Blog
Cannabis for Back Pain: What the First Big Trial Found
Cannabis for Back Pain: What the First Big Trial Found
Cannabis for Back Pain: What the First Big Trial Found
For about twenty years, anyone asking whether cannabis helps back pain ran into the same wall. There were trials. They were careful. Almost none of them had studied back pain.
That changed in late 2025, and almost nobody noticed.
A phase 3 trial, the large-scale human study a drug must pass before approval, enrolled 820 adults with chronic low back pain in Nature Medicine. It gave them a standardized cannabis extract or a placebo for twelve weeks, and it did not require anyone to have nerve pain to take part. It hit its primary target, the specific pain outcome the trial was built to measure. The cannabis group improved more than the placebo group.
Then you look at by how much, and the picture gets more interesting than either side of this argument usually admits.
The short answer: cannabis now has one large, properly controlled trial in ordinary chronic low back pain, and it worked. Pain in the cannabis group fell 1.9 points on an 11-point scale. The gap between cannabis and placebo was 0.6 points. Most of what patients felt happened in both groups. Side effects were more common on cannabis. Whether that trade is worth making depends on your pain, what you have already tried, and what you are taking now. Those questions belong to the physician reviewing your case, not to an article.
Back Pain Is Not One Problem
The National Institute of Neurological Disorders and Stroke describes pain as a biopsychosocial experience, shaped by biology, psychology, and a person’s circumstances together, which is why it concludes that treating pain is “both complex and individual.” Back pain shows why. The word covers more than one mechanism.
Nociceptive pain is tissue pain. A strained muscle, an irritated joint, a disc under load. Damage or stress somewhere in the structure, and nerves reporting it accurately.
Neuropathic pain is the nerve itself misfiring. Sciatica is the familiar version: a compressed root sending burning or electric signals down an uninjured leg. The alarm is broken rather than the building.
Both can run in the same patient at once, in proportions that differ from case to case. That distinction runs straight through the research literature on cannabis and pain. It matters before any of the numbers do.
Which Kind of Back Pain Does the Cannabis Evidence Cover?
Until 2025, almost all of it covered the nerve kind. So the first useful question is how much back pain that is.
A 2023 study in Pain Practice screened 1,957 patients with low back pain at a hospital pain clinic using painDETECT, a standard questionnaire for identifying a neuropathic component. 255 of them, 13.0 percent, reported neuropathic-like symptoms.
Treat that number carefully. It comes from one pain clinic in Seoul using one screening tool, and other studies using the same questionnaire in different populations have reported figures several times higher. The honest read is directional rather than precise: a minority of back pain patients carry a clear neuropathic component, and for the rest, the older evidence base was describing somebody else’s condition.
The same study found something more useful than the percentage. Time did not predict it. Patients with pain for ten years were no more likely to have a neuropathic component than patients with pain for three months, so you cannot reason your way to an answer from how long you have hurt. What did associate with it: a history of lumbar surgery, severe maximum pain, opioid use, lumbosacral radiculopathy (nerve root compression causing leg or foot symptoms), and sleep disturbance.
If none of those describes you, that does not close the question. It means the older literature gives you less to work with, not that cannabis is irrelevant to your case. That is the gap the 2025 trial was built to fill.
What no article can tell you is how your own treatment history, your current medications, and what you have already tried bear on whether cannabis is a reasonable next step. That is a clinical judgement, not a self-assessment quiz.
Find out whether your back pain qualifies for a medical cannabis evaluation
Why the Older Evidence Was Never About Your Back
In January 2026, Cochrane published an update to the most rigorous review anyone has done of cannabis for chronic pain. Randomized trials only, with patients and researchers kept unaware of who received cannabis and who received placebo. Minimum two weeks of treatment. Twenty-one studies, 2,187 participants. The reviewers graded the certainty of every finding.
Their inclusion rule excluded any trial on non-neuropathic pain.
That was the correct methodological choice. Mixing nerve pain with muscle pain would produce an answer about neither. But it means the review everyone cites as the last word on cannabis and pain is not a review of back pain, and quoting it as though it were is the most common mistake in the genre.
Within its own territory, it found less than the promotion suggests. Across all three formulations, THC-dominant, balanced THC and CBD, and CBD-dominant, none produced clear evidence of pain relief of at least half. The reviewers rated the certainty of those findings very low, a rating meaning the true effect could still turn out to be nothing at all. Two softer measures nudged positive for the balanced products, and the reviewers attached the qualifier that matters: these effects were not clinically relevant. For THC-dominant products, one finding pointed the other way. They may increase nervous system adverse events, seen across five studies and 439 participants.
That was the state of play. Careful, negative, and aimed at a pain type most back pain patients do not have.
Then Someone Finally Ran the Trial
If you have been told the evidence here is too thin to act on, the study a German and Austrian team published in September 2025 is the one to read first.
820 adults with chronic low back pain. 394 received a standardized full-spectrum cannabis extract, meaning one containing several cannabinoids including THC and CBD rather than a single isolated compound. 426 received placebo. Neither patients nor investigators knew which was which. Entry did not require a neuropathic designation, which is precisely why a trial like this had never been folded into the Cochrane reviews.
It hit its primary target. Pain in the cannabis group fell 1.9 points on an 11-point scale, beating placebo by a mean difference of 0.6 points, a gap the statistics put beyond reasonable doubt. In a six-month open-label extension, where everyone knew what they were taking, pain fell further, to 2.9 points.
Now the arithmetic most coverage of this trial skipped. The cannabis group improved by 1.9 and the gap over placebo was 0.6, so the placebo group improved by roughly 1.3 points. Two thirds of what patients felt happened in the placebo arm as well. The drug effect is real, and it is the smaller part of the experience.
Two other results deserve to travel with the headline. Adverse events were more common on cannabis than placebo, 83.3 percent against 67.3 percent, though the authors describe them as mostly mild to moderate and transient, with no signs of dependence or withdrawal. And a later randomized-withdrawal phase, designed to test whether people relapsed when the extract was taken away, did not meet its primary endpoint.
The honest summary is that it works, modestly, with more side effects than placebo, in a trial where most of the improvement was not attributable to the drug.
One disclosure this article owes you. The trial was sponsored by the company developing the extract, several authors declare payments from cannabis manufacturers, and the statistician is employed by the outside firm hired to analyze the data. None of that invalidates the methodology. What it means is that you are reading a sponsored trial, and independent replication has not arrived yet. Weight it accordingly.
So Why Does the Internet Sound So Sure?
Researchers have traced it to two things.
In 2022, a team at the Karolinska Institute pooled 20 placebo-controlled cannabinoid pain trials across 1,459 people and measured something unusual: whether media coverage tracked the results.
The first finding was the placebo response. Patients on placebo reported a moderate to large reduction in pain. And in the better-run trials, the ones at low risk of bias, the placebo responses were larger rather than smaller. Rigorous design made the placebo effect more visible. It did not make the drug more effective.
That does not mean those patients imagined feeling better. A placebo response is a measured reduction in reported pain, and reported pain is the thing being treated. It means that when someone takes cannabis for back pain and feels relief, a substantial share of that relief comes from expectation, context, and attention rather than from the compound. The relief is real. The cause is not what most people assume. The 2025 trial’s own numbers show the same pattern from the inside.
The second finding is the one worth sitting with. Media attention across these trials was proportionally high, with a strong positive bias, and was not associated with the clinical outcomes.
Coverage was enthusiastic whether or not the trial worked. That is the mechanism behind the gap you have probably noticed: a settled-sounding internet, and clinical guidelines refusing to settle.
What Happens Outside the Trials
The observational data tells a different story.
Orthopedic researchers at Rabin Medical Center in Israel followed 241 patients with chronic low back pain that had not responded to standard treatment for five years. These were not mild cases. Mean pain duration was 15.1 years, and every one of them had documented failure of at least a year of opioids, anticonvulsants, antidepressants, NSAIDs (anti-inflammatories such as ibuprofen), and physiotherapy. In clinical terms, “failure” means those treatments did not control the pain adequately, not that patients had stopped taking them.
Pain scores fell by more than five points. At year five, 89.2 percent had achieved at least a 30 percent reduction in pain and 77.2 percent at least 50 percent. Just under 93 percent were still on cannabis at five years. Adverse events ran about one per patient-year and 99.8 percent of graded events were mild.
Those numbers are far larger than anything the randomized trials produced, and the authors explain why the comparison needs care. These associations, they wrote, cannot be interpreted causally in the absence of a concurrent randomized control arm. Several things could produce the same result. A real pharmacological effect is one. So is the natural improvement that tends to follow when people enter a study at their worst point. So are expectation and self-selection, in a treatment every patient actively chose. So are general shifts in how people report pain over five years.
There was no control group. Every patient knew what they were taking. Every patient had chosen it. Given what the Karolinska team measured about expectation, that is not a small caveat.
Cannabis and Opioids, Handled Carefully
A second phase 3 trial randomized 384 chronic low back pain patients to the same cannabis extract or to commercially available opioids for roughly six months. Its primary endpoint was constipation, not pain, and on that it was decisive. Patients on the extract were four times less likely to develop constipation and three times less likely to need laxatives.
On pain it was far closer. Average reduction over six months was 2.50 points with the extract against 2.16 with opioids, a difference of 0.34 points that only just cleared the conventional threshold for statistical significance. The trial’s own week-27 comparisons were not significant at all. The tolerability result is the strong one here. The pain result is not, and nobody should read it as cannabis outperforming opioids for pain.
The wider picture agrees. A 2024 analysis in *BMJ Open* compared treatments across 90 trials and 22,028 patients, including treatments never tested head to head, and found probably little to no difference between cannabis and opioids for physical functioning, with reasonable confidence in that finding. For pain relief it found little to no difference too, though with weaker confidence. What it did find was that cannabis produced fewer discontinuations due to adverse events than opioids.
In the Israeli five-year cohort, opioid use alongside cannabis fell from 100 percent of patients at baseline to 4.6 percent at year five. The authors flag this outcome as hard to bias, since whether a patient is still filling a prescription is harder to misreport than a pain score. That does not make it a controlled observation. It remains a before-and-after count in a population that chose this treatment, with no comparison group.
An earlier California study looked at the same question in 180 back pain patients, analyzing the 61 on prescription opioids. Just over half, 50.8 percent, stopped opioids entirely. It took a median of 6.4 years. And among those who did not stop, most did not hold steady: 17 of 29 increased their use.
Forty-eight percent of patients felt that cannabis had helped them reduce their opioid intake, and that feeling did not predict actual reduction. Patients believing it was working were not the ones for whom it worked.
We have written before about how cannabis changes pain in the brain, and this is that finding showing up in prescription records. Cannabis can reliably change how pain feels. Whether it changes what you take is a different matter.
None of this makes cannabis a substitute for a prescribed medication, and nothing here should be read as a reason to stop one. Opioid tapering is often difficult, sometimes dangerous, and belongs entirely to the physician managing your prescription. If you want to raise it with them, our guide to the opioid conversation covers what that discussion usually involves and what to ask.
What About CBD Cream for Back Pain?
It is one of the most searched questions in this area, and it has the thinnest answer.
The trials above used whole-plant extracts taken by mouth, not creams. The Cochrane review’s CBD-dominant arm found no clear evidence of meaningful pain relief across five studies and 208 participants, and that was oral CBD in nerve pain, so it is not a direct test of a cream on a sore lower back either.
For topical CBD applied to the back specifically, there is no trial worth citing. Topicals carry a low risk profile, so the decision to try one is a small one. What a cream cannot do is reach a structural or neurological source of pain. If that is what drives your symptoms, knowing it changes what you should reasonably expect any topical to do, and nobody can honestly promise you otherwise.
Whether your back pain has that kind of source is one of the things an evaluation establishes.
See whether your pain history qualifies for a medical cannabis certification
Why We Will Not Name a Strain or a Product
No clinic can honestly answer that question.
No trial has compared strains or cultivars for back pain, so any answer would be invented. Strain names are not standardized between growers, the same name carries different chemistry from different shelves, and the indica and sativa labels predict little about what is in a product. A recommendation would be a guess wearing a lab coat.
Note what the 2025 trials actually used: a single standardized extract, made to a consistent specification, given at a controlled amount. That is not what sits on a dispensary shelf, and it is a real limit on how far those results transfer to a product chosen by name.
What a physician can do is more useful. They can work out whether your pain has a neuropathic component, which sets how much of the older evidence applies to you. That evaluation also covers everything you currently take, because the interactions here are worth watching. Knowing what to watch for means you will not mistake a change in how you feel for a change in your spine. For patients weighing this option, the evaluation at Compassionate Clinics happens over video with a licensed physician who reviews your pain history and your current medications, then tells you whether you qualify in your state.
The Honest Verdict
If you have chronic back pain and have been reading about cannabis long enough to lose track of what to believe, the position is less settled than the internet suggests and less empty than the guidelines make it sound.
There is now one large, well-run trial in ordinary back pain, and cannabis beat placebo in it. The margin was 0.6 points on an 11-point scale, most of the improvement showed up in the placebo group too, side effects were more common, and the withdrawal phase did not confirm the effect. Against opioids it is clearly gentler on the gut and, at best, marginally better on pain. The older reviews, which studied nerve pain, found less than that.
Against all of it, patients who had failed every standard treatment report substantial and durable relief over five years and come off opioids at rates that are hard to dismiss. Those studies had no control groups.
Both of those paragraphs are true. Anyone giving you only one of them is selling you something.
That adds up to something messier than either side of this argument usually admits. It is a modest, real effect that some people will find worth the side effects and others will not, and the deciding factors are your pain type, what you have already tried, and what you are taking now. Those are not things you can determine from an article. Not even this one.
Find out which kind of back pain you have and whether cannabis is a reasonable next step
Common Questions
Does cannabis help back pain?
In one phase 3 trial of 820 patients with chronic low back pain, yes, modestly. Pain fell 1.9 points on an 11-point scale, against a 0.6-point margin over placebo, with more side effects than placebo. Older reviews covering nerve pain found no clear evidence of relief of at least half.
Does weed help back pain, or is it just the high?
The 2025 trial compared a standardized extract against a placebo, with neither patients nor investigators knowing which was which, so the result is not simply intoxication. But most of the improvement patients reported also occurred on placebo, so expectation is doing a large share of the work.
Does THC help back pain?
The trial that worked used a full-spectrum extract containing THC, not an isolated compound. In the older nerve-pain reviews, THC-dominant products showed no clear evidence of meaningful relief and may increase nervous system side effects.
Is cannabis better than the medication I take now?
One trial compared the extract directly against opioids in back pain. In that trial it was better tolerated, four times less likely to cause constipation, and only marginally better for pain. A 90-trial analysis found little to no difference in pain relief and fewer discontinuations with cannabis. None of that establishes it is the better choice for you, and any change to a prescription belongs to the physician managing it.
Will CBD cream help my back?
There is no trial of topical CBD for back pain worth citing. Topicals are low risk, so trying one is a small decision, but nobody can honestly promise it will work, and a cream cannot reach a structural source of pain.
How do I know if my back pain is the nerve kind?
A physician determines it through examination and history rather than a questionnaire you can score at home. Factors associated with a neuropathic component include nerve root compression (lumbosacral radiculopathy), previous lumbar surgery, severe peak pain, sleep disturbance, and current opioid use.
Does back pain qualify for a medical cannabis certification?
It depends on your state. Chronic pain is among the most widely listed qualifying conditions across state programs, and some states name specific spinal conditions, though the wording and the evidence each program requires vary considerably. An evaluation with a licensed physician tells you what applies where you live.
This article is for educational purposes and is not medical advice. It does not diagnose, treat, or replace care from a licensed clinician. Cannabis affects each person differently. Never stop or change a prescribed medication, including an opioid, without speaking to the physician managing it.
Sources
1. Karst M, Meissner W, Sator S, Keßler J, Schoder V, Häuser W. Full-spectrum extract from Cannabis sativa DKJ127 for chronic low back pain: a phase 3 randomized placebo-controlled trial. Nature Medicine, 2025;31(12):4189-4196. PMID 41023483.
2. Meissner W, Argoff C, Sator S, Schoder V, Karst M. VER-01 Shows Enhanced Gastrointestinal Tolerability, Superior Pain Relief, and Improved Sleep Quality Compared to Opioids in Treating Chronic Low Back Pain: A Randomized Phase 3 Clinical Trial. Pain Therapy, 2025;14(6):1765-1782. PMID 41028525.
3. Ates G, Welsch P, Klose P, Phillips T, Lambers B, Hauser W, Radbruch L. Cannabis-based medicines for chronic neuropathic pain in adults. Cochrane Database of Systematic Reviews, 2026;1(1):CD012182. PMID 41548880.
4. Kim HJ, Ban MG, Yoon KB, Yang YS, Kim SH. Neuropathic pain component and its association with time elapsed since pain onset in patients with low back pain. Pain Practice, 2023;23(6):580-588. PMID 36861853.
5. Gedin F, Blome S, Ponten M, et al. Placebo Response and Media Attention in Randomized Clinical Trials Assessing Cannabis-Based Therapies for Pain: A Systematic Review and Meta-analysis. JAMA Network Open, 2022;5(11):e2243848. PMID 36441553.
6. Jeddi HM, Busse JW, Sadeghirad B, et al. Cannabis for medical use versus opioids for chronic non-cancer pain: a systematic review and network meta-analysis of randomised clinical trials. BMJ Open, 2024;14(1):e068182. PMID 38171632.
7. Robinson D, Khatib M, Lavon E, et al. Long-Term Inhaled Cannabis Therapy for Chronic Low Back Pain: A Five-Year Retrospective Analysis of Prospectively Collected Patient-Reported Outcomes in 241 Treatment-Refractory Patients. Biomedicines, 2026;14(6):1255. PMID 42351683.
8. Takakuwa KM, Hergenrather JY, Shofer FS, Schears RM. The Impact of Medical Cannabis on Intermittent and Chronic Opioid Users with Back Pain. Cannabis and Cannabinoid Research, 2020;5(3):263-270. PMID 32923663. *(The first author discloses a financial interest in a medical cannabis company that did not exist when the study was performed.)*
9. National Institute of Neurological Disorders and Stroke. Pain. https://www.ninds.nih.gov/health-information/disorders/pain
