited to one population and formulation, and insufficient on its own to guide clinical practice.
? Join the Conversation
Have a question about how this applies to your situation? Ask Dr. Caplan →
Want to discuss this topic with other patients and caregivers? Join the forum discussion →
Have thoughts on this? Share it:
Kummalue T, et al. Phase II randomized controlled trial comparing traditional Thai cannabis-based medicine with lorazepam for insomnia treatment. Journal of Cannabis Research. 2026;8(1):46. doi: 10.1186/s42238-026-00415-x — Open Access. Published February 24, 2026. Authors affiliated with Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand. Note: Full author list should be verified at the DOI link before final publication of this post.
Frequently Asked Questions
What exactly was the cannabis-based medicine tested in this trial?
The intervention was a traditional Thai herbal formulation called Anti-Pom-Leung Fever medicine. It is a multi-herb preparation that includes cannabis among several botanical ingredients and is legally recognized in Thailand for therapeutic and research use. The study does not report the specific cannabinoid content, THC-to-CBD ratio, or the precise contribution of each herbal component. This means the results cannot be attributed to cannabis alone and cannot be directly mapped to cannabis products available in US dispensaries.
Does this study mean cannabis is as effective as lorazepam for insomnia?
The study found that the cannabis-based herbal formulation was non-inferior to lorazepam 0.5 mg over four weeks in 82 Thai patients, using the Pittsburgh Sleep Quality Index as the primary measure. Non-inferiority means the treatment performed within a predefined acceptable margin of the comparator, not that it was proven equivalent or superior. This is a promising finding, but it comes from one small trial in a specific population with a compound not commercially available in the United States. It does not support switching from a prescribed benzodiazepine to cannabis without medical guidance.
What is the Pittsburgh Sleep Quality Index and what do the scores mean?
The Pittsburgh Sleep Quality Index, or PSQI, is a validated self-report questionnaire that measures sleep quality over the past month across seven components: subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleep medication, and daytime dysfunction. Scores range from 0 to 21, and a score above 5 is generally considered to indicate poor sleep quality. In this trial, both groups started with scores indicating poor sleep and ended at or near the threshold for good sleep, suggesting substantial improvement in both arms.
Why is it significant that this trial compared cannabis to a benzodiazepine rather than a placebo?
Most prior cannabis and sleep studies compared cannabis to a placebo. That design tells you whether cannabis beats doing nothing. An active-comparator design asks a more clinically meaningful question: how does cannabis perform against something already in use? Lorazepam is a real-world prescribing option for short-term insomnia. Showing that a cannabis-based preparation performs comparably to it, even in a small trial, moves the conversation from “does cannabis do something?” to “how does cannabis fit alongside existing options?” That is a more useful question for clinical decision-making, even if this trial is too small to settle it.
What are the risks of long-term benzodiazepine use for insomnia?
Benzodiazepines like lorazepam are effective for short-term insomnia but carry meaningful risks with extended use. Tolerance develops relatively quickly, requiring higher doses to achieve the same effect. Physical dependence is common, and stopping benzodiazepines abruptly after prolonged use can cause rebound insomnia that is often worse than the original problem, along with anxiety, irritability, and in some cases withdrawal seizures. Cognitive effects and fall risk, particularly in older adults, are additional concerns. This is precisely why many clinicians and patients seek alternatives. The Thai trial’s four-week duration does not address these longer-term risks for either treatment.
Can a Massachusetts patient use cannabis for insomnia based on this study?
Massachusetts has an active medical cannabis program, and insomnia is among the conditions that can support a cannabis certification through a qualified clinician. However, the products available in Massachusetts dispensaries are not the same as the Thai multi-herb formulation studied in this trial. Matching a patient to a cannabis product for sleep requires understanding the person’s specific insomnia pattern, underlying causes, current medications, and relevant history. This trial is directionally encouraging but does not translate into a straightforward product recommendation. A consultation with a cannabis-informed clinician is the appropriate first step.
Why does the multi-herb nature of the formulation matter so much?
When a treatment is a composite of multiple active ingredients, it is not possible to determine which ingredient, or which combination, is responsible for the observed outcome. The Anti-Pom-Leung Fever medicine contains several plant-based components alongside cannabis. Any or all of them may be contributing to the sleep improvements seen in the trial. This matters because it prevents direct extrapolation to cannabis-only products. If you use a cannabis oil or a cannabis gummy for sleep, you are not using what was tested here. The clinical signal is interesting, but the translation requires careful interpretation.
What would a better study on this topic look like?
A stronger follow-up trial would enroll several hundred patients rather than 82, use a well-characterized cannabinoid preparation with a documented and consistent cannabinoid profile, include a placebo arm alongside the active comparator, follow patients for at least three to six months including a post-treatment period to assess rebound insomnia risk, report insomnia subtypes and relevant comorbidities, and recruit participants from diverse populations including non-Asian cohorts. Biomarker enrichment, such as measuring inflammatory markers at baseline, could help identify who benefits most. These design features would allow the field to draw conclusions that are both clinically meaningful and broadly applicable.
Does cannabis disrupt REM sleep, and how does that affect its use for insomnia?
THC is associated with suppression of REM sleep, the stage associated with dreaming and emotional memory processing. This is a real pharmacological effect documented in multiple studies. For some patients, less REM sleep can be subjectively helpful in the short term, particularly those with nightmare-related disorders or trauma-associated sleep disruption. For others, REM suppression can diminish sleep quality over time and contribute to cognitive fog on waking. CBD appears to have a different profile and may support sleep without the same REM effects. The Thai study used a multi-herb formulation and did not report polysomnographic data, so it cannot speak to REM architecture. This is a meaningful gap when evaluating cannabis specifically for insomnia.
Should I stop taking my sleeping medication and switch to cannabis?
No. Stopping a benzodiazepine or other sleep medication without medical supervision carries real risks, including rebound insomnia and, with certain medications, withdrawal symptoms that can be serious. This study does not support unsupervised medication changes. If you are interested in exploring cannabis as part of your sleep management, the right approach is to raise it with your clinician, discuss the evidence honestly, and make any transitions gradually and with appropriate support. CED Clinic specializes in exactly this kind of conversation, and we welcome questions from patients who want to understand their options thoughtfully.
