Poster Presentation Sydney Spinal Symposium 2023

The comparative effectiveness of lumbar fusion surgery and spinal decompression surgery for lumbar spinal stenosis: protocol for a target trial emulation   (#23)

Harrison J Hansford 1 2 , Margreth Grotle 3 , James H McAuley 1 2 , Matthew D Jones 1 2 , Aidan G Cashin 1 2
  1. Centre for Pain IMPACT, Neuroscience Research Australia, Sydney, Australia, Sydney, Australia
  2. UNSW, Sydney, NSW, Australia
  3. Center for Intelligent Musculoskeletal Health, Department of Rehabilitation Science and Health Technology, Oslo Metropolitan University, P.O. Box 4, St. Olavs plass, NO-0130, Oslo, Norway

Aims: To emulate a pragmatic trial comparing spinal decompression surgery to additional spinal fusion surgery for people with spinal stenosis. We will also seek to identify whether treatment effects from spinal fusion surgery compared to decompression alone are superior within selected subgroups within this target trial emulation.

Methods: Data from the Norwegian registry for spine surgery (NORSpine) will be used to emulate a pragmatic randomised trial comparing spinal decompression surgery to additional spinal fusion surgery for people with spinal stenosis with spondylolisthesis. Participants will be adults diagnosed with lumbar spinal stenosis with spondylolisthesis of ≥3mm, reporting continuous pain in the back/legs for ≥3months, and a pain intensity ≥1 in the last week. Participants will be excluded if they have received diagnoses of cauda equina syndrome, isthmic spondylolisthesis, scoliosis/kyphosis, and previous surgery at the same level. We will use inverse probability weighting to emulate randomisation, with logistic regression including many clinical and sociodemographic characteristics. Our primary outcome will be the Oswestry disability index questionnaire v2. We will emulate the intention to treat and per-protocol effects.

Subgroups: We will conduct six subgroup analyses, selected for whom fusion surgery may be superior to decompression. These include patients with back pain intensity ≥5/10 compared to patients with back pain intensity <5/10; Patients with a BMI ≥30 compared to patients with a BMI <30; Patients who are female compared to patients who are male; Older patients (>65 years old) compared to younger patients (<65 years old); Patients who have an ASA grade ≥3 compared to patients with an ASA grade <3; and patients who have low anxiety and depression (i.e., less than moderately anxious/depressed) compared to patients with moderate to high anxiety and depression (i.e., moderately anxious/depression or more).

Sensitivity Analysis: to assess the robustness of our findings to potential violations of assumptions, we will use alternative modelling approaches to emulate randomisation and calculate e-values to identify the extent of residual confounding that would have to be present in our analysis to 'explain away' our findings.

Note: The aim is to discuss with the spinal health community whether other subgroups may be of interest