# Cannabis Oil for Fibromyalgia: New RCT Results Explained

> **Author:** Benjamin Caplan, MD (ABFM Board Certified, MA Lic. #240562, NPI: 1962722637)
> **Publisher:** CED Clinic (https://cedclinic.com)
> **Canonical URL:** https://cedclinic.com/cannabis-oil-fibromyalgia-rct-results/
> **Published:** 2026-06-07T23:50:59-04:00 | **Last Modified / Clinically Audited:** 2026-09-28T10:43:14-04:00
> **Clinical Summary:** A cannabis oil pilot trial in fibromyalgia shows why the fibromyalgia placebo response, and failed blinding, complicate every result in this field.

---

CED Clinical Relevance #79Strong Clinical RelevanceFibromyalgia patients are among the most common presentations in [cannabis medicine](https://cedclinic.com/essential-oils-industrial-hemp-antimicrobial-benefits/), and they are also the population in which expectation effects are largest. Understanding why placebo responses run this high changes how a clinician should interpret both the trial and the patient sitting across the desk.

A small randomized trial of 1:1 THC to CBD cannabis oil in fibromyalgia produced a promising result for the active arm. It also produced two numbers that deserve equal attention: 40 percent of the placebo group reached a clinically meaningful pain reduction by week 12, and 75 percent of participants in each arm correctly guessed which treatment they had been assigned.

Clinical Insight | CED Clinic

In this randomized, double-blind, placebo-controlled pilot trial, 70 percent of participants receiving 1:1 THC to CBD cannabis oil reported at least a 30 percent pain reduction both after titration and at week 12. In the placebo arm the corresponding figures were 20 percent and 40 percent.

The authors also report that 75 percent of participants in each group correctly guessed their allocation, and state directly that this high rate of correct guessing may have introduced expectation bias into the patient-reported outcomes. Both findings come from the same paper, and reporting one without the other misrepresents it.

FibromyalgiaPlacebo ResponseTrial DesignCannabis OilEvidence Interpretation
AudiencePatients with fibromyalgia, caregivers, and cliniciansPrimary TopicPlacebo response and blinding failure in a randomized cannabis oil trial for fibromyalgiaSource[Read the full source](https://doi.org/10.1155/prm/7311235)
Why This Matters

Fibromyalgia is diagnosed by symptom report and measured by symptom report. Every outcome in a fibromyalgia trial passes through the patient’s own perception, which is precisely the channel that expectation acts on. That makes placebo response rates in this condition unusually high, and it makes trial interpretation unusually demanding.

Cannabis compounds the problem, because THC announces its presence. A participant who feels a psychoactive effect has been unblinded by the drug itself, no matter how carefully the placebo oil was matched. Any reader of cannabis pain trials needs to understand this, and most published coverage skips it entirely.

What the Trial Did

Inna Kurlyandchik, Evelin Tiralongo, Romy Lauche, Gerald Tracey, Jennifer Ng, and Janet Schloss, working at the National Centre for Naturopathic Medicine at Southern Cross University, Griffith University, and Gold Coast University Hospital, published the trial in *Pain Research and Management* on May 16, 2026 (2026;2026(1):e7311235). It was registered as ACTRN12623000345684.

Twenty-four adults with fibromyalgia, all female, were randomized to a 1:1 THC to CBD cannabis oil at 10 mg per millilitre of each cannabinoid, or to a matched placebo. Participants completed a four-week dose titration followed by twelve weeks of stable dosing. Mean age was 55.25 years in the treatment arm and 52.25 years in the placebo arm.

The primary objective was feasibility rather than efficacy: whether procedures worked, whether participants could be recruited and retained, and whether the intervention was tolerable enough to test at scale. Efficacy outcomes were secondary and explicitly preliminary.

The Feasibility Answer Was Yes, With One Exception

Retention was 91.7 percent, with 22 of 24 participants completing, and every participant took at least 90 percent of prescribed doses. For a sixteen-week trial in a condition marked by fatigue and symptom variability, that is a strong operational result.

Tolerability held up as well. Adverse events were mostly mild and no serious events occurred. The most common events in the treatment arm were somnolence, dizziness, fatigue, nausea, and dry mouth. Notably, the placebo arm reported its own burden, led by headache, diarrhea, nausea, and flu-like symptoms, with 121 adverse events recorded across both groups.

Recruitment was the exception. Of 77 people prescreened, 24 were randomized, a screening-to-enrolment ratio of roughly three to one. The trial reached only 66.7 percent of its 36-participant target, a shortfall the authors attribute mainly to geographic and legal barriers.

The Placebo Arm Is the Story

In the placebo group, 20 percent of participants reached at least a 30 percent pain reduction after titration. By week 12 that figure had doubled to 40 percent. Nothing pharmacologically active produced that change.

The authors address it directly, describing the substantial placebo response as consistent with the high placebo effects typical of fibromyalgia trials, where response rates can exceed 30 to 40 percent. This is not a flaw peculiar to this study. It is a structural feature of the condition and of how it is measured.

The practical consequence is arithmetic. When the comparator arm climbs from 20 to 40 percent over twelve weeks, the margin available for an active drug to demonstrate itself narrows considerably. A trial powered on an assumption of a low placebo response will be underpowered in practice, which is one reason fibromyalgia drug development has such a long record of failed phase 3 programs.

Blinding Did Not Hold, and the Authors Say So

Seventy-five percent of participants in each group correctly guessed their allocation. The authors write that despite implemented blinding procedures, this high rate of correct allocation guessing may have introduced expectation bias on subjective patient-reported outcomes, and attribute it to the distinctive psychoactive effects of THC.

They describe this as a well-documented inherent limitation of blinding in THC-containing cannabis trials, which it is. It is also rarely reported. Most cannabis trials either do not collect allocation guesses or do not publish them, which means readers cannot judge how much of a reported effect survives the unblinding problem.

Collecting and publishing that number is a mark of a carefully run study rather than a weakness in it. The appropriate response is to weight the results accordingly, not to discount the trial for being honest about something its peers usually leave unmeasured.

What the Efficacy Signals Actually Were

Secondary outcomes suggested medium to large between-group effects favoring cannabis for pain reduction, sleep quality, and fibromyalgia impact. Reported partial eta squared values were 0.180 at week 1 and 0.085 at week 12 for average daily pain score, 0.184 for the Pittsburgh Sleep Quality Index, and 0.132 for the Fibromyalgia Impact Questionnaire Revised total.

On the Fibromyalgia Impact Questionnaire, clinically meaningful improvement against a predefined minimal clinically important difference of 45.5 percent occurred in 40 percent of cannabis-treated participants versus 10 percent on placebo. Fatigue and anxiety or depression showed no significant changes.

The pattern, pain and sleep responding while fatigue and mood did not, matches what larger cannabinoid reviews have reported across other conditions. With 24 participants and known unblinding, this trial cannot confirm that pattern. It is consistent with it.

Study SnapshotStudy TypeRandomized, double-blind, placebo-controlled pilot trial; primary objective feasibility and safetyParticipants24 adults with fibromyalgia, all female; mean age 55.25 (treatment) and 52.25 (placebo) yearsIntervention1:1 THC:CBD cannabis oil, 10 mg/mL of each cannabinoid, versus matched placeboSchedule4-week dose titration followed by 12 weeks of stable dosingRecruitment77 prescreened, 24 randomized (31.2%); 66.7% of the 36-participant target reachedRetention and Adherence91.7% retention (22 of 24); all participants took at least 90% of prescribed dosesPain Response (30% threshold)Cannabis 70% post-titration and at week 12; placebo 20% post-titration and 40% at week 12FIQR ResponseClinically meaningful improvement (predefined MCID 45.5%) in 40% on cannabis vs 10% on placeboBlinding75% of participants in each group correctly guessed their allocationSafety and FundingNo serious adverse events; 121 adverse events across both arms, mostly mild. Study drug supplied by Little Green Pharma; lead author’s scholarship jointly funded by Southern Cross University and Little Green PharmaJournal and RegistrationPain Research and Management 2026;2026(1):e7311235. PMID 42142029 / DOI 10.1155/prm/7311235. ACTRN12623000345684
How Strong Is This Evidence?

As a feasibility trial this is good evidence, and feasibility was what it set out to test. Retention of 91.7 percent, adherence above 90 percent of doses in every participant, and no serious adverse events answer the operational questions a larger trial would need answered first. The recruitment shortfall is also useful information, because it tells the next investigator what to plan for.

As efficacy evidence it is preliminary by design and further limited by two features the authors name themselves: 24 participants, and a 75 percent correct allocation-guessing rate in each arm. Effect sizes derived from self-reported outcomes in a trial where most participants knew their assignment should be read as hypothesis-generating. The authors say as much.

Where This Paper Deserves Skepticism

The funding relationship deserves to be visible. The lead author’s doctoral scholarship was jointly funded by Southern Cross University and Little Green Pharma, the company that supplied both the investigational product and the placebo. The authors disclose this, and per their disclosures the company had no role in design, data, or analysis. Disclosure is the right response; it does not make the relationship disappear.

All 24 participants were female. That reflects the epidemiology of fibromyalgia, which is diagnosed far more often in women, and it still means nothing here speaks to men with the condition. Mean age in the mid-fifties also means the results do not extend cleanly to younger patients.

The recruitment shortfall is more than a power problem. Reaching 66.7 percent of target because of geographic and legal barriers means the enrolled sample consists of people willing and able to overcome those barriers, which is a selection filter on motivation and expectation in a trial whose outcomes are expectation-sensitive.

What This Paper Does Not Show

This trial does not establish that 1:1 THC to CBD cannabis oil relieves fibromyalgia pain. It was not designed or powered to do so, and the authors state that preliminary safety and efficacy signals warrant confirmation in larger, adequately powered trials.

It does not separate pharmacological effect from expectation effect. With 75 percent of participants in each arm correctly identifying their allocation, the between-group differences in self-reported pain, sleep, and fibromyalgia impact cannot be attributed cleanly to the drug.

It does not identify who responds. Twenty-four participants cannot support subgroup analysis, so nothing here indicates which fibromyalgia patients are likely to benefit, at what dose, or for how long. It also says nothing about effects beyond sixteen weeks.

How This Fits With the Broader Clinical Conversation

High placebo response is the defining methodological problem of fibromyalgia research, and it has sunk drug programs well outside cannabis. A condition whose every endpoint is a self-report, in a patient population with long histories of treatments that did not work and correspondingly strong hopes for one that might, is close to a worst case for detecting a modest drug effect.

Trialists have developed responses to this. Enriched enrollment designs with open-label run-in periods, used to identify and exclude high placebo responders before randomization, are one option the authors themselves recommend for future work. Objective or semi-objective endpoints, such as actigraphy for sleep, are another. Neither is a complete solution and both add cost.

For cannabis specifically there is an additional layer, because the active drug identifies itself. Until trials in this field routinely collect and publish allocation-guessing data, readers have no way to judge how much of a reported benefit reflects pharmacology. This trial published that number, which makes it more useful than many larger studies that did not.

Dr. Caplan’s Take

The placebo number in this trial is the one I keep coming back to. Four in ten women on an inert oil reached a 30 percent pain reduction by week 12. That is not a measurement error and it is not participants being suggestible. In a condition driven substantially by central pain processing, expectation, attention, and the experience of being carefully monitored by a clinical team are doing real physiological work.

That has two implications and they pull in different directions. It means I should be cautious about attributing every improvement my fibromyalgia patients report to the cannabis they started. It also means the structured attention that comes with a careful treatment plan is part of the treatment, and I would rather patients get that attention alongside a medication they tolerate than not get it at all.

What I appreciate about this paper is that the authors reported the allocation-guessing data. Most cannabis trials do not, and it is the single most informative number for judging a self-reported outcome in a study where the active drug is psychoactive. A team that publishes an inconvenient number is a team I read more carefully, not less.

Clinical Bottom Line

The active arm outperformed placebo on pain, sleep, and fibromyalgia impact, and 40 percent of the placebo arm also reached a meaningful pain reduction while 75 percent of everyone guessed their assignment correctly. Treat this as a well-run feasibility study that justifies a larger trial, and be explicit with patients that expectation is a real and substantial contributor in this condition.

What a Careful Reader Should Take Away

Read the placebo column before the treatment column. The finding to carry forward is that a properly conducted fibromyalgia trial is feasible and that expectation effects in this condition are large enough to swallow a modest drug effect. The finding not to carry forward is any efficacy claim, because with 24 participants and 75 percent correct allocation guessing, this design cannot separate the drug from the expectation of the drug.

Evidence Interpretation Guide

Table of Contents

- [How to read a trial when the comparator arm also gets better](#How_to_read_a_trial_when_the_comparator_arm_also_gets_better)
- [Placebo, Blinding, and Cannabis in Fibromyalgia, Seen From Eight Angles](#Placebo_Blinding_and_Cannabis_in_Fibromyalgia_Seen_From_Eight_Angles)[A placebo response is not imaginary](#A_placebo_response_is_not_imaginary)
- [Set expectations before you set a dose](#Set_expectations_before_you_set_a_dose)
- [The unblinding rate caps what this can show](#The_unblinding_rate_caps_what_this_can_show)
- [Twenty-four participants and a sponsor-supplied product](#Twenty-four_participants_and_a_sponsor-supplied_product)
- [The prior literature is thinner than it appears](#The_prior_literature_is_thinner_than_it_appears)
- [What the protocol required](#What_the_protocol_required)
- [Designs that could actually answer this](#Designs_that_could_actually_answer_this)
- [Access rules shaped who could enroll](#Access_rules_shaped_who_could_enroll)
- [Frequently Asked Questions](#Frequently_Asked_Questions)[Related](#Related)

## How to read a trial when the comparator arm also gets better

The Research QuestionIs a randomized, double-blind trial of 1:1 THC to CBD cannabis oil in fibromyalgia feasible, safe, and tolerable enough to justify a larger study?The Patient QuestionIf people on the placebo also felt better, how do I know whether cannabis is doing anything for my fibromyalgia?The Bottom LineFeasibility was demonstrated with 91.7 percent retention and no serious adverse events. Efficacy was not, because 40 percent of the placebo arm improved meaningfully and 75 percent of participants knew which arm they were in.
CED Perspective Lens

## Placebo, Blinding, and Cannabis in Fibromyalgia, Seen From Eight Angles

One small pilot trial, read through the lenses that matter in clinical practice.

Lens OverviewA 24-participant randomized trial of 1:1 THC to CBD cannabis oil in fibromyalgia reported 70 percent of the cannabis arm reaching a 30 percent pain reduction, alongside 40 percent of the placebo arm by week 12 and 75 percent of participants correctly guessing their allocation.

#### A placebo response is not imaginary

Forty percent of the women on inert oil in this trial reported at least a 30 percent reduction in pain by week 12. Their relief was real to them and measurable on the same scale used for everyone else. Placebo response in fibromyalgia reflects genuine changes in how the nervous system processes pain, not participants making things up.

What that means for you is that feeling better after starting something new is good news and weak evidence about the something. Giving a treatment a defined trial period, tracking symptoms in writing, and reviewing with a clinician is how you tell the difference over time.

Lens takeawayRelief is real even when the cause is uncertain. Track it, and give the treatment a defined review date.

#### Set expectations before you set a dose

In a condition with a 30 to 40 percent placebo response rate, a patient who improves after starting cannabis has not demonstrated that cannabis worked. Framing that at the outset, before the first dose, protects the patient from concluding a treatment failed when a regression toward baseline occurs months later.

The practical structure is a defined trial period, a written symptom record, and a prespecified decision point. That approach borrows from trial design and it works in clinic for exactly the same reason it works in research.

Lens takeawayDefine the trial period and the stopping rule before starting, not after the first disappointment.

#### The unblinding rate caps what this can show

Seventy-five percent of participants in each arm correctly identified their allocation. Every efficacy outcome in this trial was self-reported. Those two facts together mean the between-group differences cannot be attributed confidently to pharmacology.

This is not a criticism of the investigators, who measured and published the problem. It is a reason to read the effect sizes as upper bounds rather than as estimates.

Lens takeawaySelf-reported outcomes plus known unblinding equals effect sizes that should be read as ceilings.

#### Twenty-four participants and a sponsor-supplied product

Recruitment reached 66.7 percent of target, leaving 24 randomized participants and correspondingly wide confidence intervals around every estimate. Subgroup analysis is impossible and the risk that a chance imbalance drove a result is not small.

Little Green Pharma supplied the investigational product and the placebo and jointly funded the lead author’s scholarship. The authors disclose this and report that the company had no role in design, data, or analysis. Readers should note it and weight accordingly.

Lens takeawaySmall sample, industry-supplied product, disclosed funding relationship. Note all three.

#### The prior literature is thinner than it appears

Earlier fibromyalgia [cannabis research](https://cedclinic.com/pediatric-cannabis-certification-care-children/) is largely observational. Habib and Artul, in the Journal of Clinical Rheumatology in 2018, reported registry data on 26 patients with reported improvement across questionnaire parameters. Sagy and colleagues, in the Journal of Clinical Medicine in 2019, followed 367 fibromyalgia patients prospectively for six months and reported median pain falling from 9.0 to 5.0, with a 70.8 percent response rate.

Those designs have no comparator arm at all, which means the 40 percent placebo response documented here sits inside every one of their reported benefits, unmeasured and unremovable.

Lens takeawayUncontrolled studies cannot subtract a placebo response they never measured.

#### What the protocol required

Participants titrated over four weeks before twelve weeks of stable dosing, at 10 mg per millilitre each of THC and CBD. Adverse events were mostly mild and concentrated in the familiar profile: somnolence, dizziness, fatigue, nausea, dry mouth.

The placebo arm reported its own adverse events, led by headache and diarrhea, which is a reminder that not every symptom appearing after a new treatment is caused by it. That applies in clinic as much as in a trial.

Lens takeawaySlow titration, a stable maintenance period, and awareness that side effects appear on placebo too.

#### Designs that could actually answer this

The authors recommend enrichment strategies such as open-label run-ins to identify and exclude high placebo responders before randomization. Adding semi-objective endpoints, actigraphy for sleep rather than questionnaire alone, would reduce dependence on self-report.

Routine collection and publication of allocation-guessing data should become standard in any trial of a psychoactive compound. Without it, a reader cannot assess the one bias most likely to be driving the result.

Lens takeawayEnriched enrollment, semi-objective endpoints, and published blinding data are the three fixes available now.

#### Access rules shaped who could enroll

The recruitment shortfall was attributed mainly to geographic and legal barriers, including restrictions affecting participants in a THC trial. Regulatory friction did not just slow this study down; it changed which patients were able to take part.

In Massachusetts, fibromyalgia is not named in the statutory list of debilitating conditions. It can qualify when a certifying healthcare provider determines in writing that it is debilitating for that patient, which places the decision with the clinician rather than with a fixed list.

Lens takeawayRegulatory friction selects the sample. In Massachusetts, fibromyalgia qualifies through provider determination rather than by statute.
Related Reading at CED ClinicContinue exploring the evidence[Cannabis Oil for Fibromyalgia: New RCT Results Explained](https://cedclinic.com/cannabis-oil-fibromyalgia-rct-results/)

Our primary coverage of this trial, with the full efficacy and safety breakdown alongside the study limitations.

[Read the full trial coverage](https://cedclinic.com/cannabis-oil-fibromyalgia-rct-results/)[Cannabis Oil and Fibromyalgia: Clinical Benefit Without Blood Biomarker Change](https://cedclinic.com/oral-thc-cbd-fibromyalgia-trial-no-blood-biomarker-shift/)

The companion biomarker analysis from the same participants, which looked for an objective signal and did not find one.

[See the biomarker findings](https://cedclinic.com/oral-thc-cbd-fibromyalgia-trial-no-blood-biomarker-shift/)[What the 12-Week Pilot Actually Found About Topical CBD for Fibromyalgia Pain](https://cedclinic.com/topical-cbd-fibromyalgia-pain/)

A different cannabinoid route in the same condition, and another pilot where placebo interpretation is central.

[Compare the pilot data](https://cedclinic.com/topical-cbd-fibromyalgia-pain/)

Join the Conversation

Have a question about how this applies to your situation? [Ask Dr. Caplan](https://benjamincaplan.com/ask-dr-caplan)

Want to discuss this topic with other patients and caregivers? [Join the forum discussion](https://cedclinic.com/interact/discussion-forums/forum/cannabis-industry/)

## Frequently Asked Questions

How large is the placebo response in fibromyalgia trials?

Large enough to obscure a modest drug effect. In this randomized trial, 20 percent of the placebo group reached at least a 30 percent pain reduction after titration, rising to 40 percent by week 12. The authors describe this as consistent with high placebo effects typical of fibromyalgia trials, where response rates can exceed 30 to 40 percent. Every endpoint in the condition is self-reported, which is the channel expectation acts on.

Does a placebo response mean the pain relief was not real?

No. Placebo response in chronic pain reflects measurable changes in how the nervous system processes pain signals, driven by expectation, attention, and the experience of structured clinical care. Participants reporting relief on placebo are not fabricating it. What a placebo response means is that the relief cannot be attributed to the pharmacology of the drug being tested.

What did this cannabis oil trial find?

Seventy percent of participants on 1:1 THC to CBD cannabis oil reported at least a 30 percent pain reduction both after titration and at week 12, versus 20 percent and 40 percent on placebo. Clinically meaningful improvement on the Fibromyalgia Impact Questionnaire occurred in 40 percent on cannabis versus 10 percent on placebo. Sleep quality improved. Fatigue and anxiety or depression showed no significant change.

Why does blinding fail in cannabis trials?

Because THC produces effects the participant can feel. In this trial 75 percent of participants in each group correctly guessed their allocation, and the authors attribute that to the distinctive psychoactive effects of THC, describing it as a well-documented inherent limitation of blinding in THC-containing trials. Once participants know their assignment, self-reported outcomes carry expectation bias that cannot be separated from drug effect.

Was this trial large enough to prove cannabis helps fibromyalgia?

No, and it was not designed to. With 24 participants, the primary objective was feasibility: whether the procedures worked, whether people could be recruited and retained, and whether the intervention was tolerable. The authors state that preliminary safety and efficacy signals warrant confirmation in larger, adequately powered trials. Efficacy conclusions should not be drawn from a sample this size.

Was the cannabis oil safe in this trial?

It was reasonably well tolerated. No serious adverse events occurred in either group, and 22 of 24 participants completed the trial. The most common events in the cannabis arm were somnolence, dizziness, fatigue, nausea, and dry mouth. The placebo arm reported its own events, led by headache and diarrhea, with 121 adverse events recorded across both groups and most classified as mild.

How does this compare with earlier fibromyalgia cannabis research?

Earlier work was mostly uncontrolled. Habib and Artul reported registry data on 26 patients in the Journal of Clinical Rheumatology in 2018. Sagy and colleagues followed 367 fibromyalgia patients prospectively for six months in the Journal of Clinical Medicine in 2019, reporting median pain falling from 9.0 to 5.0. Neither had a placebo arm, so the placebo response documented here is embedded in their results, unmeasured.

Can I get medical cannabis for fibromyalgia in Massachusetts?

Fibromyalgia is not named in the Massachusetts list of debilitating conditions, but the program allows a certifying healthcare provider to determine in writing that another condition is debilitating for a given patient. That places the decision with the clinician rather than with a fixed list. A physician experienced in cannabis medicine can assess whether certification is appropriate given your history and medications.

 

### Related

---

### Verification & Medical Governance

- **Lead Clinical Author:** Benjamin Caplan, MD
- **Credentials:** Board Certified in Family Medicine (ABFM), Former CMO Canna Care Docs (250,000+ patients supervised), Founder & CMO CED Clinic & CED Foundation, Author of *The Doctor-Approved Cannabis Handbook* (Penguin Random House).
- **Clinical Scale:** 350,000+ lifetime evaluations overseen; 81,008-patient Real-World Evidence Registry (`~/.ced_rwe_vault.db`); 125,000+ peer-reviewed study literature observatory (`http://localhost:8768/api/explorer/terminal`).
- **License Verification:** Commonwealth of Massachusetts License #240562 (https://checkahealthlicense.mass.gov/).
- **NPI Registry:** 1962722637 (https://npiregistry.cms.hhs.gov/provider-view/1962722637).
- **Ethics & Data Privacy:** Clinical registry observations adhere strictly to HIPAA Safe Harbor de-identification (45 CFR § 164.514(b)) and criteria for secondary research exemption under 45 CFR § 46.104(d)(4).