Portenoy 2012.

Study characteristics
Methods	Purpose of the study: to reduce moderate‐to‐severe cancer pain despite stable opioid regimen Study setting: USA, number of study centres not reported Study period: not reported Study design: randomised, double‐blind, placebo‐controlled, parallel group, graded‐dose design Study duration: 5‐ to 14‐day baseline period, a 5‐week double‐blind titration and treatment period, and a poststudy visit after 2 weeks. Maximum duration 9 weeks
Participants	Inclusion criteria: active cancer and chronic pain that was moderate or severe despite a stable opioid regimen that could not be made more effective by further opioid dose titration Type of cancer: breast, gastrointestinal, lung, prostate, other Exclusion criteria: receiving long‐term methadone therapy for pain, major psychiatric or cardiovascular disorder, epilepsy, or significant renal or hepatic impairment, pregnancy, lactating or not using adequate contraception, had received or were to receive radiotherapy, chemotherapy or hormonal therapy, usage of marijuana, CB‐based medications or rimonabant within 30 days of study entry unwilling to abstain for the duration of the study Oromucosal spray; THC:CBD extract: THC 2.7 mg and CBD 2.5 mg Low‐dose group: 91 participants; 49.4% men; mean age 59 (SD 12.3) years; Caucasian 73.6%, Black 12.1%, Hispanic 11.0%, Asian 0%, other 3.3%; type of cancer pain: neuropathic (8.8%), somatic (1.1%), visceral (22.0%), mixed (46.2%), bone (22.0%), other (0%); mean pain 5.8 (SD 1.3); daily morphine equivalent maintenance 120 mg/day; daily breakthrough morphine equivalent not reported; previous cannabis use: 12.1% Medium‐dose group: 88 participants; 55.7% men; mean age 59 (SD 13,1) years; Caucasian 84.1%, Black 6.8%, Hispanic 8.0%, Asian 1.1%, other 0%. Type of cancer pain: neuropathic (13.6%), somatic (14.8%), visceral (12.5%), mixed (42.0%), bone (17.0%), other (0%); mean pain 5.8 (SD 1.2); daily morphine equivalent maintenance 120 mg/day; daily breakthrough morphine equivalent not reported; previous cannabis use 12.5% High‐dose group: 90 participants; 53.3% men; mean age 58 (SD 11.2) years; Caucasian 75.6%, Black 11.1%, Hispanic 7.8%, Asian 1.1%, Other 4.4%; type of cancer pain: neuropathic (7.8%), somatic (7.8%), visceral (11.1%), mixed (35.6%), bone (37.8%), other (0%); mean pain 5.8 (SD 1.2); daily morphine equivalent maintenance 180 mg/day; daily breakthrough morphine equivalent not reported; previous cannabis use: 11.1% Placebo: 91 participants, 48.3% men; mean age 56 (SD 12.2) years; Caucasian 75.8%, Black 6.6%, Hispanic 13.2%, Asian 0%, Other 4.4%; type of cancer pain: neuropathic (12.1%), somatic (12.1%), visceral (14.3%), mixed (42.9%), bone (18.7%), other (0%); mean pain 5.7 (SD 1.2); daily morphine equivalent maintenance 120 mg/day; daily breakthrough morphine equivalent not reported; previous cannabis use: 6.6%
Interventions	THC/CBD extract 2.7 mg/2.5 mg, oromucosal spray, flexible dosage:low‐dose group: 1–4 actuations/day; medium‐dose group: 6–10 actuations/day, high‐dose group: 11–16 actuations/day Placebo, oromucosal spray: 1–16 actuations/day; flexible dosage Rescue medication: not reported Allowed cotherapies: all opioids typically used for severe cancer pain except for methadone were allowed.
Outcomes	Proportion of participants reporting no worse than mild pain on treatment at 14 days after start of treatment (typically < 30/100 mm on a 100‐mm VAS or < 3 on an 11‐point NRS): not assessed Patient impression to be much or very much improved: PGIC Withdrawal due to adverse events: adverse events as reported by the participant Combined responder: not assessed Pain relief ≥ 30%: BPI‐SF NRS 0–10, daily Pain relief ≥ 50%: BPI‐SF NRS 0–10, daily. Calculated by imputation method. Mean pain intensity: BPI‐SF NRS 0–10, daily Sleep problems: Sleep Disruption Score 0–10, last 24 hours Depression: Montgomery Åsberg Depression Rating Scale Anxiety: not assessed Maintenance opioid therapy dose: assessed, but not reported in detail. Data not suited for meta‐analysis Breakthrough opioid therapy dose: assessed, but not reported in detail. Data not suited for meta‐analysis Withdrawals due to lack of efficacy: not reported Nervous system disorders adverse effects: participant‐reported adverse events Psychiatric disorders adverse effects: participant‐reported adverse events Any serious adverse event: participant‐reported serious adverse events
Notes	Funding: in part by the Huntsman Cancer Foundation (S.W.). GW Pharmaceuticals produces nabiximols, which is licenced in Canada as an adjunctive analgesic treatment in adults with advanced cancer. GW Pharmaceuticals and Otsuka funded the study. Conflicts of interest: not declared
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly assigned by computer using a block approach.
Allocation concealment (selection bias)	Unclear risk	No details provided.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Placebo was similar colour, no adjustment for taste described.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No details provided.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	ITT analysis by last observation carried forward.
Selective reporting (reporting bias)	Low risk	NCT00530764.
Selection bias	High risk	Number of participants with previous cannabis use double in all nabiximols groups as in placebo groups.