Seven RCTs (298 unique patients: 222 treated with cannabis preparations (many of whom also crossed over to placebo) and 76 treated with placebo alone) were included. Four of these were crossover trials.
The quality of all 7 included studies was rated as good. Five studies scored 5, one scored 4 and one scored 3. Both the chi-squared and I-squared statistical tests indicated that there was no significant statistical heterogeneity between the studies. The Begg-Mazumdar test showed no evidence of publication bias.
There were no statistical differences at baseline between pain scores of patients treated with cannabinoids and those receiving placebo.
Cannabis preparations were more effective in reducing pain scores than placebo for the treatment of MS-related or neuropathic pain, with a difference in effect size of 0.8 points; the difference was statistically significant (p=0.029). The difference from baseline was 1.7 (SE 0.7, p=0.018) for cannabidiol/THC buccal spray (6 studies, n=196), 1.5 (SE 0.7, p=0.044) for cannabidiol alone (5 studies, n=41), 1.5 (SE 0.6, p=0.013) for dronabinol (3 studies, n=91) and 1.6 (SE 0.4, p<0.001) for all cannabinoids (14 trial arms, n=328).
Data for placebo groups (10 studies, n=250) reported an average reduction in pain scores of 0.8 points, which was statistically significant (p=0.023). A post hoc analysis removing 2 studies that allowed free use of ‘rescue’ medications, including analgesics, lowered the placebo effect to 0.6 points, which was not statistically significant.
Dizziness was the most commonly reported adverse event for all cannabinoid treatments (32.5% +/- 16.4) and for placebo (10.1% +/- 3.8). In studies of cannabidiol/THC buccal spray, 39% (+/- 16.3) reported dizziness.
Withdrawals due to adverse events were similar between patients receiving cannabis (5.5%, 14 out of 255) and those treated with placebo (5.1%, 13 out of 253).