Ten-year outcomes of medical cannabis for chronic low back pain: opioid reduction, pain relief, and functional improvement in 1,000 patients

Abstract Purpose To evaluate 10-year outcomes of medical cannabis therapy on opioid use, pain intensity, functional disability, concomitant medication use, and adverse events in cannabis-naïve chronic low back pain (CLBP) patients, with particular attention to clinically meaningful outcomes defined by established minimal clinically important difference (MCID) thresholds. Methods Single-center longitudinal observational study of 1,000 consecutive cannabis-naïve CLBP patients from a registry database (2015–2024), reported following STROBE guidelines. Primary outcome: morphine milligram equivalents (MMEQ). Secondary outcomes: Numeric Rating Scale (NRS) for pain intensity, Oswestry Disability Index (ODI) for functional disability, cannabis dosing patterns, adverse events, and concomitant medication use. Pre-specified MCID thresholds: ≥50% MMEQ reduction, ≥ 30% NRS reduction, ≥ 10-point ODI improvement. Statistical analyses included paired t-tests for completers and linear mixed-effects models for sensitivity analyses. Results Of 1,000 enrolled patients, 638 (63.8%) completed 10-year follow-up. Among completers, MMEQ decreased from 62.8 ± 35.7 to 6.4 ± 7.1 mg/day (− 89.8%, p < 0.001); NRS from 8.71 ± 1.23 to 1.37 ± 1.71 (− 84.2%, p < 0.001); ODI from 52.9 ± 11.9% to 36.8 ± 14.9% (16.1-point reduction, − 30.4%, p < 0.001). MCID responders: 91.2% for MMEQ, 96.6% for NRS, 62.1% for ODI. Substantial polypharmacy reductions occurred: tramadol/tapentadol − 84.0% points (pp), benzodiazepines − 73.5 pp, SSRIs − 71.9 pp, gabapentinoids − 30.7 pp. True tolerability adverse events occurred in 11.4% of visits; serious psychiatric events in 0.02%. Conclusions In this uncontrolled observational study, medical cannabis therapy was associated with reductions in opioid use, pain intensity, and functional disability over 10 years, accompanied by polypharmacy reduction and acceptable tolerability. Effect sizes substantially exceeded RCT benchmarks (e.g., − 0.6 NRS difference vs. placebo in a recent phase 3 trial), suggesting observational biases contribute to these findings. These hypothesis-generating data warrant validation in randomized controlled trials.