Medical cannabis was associated with sustained reductions in chronic pain over two years in a large new study involving 3,744 patients, adding important long-term evidence to a field where most controlled cannabis trials have lasted only a few weeks or months. Researchers following patients in the UK Medical Cannabis Registry found improvements in pain severity, the extent to which pain interfered with everyday life, sleep, anxiety and overall health-related quality of life after patients began treatment with prescription cannabis-based medicinal products. More than half of participants achieved improvements considered clinically meaningful on the study’s pain measurements after 24 months.
Published in the Journal of Clinical Pharmacology in October 2026, the analysis is particularly significant because it did more than calculate an average response. Researchers used statistical clustering to identify five different pain trajectories, revealing substantial differences in how individual patients responded over time. Every group nevertheless ended the two-year observation period with lower average pain severity than it had at the beginning. The findings do not prove cannabis caused the improvements because the study was observational and lacked a placebo group, but they provide one of the largest examinations yet of how chronic-pain patients fare during long-term medical cannabis treatment.
Researchers Followed 3,744 Chronic-Pain Patients for Two Years
The study was led by Simon Erridge and colleagues from the Medical Cannabis Research Group at Imperial College London and other UK institutions. Participants were enrolled in the UK Medical Cannabis Registry and had chronic pain as their primary reason for receiving cannabis-based medicinal products. Rather than limiting the study to a single condition such as arthritis, neuropathy or back pain, the registry included patients with a range of chronic-pain diagnoses, allowing researchers to examine how treatment outcomes varied across a large real-world population.
Patients completed standardized questionnaires before beginning treatment and repeatedly during the following 24 months. Investigators measured pain severity and pain interference using the Brief Pain Inventory-Short Form, while additional questionnaires assessed general health-related quality of life, anxiety and sleep. The researchers then used longitudinal k-means clustering, a statistical method capable of grouping people according to the pattern of their pain scores over time rather than simply comparing the entire population at the beginning and end of treatment.
An earlier analysis of the same 3,744-patient chronic-pain cohort reported an average age of approximately 46 years. About 57 percent were already using cannabis when they entered the registry, an important detail when interpreting the findings because many participants already had experience with the drug and may have been more likely to tolerate or perceive benefit from cannabis treatment. The research nevertheless provides unusually long follow-up data for a medical cannabis population.
Pain Improvements Were Still Present at 24 Months
The central finding was that pain outcomes improved and remained improved through the two-year follow-up. The newly published analysis reported a moderate effect-size improvement in Brief Pain Inventory pain severity and pain interference at 24 months. Overall health-related quality of life also improved significantly. More than half of patients achieved the minimal clinically important difference on each pain measure—meaning the change was large enough to be considered meaningful to patients rather than merely statistically detectable.
Earlier data from this cohort provide a useful indication of the magnitude involved. Average Brief Pain Inventory severity scores fell from approximately 5.97 before treatment to 5.05 after one month, 4.85 after three months, 4.74 after six months and roughly 4.81 after 24 months. Pain interference, the Short-Form McGill Pain Questionnaire and pain visual analogue scores also showed significant improvements during follow-up. The largest changes appeared relatively early and were then generally maintained rather than continuing to decline indefinitely.
That pattern could be clinically meaningful for people with persistent pain because chronic pain treatment is often less about eliminating symptoms completely than reducing them enough to improve sleep, mobility, work, social activity and everyday functioning. A therapy that lowers pain and maintains the effect for two years may therefore matter even if average pain scores remain well above zero.
Five Different Pain-Response Patterns Emerged
One of the most interesting aspects of the research was the discovery of five distinct pain trajectories. Medical cannabis did not produce one uniform response across thousands of patients. Some groups experienced substantially stronger improvements, while others changed more gradually or achieved comparatively modest reductions. Importantly, all five clusters finished the study with average pain severity below where they started.
The differences may help explain why medical cannabis studies often produce conflicting conclusions. When results from every patient are averaged together, a substantial response among one subgroup can be diluted by little or no response in another. Conversely, enthusiastic reports from patients who respond extremely well can make a treatment appear more universally effective than it really is. The trajectory analysis suggests both phenomena may be occurring with cannabis: some chronic-pain patients appear to obtain meaningful and lasting benefit, while the magnitude of improvement varies greatly between individuals.
Researchers found that baseline patient characteristics were more informative in predicting those trajectories than the specific cannabis product or dose prescribed. Sleep quality, anxiety and the underlying pain indication showed stronger relationships with treatment response than product-related variables. That finding points toward a more individualized model of cannabinoid medicine in which identifying the right patient may ultimately prove at least as important as choosing a particular THC-to-CBD ratio.
Sleep and Anxiety May Be Closely Connected With Pain Improvement
Chronic pain rarely exists in isolation. Persistent pain can interfere with sleep, while poor sleep can make pain feel more intense the following day. Anxiety can increase attention to painful sensations and contribute to muscle tension, avoidance behavior and reduced physical activity. Over months or years, pain, anxiety and insomnia can reinforce one another in a cycle that substantially reduces quality of life.
The new registry study found improvements not only in pain but also in sleep, anxiety and overall health-related quality of life. Researchers reported that baseline sleep and anxiety were among the variables most strongly associated with a patient’s eventual pain trajectory. This does not establish that cannabis reduced pain because it improved sleep, or vice versa, but it suggests successful treatment may involve several interconnected symptoms rather than a purely analgesic effect.
Similar results have appeared in other UK Medical Cannabis Registry studies. A 2026 analysis of 240 people with hypermobility-associated chronic pain found sustained improvements in pain, sleep and health-related quality of life over 24 months, with more than half of patients reaching clinically meaningful improvements on several pain measures. Another 2026 study of 192 patients with inflammatory arthritis reported significant pain and sleep improvements throughout two years of follow-up.
Opioid Prescriptions Also Declined During Follow-Up
An earlier presentation of the 3,744-patient dataset reported that prescribed opioid exposure declined during cannabis treatment. Average opioid dosage, measured in oral morphine equivalents, fell from approximately 33.09 milligrams per day at baseline to 26.77 milligrams per day after 24 months. Researchers described the change as an association rather than evidence that medical cannabis directly caused patients to reduce opioids.
The distinction matters because opioid doses may change for many reasons. Physicians could adjust medications independently of cannabis, patients whose health improved may have needed fewer analgesics, or people motivated to use medical cannabis may also have been motivated to reduce opioid therapy. Without randomized assignment, the study cannot isolate cannabis as the reason for the decline.
Even so, the finding is consistent with several other observational studies suggesting some chronic-pain patients reduce opioid consumption after beginning cannabis treatment. A 2026 two-year study of patients with hypermobility-associated pain similarly found reductions in concomitant opioid prescriptions at later follow-up points. Whether cannabis can reliably produce an opioid-sparing effect remains a major research question, especially because high-quality randomized human evidence has been much less consistent than surveys and registry studies.
Most Reported Adverse Events Were Mild or Moderate
Safety becomes increasingly important when a medication is used for years rather than days. In the new analysis, 538 of the 3,744 patients—14.37 percent—reported at least one adverse event during follow-up. Most events were categorized as mild or moderate rather than severe.
Cannabis medicines can nevertheless produce clinically important side effects. THC-containing products can cause dizziness, fatigue, impaired attention, altered perception, anxiety, increased heart rate and intoxication. Some people develop tolerance or physical dependence with repeated THC exposure, and long-term heavy use can lead to cannabis use disorder. Older adults may also face increased concerns about falls and medication interactions.
A separate two-year longitudinal study published in 2026 examined cannabis use disorder risk among 1,453 people with chronic pain. Cannabis users with higher risk scores tended to have worse mental-health symptoms and more severe pain at baseline, although their pain did not worsen more rapidly over two years than that of nonusers. The findings reinforce the need to evaluate both benefits and dependence-related risks when cannabis becomes a long-term pain treatment.
The New Results Are Stronger on Duration Than on Causation
The 24-month follow-up addresses one of the largest gaps in previous cannabis pain research: duration. A major BMJ systematic review analyzed 32 randomized clinical trials involving 5,174 chronic-pain patients and concluded that non-inhaled medical cannabis produced small improvements in pain, physical functioning and sleep compared with placebo. However, trials lasted only one to about five and a half months, leaving researchers with little rigorous evidence about whether benefits persisted for years.
The new registry study provides precisely the kind of long-term information randomized trials have generally lacked, but it sacrifices the experimental control those trials provide. There was no placebo group, patients knew they were receiving cannabis, and clinicians could modify treatment throughout follow-up. Chronic pain naturally fluctuates, and patients often seek new treatments when symptoms are particularly severe, meaning some improvement can occur through regression toward the average even without an effective intervention.
The strongest interpretation is therefore that prescription cannabis was associated with sustained improvement rather than proven to have caused it. The authors themselves emphasized that high-quality randomized controlled trials remain necessary to establish efficacy and safety conclusively.
Existing Research Suggests Cannabis Has Real but Usually Modest Analgesic Effects
The broader evidence on cannabis and pain has gradually become more nuanced. The National Academies concluded in its landmark review that substantial evidence supported cannabis or cannabinoids for chronic pain in adults. Subsequent randomized-trial reviews have generally agreed that a therapeutic effect exists but have characterized its average magnitude as small.
That does not necessarily conflict with the stronger improvements seen in some registry patients. Randomized trials estimate the average effect attributable specifically to treatment compared with placebo. Observational registries capture the overall change a patient experiences after starting therapy, which can include pharmacological effects, placebo effects, medication changes, improved sleep and lifestyle changes occurring at the same time.
The new trajectory approach may help connect those two bodies of evidence. If only a portion of patients obtain a large response, an average randomized-trial effect could remain modest even though the treatment is highly useful for selected individuals. Identifying those responders prospectively is likely to become an increasingly important goal of cannabis research.
The Registry’s Structure Is Another Important Limitation
The UK Medical Cannabis Registry is a prospective real-world research database that follows patients receiving prescription cannabis products. It has generated a growing number of studies covering chronic pain, arthritis, migraine, endometriosis, cancer symptoms, anxiety and other conditions. The approach offers researchers access to populations far larger and longer-followed than most cannabis clinical trials.
The registry is privately managed through Curaleaf Clinic, and a number of investigators involved in registry publications are employed by or otherwise affiliated with the clinic. That does not determine whether the findings are correct, but independent replication is important, particularly when observational evidence is being used to influence treatment decisions. Earlier registry publications have explicitly acknowledged these affiliations and the inability of the observational design to demonstrate causation.
Another limitation is that more than half of the chronic-pain patients were already cannabis users when they entered the study. People who previously tried cannabis and found it intolerable or ineffective may therefore be underrepresented, potentially making the treatment appear more successful or better tolerated than it would in a completely cannabis-naïve population.
Long-Term Data Strengthen the Case for More Definitive Trials
The 3,744-patient study does not settle the debate over medical marijuana for chronic pain, but it substantially improves the amount of long-term information available. Pain severity and interference remained below baseline after two years, quality of life improved, more than half of patients experienced clinically meaningful changes in pain, opioid prescribing declined in earlier analyses of the same cohort and adverse events were reported by fewer than one in six participants.
Perhaps more importantly, the study demonstrates that “does cannabis work for pain?” may be the wrong question. The five different pain trajectories show that patients do not respond uniformly. A more useful question may be which patients are most likely to improve, what combination of THC, CBD and delivery method best fits their condition, and how clinicians can identify people at increased risk of side effects or dependence before treatment begins.
For patients living with persistent pain that has not responded adequately to conventional treatment, the findings offer evidence that medical cannabis can be associated with benefits lasting far beyond the short follow-up periods typical of clinical trials. They do not establish cannabis as a universal solution, nor do they justify replacing proven therapies indiscriminately. What they provide is a stronger argument for large, independent randomized studies designed to determine whether the long-term improvements seen in thousands of real-world patients can be reproduced under controlled conditions—and which chronic-pain patients are most likely to benefit.
