Can Medical Cannabis Reduce Opioid Use? What the Evidence Can Show

Medical cannabis collection · Sources reviewed 11 October 2026

Using fewer opioid tablets sounds like a straightforward outcome, but it does not automatically mean better pain control, fewer harms or less overall medication. The useful question is whether adding or substituting a specified cannabis medicine makes a person’s treatment safer and more effective. Research has not yet supplied a dependable general answer.

Why some trials cannot answer the tapering question

A 2021 systematic review included five randomised trials, all in cancer pain, alongside observational studies. The randomised trials asked participants to maintain their opioid doses. That makes their results indirect evidence for a question about reducing opioids: the study instructions restricted the very change being investigated. The review rated the evidence about opioid reduction as very uncertain. Noori and colleagues, 2021.

The observational studies suggested reductions, but people who choose cannabis treatment can differ from people who do not. Their clinicians, other treatments and reasons for changing opioids can also differ. The contrast is therefore not simply that trials are negative and everyday experience is positive. The designs are often asking different questions with different vulnerabilities.

What the newer cancer-pain review found

A 2026 systematic review included 15 studies and pooled ten. Controlled comparisons did not establish significant reductions in total, maintenance or breakthrough opioid use. Some uncontrolled before-and-after analyses suggested reductions with particular formulations, but these depended on limited studies. The overall certainty was low. Creangă-Murariu and colleagues, 2026.

“should not be considered a dependable opioid-sparing strategy”

Cancer-pain systematic review, 2026

This conclusion concerns cannabinoids as a means of reducing opioid use in cancer pain; it does not answer every question about symptom relief.

A direct trial found fewer tablets, with important trade-offs

The 2024 SPIRAL randomised trial reported on 81 people with fibromyalgia treated with oxycodone, inhaled cannabis or their combination for six weeks. Treatment was open-label, meaning participants knew what they received. The combination group consumed 35% fewer oxycodone tablets than the oxycodone-only group, but its overall drug load was highest. SPIRAL trial results, 2024.

Adverse effects led to early withdrawal in 31% of participants allocated to the cannabis-containing groups, compared with 13% in the oxycodone group. Among those completing treatment, the primary composite adverse-effect score did not differ between groups. A reduction in one medicine therefore coexisted with substantial tolerability problems and greater combined treatment exposure.

This is a useful positive signal about one outcome. It is not proof that combining the treatments reduces overall harm, nor a template for unsupervised substitution.

Why a large observational reduction is still uncertain

A 2025 study followed separate pain-clinic cohorts receiving cannabis with opioids or opioids alone. It reported marked reductions in opioid use in the cannabis cohort. However, treatment was not randomised, the cohorts came from different clinics, and 46 of 102 people in the cannabis cohort dropped out versus one of 53 controls. The one-year observational study.

Such unequal loss matters: the eventual result may disproportionately describe people who found treatment acceptable or remained engaged with that clinic. An encouraging association warrants investigation, while leaving uncertainty about what would happen in a comparable group offered the alternative.

Similar average pain effects do not establish safe substitution

A 2024 network meta-analysis compared the chronic-pain evidence for cannabis medicines and opioids. It suggested similar average benefits for some outcomes, with fewer discontinuations due to adverse effects for cannabis. Much of the comparison was indirect, through other treatments or placebo. It was not a trial of a monitored opioid-to-cannabis switch. Jeddi and colleagues, 2024.

Average results across separate trial populations cannot determine the best transition for a particular patient. They also leave questions about longer-term outcomes and people with complex medical conditions.

What better evidence would look like

People already taking opioids should discuss any proposed reduction with their prescriber. Abrupt stopping is not an appropriate way to test whether cannabis works. Physical dependence during prescribed treatment also needs to be distinguished from an opioid use disorder; reducing a prescription and treating an addiction are not interchangeable research questions.

Educational information, not an individual prescribing plan. Decisions about medicines belong with the treating clinician. External reference links open in a new tab.