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

Khatib M.; Robinson D.; Lavon E.; Qawasmi F.; Abu Rashed W.; Murad H.; Yassin M. · European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society · 2026

**Background & Methods** This was a single-center longitudinal observational study of 1,000 cannabis-naïve patients with chronic low back pain (CLBP) enrolled from a registry database between 2015–2024, with 10-year follow-up assessment. The study evaluated medical cannabis therapy's effects on opioid use (morphine milligram equivalents), pain intensity (Numeric Rating Scale), functional disability (Oswestry Disability Index), concomitant medication use, and adverse events, with pre-specified minimal clinically important difference (MCID) thresholds to define clinically meaningful response. **Key Findings** - **Opioid Reduction:** Among 638 completers (63.8% retention), morphine milligram equivalents decreased 89.8% from 62.8±35.7 mg/day to 6.4±7.1 mg/day (p<0.001), with 91.2% of patients achieving ≥50% MMEQ reduction (MCID threshold). - **Pain and Function Improvement:** Pain intensity (NRS) decreased 84.2% from 8.71±1.23 to 1.37±1.71 (p<0.001), with 96.6% achieving ≥30% reduction; functional disability (ODI) improved by 16.1 points (−30.4%), exceeding MCID in 62.1% of responders. - **Polypharmacy Reduction:** Substantial decreases in concomitant medications occurred: tramadol/tapentadol (−84.0 percentage points), benzodiazepines (−73.5 pp), SSRIs (−71.9 pp), and gabapentinoids (−30.7 pp). **Dosage & Administration** Not reported. **Safety & Adverse Effects** Tolerability adverse events occurred in 11.4% of visits, with serious psychiatric events in only 0.02% of encounters, indicating an acceptable safety profile over the 10-year observation period. **Evidence Quality** This uncontrolled observational study has significant limitations that substantially reduce evidence quality. The 36.2% loss to follow-up introduces potential selection bias favoring treatment responders. The absence of a control group, randomization, or blinding prevents causal attribution; the authors acknowledge that observed effect sizes (e.g., −84.2% NRS reduction) substantially exceed typical RCT benchmarks (−0.6 NRS vs. placebo) and likely reflect observational bias rather than true drug effect. The study is explicitly hypothesis-generating and requires validation in randomized controlled trials. Level of evidence: IV (uncontrolled observational study).