Effects of cannabinoid administration for pain: A meta-analysis and meta-regression

Abstract

Chronic pain states have resulted in an overreliance on opioid pain relievers, which can carry significant risks when used long term. As such, alternative pain treatments are increasingly desired. Although emerging research suggests that cannabinoids have therapeutic potential regarding pain, results from studies across pain populations have been inconsistent. To provide meta-analytic clarification regarding cannabis's impact on subjective pain, we identified studies that assessed drug-induced pain modulations under cannabinoid and corresponding placebo conditions. A literature search yielded 25 peer-reviewed records that underwent data extraction. Baseline and end-point data were used to compute standardized effect size estimates (Cohen's d) across cannabinoid administrations (k = 39) and placebo administrations (k = 26). Standardized effects were inverse-variance weighted and pooled across studies for meta-analytic comparison. Results revealed that cannabinoid administration produced a medium-to-large effect across included studies, Cohen's d = -0.58, 95% confidence interval (CI) [-0.74, -0.43], while placebo administration produced a small-to-medium effect, Cohen's d = -0.39, 95% CI [-0.52, -0.26]. Meta-regression revealed that cannabinoids, β = -0.43, 95% CI [-0.62, -0.24], p < .05, synthetic cannabinoids, β = -0.39, 95% CI [-0.65, -0.14], p < .05, and sample size, β = 0.01, 95% CI [0.00, 0.01], p < .05, were associated with marked pain reduction. These outcomes suggest that cannabinoid-based pharmacotherapies may serve as effective replacement/adjunctive options regarding pain, however, additional research is warranted. Additionally, given demonstrated neurocognitive side effects associated with some constituent cannabinoids (i.e., THC), subsequent work may consider developing novel therapeutic agents that capitalize on cannabis's analgesic properties without producing adverse effects. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Figure 1.. Literature Search and Review Pipeline.

Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) pipeline diagram showing search and review results. A preliminary search produced 949 records, with an additional 5 assembled from additional resources (e.g., narrative reviews), totaling 954 records overall. During abstract review, 899 records were removed from the meta-analysis pipeline. During complete manuscript review, an additional 30 records were discarded based on study exclusion criteria. Finally, the 25 remaining records underwent data extraction and subsequent meta-analytic assessment.

Figure 2.. Pooled Cannabinoid Administration Effect.

Study-level standardized effect size estimates (Cohen’s d) were computed for each cannabinoid administration across included studies. Circle sizes are proportional to small, medium, and large effect size estimate interpretations (Cohen, 1988). Study-level estimates were inverse-variance weighted and pooled to determine a representative estimate. When considering overall pain reduction effects, cannabinoid administration was associated with a medium-to-large effect across studies, Cohen’s d = −0.58, 95% CI (−0.74, −0.43). N, sample size; ES, standardized effect size estimate; CI, confidence interval.

Figure 3.. Pooled Placebo Administration Effect.

Study-level standardized effect size estimates (Cohen’s d) were computed for each placebo administration across included studies. Circle sizes are proportional to small, medium, and large effect size estimate interpretations (Cohen, 1988). Study-level estimates were inverse-variance weighted and pooled to determine a representative estimate. When considering overall pain reduction effects, placebo administration was associated with a small-to-medium effect across studies, Cohen’s d = −0.39, 95% CI (−0.52, −0.26). N, sample size; ES, standardized effect size estimate; CI, confidence interval.

Figure 4.. Bivariate Relationship Between Effect Size Estimates and Significant Predictors.

Meta-regression results revealed that, when controlling for other explanatory variables, drug administration conditions were linked with pain reduction among included studies, such that cannabinoids (whole-plant cannabis and whole-cannabis extracts) β = −0.43, 95% CI (−0.62, −0.24), p < 0.05, and synthetic cannabinoids (Dronabinol, Nabilone, and CT3) β = −0.39, 95% CI (−0.62, −0.24), p < 0.05, performed better than placebo. Furthermore, meta-regression results showed that, when controlling for other explanatory variables, sample size was linked with pain reduction, β = 0.01, 95% CI (0.00, 0.01), p < 0.05, such that studies involving smaller samples tended to report greater pain reduction. Cannabinoids = shaded (green) circles; placebo = unshaded circles.

Figure 5.. Bivariate Relationship Between Effect Size Estimates and Sample Sex Composition (Sex Ratio).

Meta-regression results showed that, when controlling for other explanatory variables, sample sex composition was linked with a modest, albeit non-significant, effect, β = −0.64, 95% CI (−1.37, 0.09), p = 0.09, such that studies including more female participants tended to report greater pain reductions. Cannabinoids = shaded (green) circles; placebo = unshaded circles.