Table 6.

Summary of findings and GRADE recommendations

OUTCOMESCOMPARATORRELATIVE EFFECT (95% CI)CERTAINTY OF EVIDENCE (GRADE)
Pain
  • ≥ 30% reduction of painPlaceboRR = 1.37 (1.14 to 1.64)
  • Overall: Very low owing to serious risk of bias, serious inconsistency, serious indirectness, and serious imprecision

  • Smoked medical marijuana: Very low owing to serious risk of bias, serious inconsistency, serious indirectness, and serious imprecision

  • Buccal cannabinoids: Very low owing to serious risk of bias, serious inconsistency, and serious indirectness

  • First or second line for pain: Very low owing to serious risk of bias, serious inconsistency, serious indirectness, and serious imprecision

  • Third line for pain: Very low owing to serious risk of bias, serious inconsistency, and serious indirectness

  • Change in pain scale2PlaceboWMD = 0.5 (0.11 to 0.80)*Overall: Very low owing to serious risk of bias, serious inconsistency, serious indirectness, and serious imprecision
Nausea and vomiting
  • Control of nausea and vomitingPlaceboRR = 3.60 (2.55 to 5.09)Moderate owing to serious risk of bias and serious imprecision, but magnitude had large effect
AntiemeticRR = 1.85 (1.18 to 2.91)Low owing to serious risk of bias and serious inconsistency
  • Patient preference25PlaceboRR = 4.82 (1.74 to 13.36)Low owing to serious risk of bias, serious inconsistency, and serious imprecision, but magnitude had large effect
AntiemeticRR = 2.76 (1.88 to 4.03)Low owing to serious risk of bias, serious inconsistency, and serious imprecision, but magnitude had large effect
Spasticity
  • ≥ 30% improvement in spasticity28PlaceboRR = 1.37 (1.07 to 1.76)Low owing to serious risk of bias and serious publication bias
  • Change in spasticity scale2,28WMD = 0.31 or 0.76Very low owing to serious risk of bias, serious inconsistency, and serious imprecision
  • Global impression of changeRR = 1.45 (1.08 to 1.95)Low owing to serious risk of bias and serious inconsistency
Adverse events
  • Withdrawal owing to adverse eventsPlaceboNNH = 8 to 22High owing to serious risk of bias and serious imprecision, but magnitude had large effect and plausible confounding had large effect
  • GRADE—Grading of Recommendations Assessment, Development and Evaluation, NNH—number needed to harm, RR—risk ratio, WMD—weighted mean difference.

  • * Whiting et al2 article selected because it was more general pain rather than multiple sclerosis pain as examined in the study by Iskedjian et al.18

  • Patient preference is inconsistent with effect on nausea and vomiting and therefore might reflect more than control of nausea and vomiting but also the euphoria or “high” received from cannabinoids.