A Phase 3 Trial Put a Cannabis Extract Head to Head With Opioids. The Primary Endpoint Was Constipation.

In a randomized phase 3 trial of 384 adults with chronic low back pain, participants taking the standardized cannabis extract VER-01 were roughly fourfold less likely to develop constipation than those taking opioids (relative risk 0.25, 95% CI 0.09 to 0.69) and about threefold less likely to need a laxative (RR 0.34, 95% CI 0.18 to 0.65).

Pain relief favored the extract over six months, but by a mean difference of 0.34 points on an 11-point scale (95% CI 0.00 to 0.67). That is a statistically significant result and a small one, and the difference at week 27 taken alone did not reach significance.

Chronic low back pain affects more than half a billion people worldwide. Opioids remain widely prescribed for it despite limited efficacy and well-documented harms, and the most common reason patients ask about cannabis in a pain clinic is that they want off, or want to avoid starting.

Almost every cannabis pain trial answers a different question than the one patients are asking. Placebo-controlled trials tell you whether the drug beats nothing. This one tells you how it performs against the actual alternative, which is the comparison that determines whether a prescription changes.

Winfried Meissner of Jena University Hospital, Charles Argoff of Albany Medical Center, Sabine Sator of the Medical University of Vienna, Volker Schoder of Metronomia Clinical Research, and Matthias Karst of Hannover Medical School published the trial in Pain and Therapy on September 30, 2025 (2025;14(6):1765-1782). It was registered as NCT05610813 and EudraCT 2022-001358-41.

Three hundred eighty-four adults with chronic low back pain were randomized one to one, with 192 receiving VER-01, a standardized full-spectrum extract from Cannabis sativa DKJ127 L., and 192 receiving a range of commercially available opioids. After a three-week titration, participants were treated for 24 weeks and then washed out for two weeks.

The primary endpoint was not pain. It was the relative risk of constipation at 27 weeks. That design choice tells you what the sponsor believed its product’s strongest claim was, and readers should hold that in mind when interpreting the secondary analgesia results.

Participants on VER-01 were fourfold less likely to develop constipation than those on opioids, at a relative risk of 0.25 with a 95% confidence interval from 0.09 to 0.69 and a p value of 0.007. Laxative use followed the same pattern, with a relative risk of 0.34 (95% CI 0.18 to 0.65, p less than 0.001).

Opioid-induced constipation is not a cosmetic side effect. It is one of the most common reasons patients discontinue opioid therapy, it does not reliably improve with continued exposure the way sedation often does, and managing it adds a second medication to a regimen that was meant to simplify life rather than complicate it.

The mechanism is straightforward. Mu-opioid receptors in the enteric nervous system slow gut transit directly, and that effect is separable from analgesia. A drug that provides comparable pain relief through a different receptor system has a plausible reason to spare the bowel, and this trial measured that difference rather than assuming it.

Over six months of treatment, mean pain reduction was 2.50 points on the 11-point numeric rating scale with VER-01 versus 2.16 with opioids. The mean difference between groups was 0.34 points (95% CI 0.00 to 0.67, p equals 0.048). Sleep followed a similar shape, improving 2.52 points with the extract versus 2.07 with opioids (mean difference 0.45, 95% CI 0.11 to 0.79, p equals 0.009).

A confidence interval whose lower bound sits at 0.00 and a p value of 0.048 describe a result that cleared the statistical threshold by a margin close to nothing. More importantly, 0.34 points on an 11-point scale falls well below what pain researchers generally regard as a clinically important difference for an individual patient, which is usually estimated at one to two points.

The authors are also explicit that the longitudinal analysis over six months is what produced the significant difference. Secondary endpoints limited to week 27 alone were not significant. Both groups improved substantially from baseline. The honest summary is that the extract matched opioids on pain and edged slightly ahead, not that it clearly outperformed them.

The companion trial, published in Nature Medicine on September 29, 2025 by Matthias Karst and colleagues (2025;31(12):4189-4196), tested VER-01 against placebo in 820 adults with chronic low back pain, 394 on active treatment and 426 on placebo. It met its primary endpoint with a mean pain reduction of 1.9 points on the numeric rating scale in the VER-01 arm and a mean difference versus placebo of 0.6 points (95% CI -0.9 to -0.3, p less than 0.001).

That trial also included a randomized withdrawal phase, which did not meet its primary endpoint of time to treatment failure (hazard ratio 0.75, 95% CI 0.44 to 1.27, p equals 0.288), though pain did increase significantly more with placebo upon withdrawal. Adverse events in the double-blind phase were more common with VER-01 than placebo at 83.3% versus 67.3%, mostly mild to moderate and transient, with no signs of dependence or withdrawal.

Read together, the two trials tell a coherent story: a modest analgesic effect over placebo, roughly comparable analgesia to opioids, and a materially better tolerability profile in the gut. That is a useful profile. It is not a breakthrough in pain relief.

VER-01 is an investigational product manufactured to pharmaceutical standards from a proprietary cultivar with a defined and reproducible composition. A dispensary extract labeled full-spectrum is not the same product and has not been studied this way. Treating the two as interchangeable is the single most common error made when trials like this one reach patients.

The trial also used a supervised three-week titration and continued clinical oversight across 24 weeks of treatment and a two-week washout. Nobody in this trial swapped their opioid for a cannabis product on their own timeline. Abrupt opioid discontinuation carries its own risks, and any taper belongs with the prescriber who wrote the original prescription.

The comparison group received a range of commercially available opioids rather than a single fixed agent and dose, which reflects real prescribing but complicates interpretation. Different opioids differ in constipation liability, and a trial comparing against a specific agent at a specific dose would produce a different number.

For the gastrointestinal question this is good evidence. A randomized head to head design with an active comparator, a prespecified primary endpoint, and a 27-week horizon is the appropriate instrument for asking which drug constipates people more, and the confidence intervals around the constipation and laxative results sit comfortably away from 1.0.

For the analgesia question the evidence is thinner than the phrase superior pain relief suggests. A 0.34-point mean difference on an 11-point scale, with a confidence interval touching zero and a p value of 0.048, is a signal at the edge of detectability. The companion placebo-controlled trial in Nature Medicine reported a mean difference of 0.6 points over placebo, which is also modest. Nothing here supports describing the extract as a substantially stronger analgesic than opioids.

Both trials were sponsored by the company developing the product, and the primary endpoint of the head to head study was a tolerability measure on which the sponsor’s product had a strong mechanistic prior. Choosing constipation rather than pain as the primary endpoint is scientifically defensible and commercially convenient at the same time, and readers should hold both facts.

Blinding is the deeper problem. A trial comparing a THC-containing extract against opioids cannot fully conceal allocation from participants, because the two drug classes feel different. Pain and sleep were measured by patient-reported numeric rating scales, which are exactly the outcomes most vulnerable to knowing which arm you are in.

The opioid comparator was a range of agents rather than a fixed drug and dose. That mirrors clinical reality and it means the constipation comparison is against an average of prescribing practice rather than against a defined standard, which limits how precisely the result can be transferred to a specific patient on a specific opioid.

This trial does not show that a cannabis extract relieves chronic low back pain better than opioids in any way a patient would notice. A 0.34-point difference on an 11-point scale is below the threshold most pain researchers use for an individually meaningful change.

It does not show that dispensary cannabis products behave like VER-01. The extract is a standardized pharmaceutical preparation from a single proprietary cultivar with batch-to-batch consistency that retail products do not claim and are not required to demonstrate.

It does not establish that switching from an opioid to a cannabis extract is safe or effective for a patient already on long-term opioid therapy. Participants were randomized at the start rather than transitioned from an existing opioid regimen, and no taper protocol was tested.

The interesting contribution of this trial is not that cannabis beat opioids. It is that the two performed similarly on pain while differing sharply on a side effect that drives discontinuation. In chronic pain management, where most analgesic options plateau at similar modest efficacy, tolerability is frequently what determines whether a patient stays on treatment long enough to benefit.

That framing also sets a reasonable expectation for patients. The realistic promise of a cannabis-derived analgesic in chronic low back pain is not dramatic pain relief. It is comparable relief with a different and, for many people, more livable side effect profile, plus no observed dependence or withdrawal signal in the placebo-controlled companion trial.

The regulatory picture is moving alongside the science. The FDA granted VER-01 Breakthrough Therapy Designation for chronic low back pain on May 18, 2026 on the strength of these two phase 3 trials. A designation accelerates FDA engagement and review; it is not an approval, and a separate US pivotal phase 3 trial is underway with a first data readout anticipated in 2027 and a New Drug Application planned for 2028 if results support it.

The constipation finding is the one I would put in front of a patient first, because it is the part of this trial that is both large and easy to feel. Patients on long-term opioids describe the bowel effects as one of the reasons the treatment wears them down, and a fourfold difference in that risk is not a footnote.

I would be careful with the pain number. When a trial reports a 0.34-point advantage on a 0 to 10 scale, the honest translation is that the two treatments performed about the same. Overselling that as superior analgesia does the field no favors, and patients notice when a promise does not match what they experience.

The distinction I spend the most time on in clinic is between a standardized pharmaceutical extract and what is on a dispensary shelf. VER-01 is characterized batch to batch. A retail product with a similar description on the label is a different thing, and a trial result does not transfer to it. That is not a knock on the retail market; it is a statement about what was and was not studied.

Against opioids over six months, a standardized cannabis extract produced roughly equivalent pain relief with about a fourfold lower risk of constipation. Frame it to patients as a tolerability advantage rather than an analgesic upgrade, and be explicit that this product is not what they would buy at a dispensary.

Note which endpoint was primary. The trial was designed to test constipation, and it answered that question convincingly. The pain and sleep results were secondary, favored the extract by small margins, and depended on longitudinal analysis across six months rather than on the week 27 timepoint alone. Quote the constipation result with confidence and the pain result with its actual effect size attached.