Association of cannabis use disorder with clinical outcomes following lumbar spinal fusion: a propensity-matched cohort study.

Cave S.; Ferdon R.; Shissias C.; Ravinsky R.; Reitman C.; Silvestre J. · European journal of orthopaedic surgery & traumatology : orthopedie traumatologie · 2026

**Background & Methods** This retrospective, propensity-matched cohort study examined 1,266 lumbar spinal fusion patients with cannabis use disorder (CUD) matched 1:1 to 1,266 controls without CUD across a multi-institutional electronic medical record database. Patients were stratified by documented preoperative CUD diagnosis (ICD-10-CM F12) within a 1-year preoperative window, with matching performed on 37 demographic and clinical covariates. **Key Findings** - Patients with CUD demonstrated no consistent increase in coded medical complications at 30 and 90 days post-operatively after multiple comparison correction, with similar rates of pseudarthrosis, subsequent lumbar procedures, and reoperation at 2- and 5-year follow-up compared to matched controls. - CUD was significantly associated with higher postoperative opioid prescription orders and inpatient readmission rates in both single- and multi-level fusion cohorts; emergency department utilization was elevated only in single-level cases, and behavioral/psychological service utilization was higher in multi-level cases at 90 days. - No long-term outcome differences remained statistically significant after correction for multiple comparisons, suggesting that observed associations may reflect psychosocial complexity and healthcare-access factors rather than direct pharmacologic effects of cannabis. **Dosage & Administration** Not reported. **Safety & Adverse Effects** The study did not identify a consistent increase in coded perioperative medical complications in CUD patients at 30 and 90 days. However, documented concerns included higher opioid prescription orders and increased inpatient readmission rates. Increased healthcare utilization patterns (emergency department visits and behavioral health services) were observed but not consistently across both cohorts. **Evidence Quality** This study represents moderate-quality evidence with significant limitations. The retrospective design and reliance on ICD-10-coded diagnoses may underestimate true CUD prevalence and miss undocumented cases. The propensity-matching methodology strengthens causal inference but cannot eliminate unmeasured confounding. The authors acknowledge that associations with opioid use and readmission likely reflect unmeasured psychosocial factors rather than direct cannabis effects. Prospective studies with detailed cannabis exposure metrics, dose quantification, and validated substance use assessments are needed to clarify mechanisms and establish causality.

More recent cannabis research