Reduction in Narcotic Use After Lumbar Decompression and Fusion in Patients With Symptomatic Lumbar Stenosis or Spondylolisthesis

STUDY DESIGN: Retrospective cohort study. OBJECTIVES: The purpose of this study is to assess change in opioid use before and after lumbar decompression and fusion surgery for patients with symptomatic lumbar stenosis or spondylolisthesis. METHODS: A large insurance database was queried for patients with symptomatic lumbar stenosis or spondylolisthesis undergoing index lumbar decompression and fusion procedures between 2007 and 2016. This database consists of 20.9 million covered lives and includes private/commercially insured and Medicare Advantage beneficiaries. Opioid use 6 months preoperatively through 2 years postoperatively was assessed. RESULTS: The study included 13 257 patients that underwent 1-, 2-, or 3-level posterior lumbar instrumented fusion. Overall, 57.8% of patients used opioids preoperatively. Throughout the 6-month preoperative period, 2 368 008 opioid pills were billed for (51.6 opioid pills/opioid user/month). When compared with preoperative opioid use, patients billed fewer opioid medications in the 2-year period postoperatively: 33.6 pills/patient/month (8 851 616 total pills). In a multivariate logistic regression analysis, obesity (odds ratio [OR] 1.10, 95% CI 1.004-1.212), preoperative narcotic use (OR 3.43, 95% CI 3.179-3.708), length of hospital stay (OR 1.02, 95% CI 1.010-1.021), and receiving treatment in the South (OR 1.18, 95% CI 1.074-1.287) or West (OR 1.26, 95% CI 1.095-1.452) were independently associated with prolonged postoperative (>1 year) opioid use. Additionally, males (OR 0.87, 95% CI 0.808-0.945) were less likely to use long-term opioid therapy. CONCLUSIONS: This study demonstrates that reduction in opioid use was observed postoperatively in comparison with preoperative values in patients with symptomatic lumbar stenosis or spondylolisthesis that underwent lumbar decompression with fusion. Further prospective studies that are more methodologically stringent are needed to corroborate our findings.