One-year outcomes of surgical versus nonsurgical treatments for discogenic back pain: a community-based prospective cohort study

Background context: The clinical entity "discogenic back pain" remains controversial at fundamental levels, including its pathophysiology, diagnostic criteria, and optimal treatment. This is true despite availability of four randomized trials comparing the efficacy of surgical and nonsurgical treatments. One trial showed benefit for lumbar fusion compared with unstructured nonoperative care, and three others showed roughly similar results for lumbar surgery and structured rehabilitation.

Purpose: To compare outcomes of community-based surgical and nonsurgical treatments for patients with chronic back pain attributed to degeneration at one or two lumbar disc levels.

Design: Prospective observational cohort study.

Patient sample: Patients presenting with axial back pain to academic and private practice orthopedic surgeons and neurosurgeons in a large metropolitan area.

Outcome measures: Roland-Morris back disability score (primary outcome), current rating of overall pain severity on a numerical scale, back and leg pain bothersomeness measures, the physical function scale of the short-form 36 version 2 questionnaire, use of medications for pain, work status, emergency department visits, hospitalizations, and further surgery.

Methods: Patients receiving spine surgery within 6 months of enrollment were designated as the "surgical treatment" group and the remainder as "nonsurgical treatment." Outcomes were assessed at 3, 6, 9, and 12 months after enrollment.

Results: We enrolled 495 patients with discogenic back pain presenting for initial surgical consultation in offices of 16 surgeons. Eighty-six patients (17%) had surgery within 6 months of enrollment. Surgery consisted of instrumented fusion (79%), disc replacement (12%), laminectomy, or discectomy (9%). Surgical patients reported more severe pain and physical disability at baseline and were more likely to have had prior surgery. Adjusting for baseline differences among groups, surgery showed a limited benefit over nonsurgical treatment of 5.4 points on the modified (23-point) Roland disability questionnaire (primary outcome) 1 year after enrollment. Using a composite definition of success incorporating 30% improvement in the Roland score, 30% improvement in pain, no opioid pain medication use, and working (if relevant), the 1-year success rate was 33% for surgery and 15% for nonsurgical treatment. The rate of reoperation was 11% in the surgical group; the rate of surgery after treatment designation in the nonsurgical group was 6% at 12 months after enrollment.

Conclusions: The surgical group showed greater improvement at 1 year compared with the nonsurgical group, although the composite success rate for both treatment groups was only fair. The results should be interpreted cautiously because outcomes are short term, and treatment was not randomly assigned. Only 5% of nonsurgical patients received cognitive behavior therapy. Nonsurgical treatment that patients received was variable and mostly not compliant with major guidelines.