Systematically Testing the Evidence on Marijuana

Search: 

Screening:

Study Selection Criteria

Abbreviations. KQ: Key Question; PICOTS: population, interventions, comparators, outcomes, timing, setting; RCT: randomized controlled trial

Data synthesis: 

• Random effects meta-analyses were performed when there was sufficient homogeneity among studies.
• Risk of bias assessment: 
  • Cochrane Back Review Group methods for RCTs
  • US Preventive Services Task Force methods for observational studies

Assessing the strength (certainty) of the body of evidence: 

• AHRQ Methods Guide approach for all primary comparisons and outcomes. 
• Of note, the AHRQ methods are very similar to GRADE with the following exceptions: the term strength of evidence is used rather than certainty of evidence, and the term “insufficient” evidence is used in place of the term “very low” strength (certainty) evidence.

Living systematic review: 

• Abridged surveillance reports are done quarterly, with availability of new studies that are likely to significantly impact findings triggering an updated report.

More detailed information about systematic review methods is available here.

(Searched as of April 2025)

The tables below summarize the evidence for cannabinoids for chronic pain from 29 trials. For the full results see link below.

• Treatments were grouped according to the ratio of THC to CBD and to the source of the cannabinoids (synthetic or extracted from whole plant)
• Effect sizes are classified as no effect, small, moderate, or large according to methods used in other systematic reviews of chronic pain.
• Strength of evidence is indicated by plus signs, except where the effect is unclear.

Definitions of effect sizes

Abbreviations: MD = mean difference; OR = odds ratio; RR = relative risk; SMD = standardized mean difference.

Example interpretation: 

For the outcome of pain severity: From 8 trials, a synthetic high THC oral treatment had a small effect on the severity of pain in adults with chronic pain, and the confidence in this finding is low.

Benefits of cannabinoids for chronic pain compared with placebo in the short term (4 weeks to < 6 months)

Abbreviations: CBD = cannabidiol; CBDV = cannabidivarin; CBN = cannabinol; SOE = strength of evidence; THC = tetrahydrocannabinol.
^a  ≥30% improvement from baseline.
^b Potential effect: SOE of low or higher; findings indicate at least a small magnitude of effect but not statistically significant.
^cThe pooled difference of -0.42 (95% CI -0.73 to -0.16) was just below the threshold for a small effect
^dComparison was “usual care.”
Effect size: none (i.e., no effect/no statistically significant effect), small, moderate, or large increased risk; SOE: [+] = low, [++] = moderate, [+++] = high.

Harms of cannabinoids for chronic pain compared with placebo in the short term (4 weeks to < 6 months)

Abbreviations: CBD = cannabidiol; CBDV = cannabidivarin; SAE = serious adverse event; SOE = strength of evidence; THC = tetrahydrocannabinol; WAE = withdrawal due to adverse event.
^a potential effect: SOE of low or higher; findings indicate at least a small magnitude of effect but not statistically significant.
^b Comparison was “usual care.”
Effect size: none (i.e., no effect/no statistically significant effect), small, moderate, or large increased risk; SOE: [+] = low, [++] = moderate, [+++] = high.

• An oromucosal spray with comparable amounts of THC and CBD probably reduces pain (but does not improve function)to a small extent in adult patients with neuropathic pain.
• Oral CBD alone (without THC; synthetic or purified) probably does not affect pain or function.
• The strength of evidence for all other benefits is low or insufficient.
• Compared with placebo, cannabis-related interventions resulted in greater risk of common, but not severe adverse events (e.g., dizziness, nausea, sedation), and withdrawals from studies due to adverse events.
• No studies were included that examined the evidence among children or adolescents or on subacute pain.

• Whole plant formulations, high THC:CBD, and extracted or nonoral low-THC:CBD products remain understudied.
• No studies examined effects beyond 6 months.
• There was relatively little evidence examining the harms of many preparations.