Predicting recovery after lumbar spinal stenosis surgery: a historical cohort study using data from the Canadian Spine Outcomes Research Network (CSORN) — The International Society for the Study of the Lumbar Spine
  1. Schedule
  2. Special Poster Session 1B. SURGERY AND CLINICAL OUTCOME
  3. Predicting recovery after lumbar spinal stenosis surgery: a historical cohort study using data from the Canadian Spine Outcomes Research Network (CSORN)

Predicting recovery after lumbar spinal stenosis surgery: a historical cohort study using data from the Canadian Spine Outcomes Research Network (CSORN) (#110)

Luciana Macedo 1 , Fatemeh Jazinizadeh 2 , Lisa Carlesso 1 , Janie Astephen Wilson 3 , Douglas Gross 4 , Charles Fisher 5 , Hamilton Hall 6 , Neil Manson 7 , Ken Thomas 8 , Greg McIntosh 9 , Brian Drew 10 , Y. Raja Rampersaud 6
  1. School of Rehabilitation Science, McMaster University, Hamilton, ON, Canada
  2. School of Biomedical Engineering, McMaster University, Hamilton, ON, Canada
  3. Dalhousie University, Halifax, NS, Canada
  4. Department of Physical Therapy, University of Alberta, Edmonton, AB, Canada
  5. Vancouver General Hospital, University of British Columbia, Vancouver, BC, Canada
  6. Department of Surgery, University of Toronto, Toronto, ON, Canada
  7. Canada East Spine Centre, Saint John, NB, Canada
  8. Cumming School of Medicine, University of Calgary, Calgary, AB, Canada
  9. Canadian Spine Society, Toronto, ON, Canada
  10. Department of Surgery, McMaster University, Hamilton, ON, Canada

Introduction: Decompressive surgery, with or without spinal fusion, is recommended for persons with symptomatic lumbar spinal stenosis (SLSS) symptoms for whom conservative management has failed. However, significant persistent pain, functional limitations, and narcotic use can affect up to one third of patients post-surgery. The aim of this study was to identify predictors of outcomes 1-year post SLSS surgery with a focus on modifiable predictors.

Methods:  The Canadian Spine Outcomes Research Network (CSORN) is a database of prospectively collected data on pre- and postsurgical outcomes among surgical patients. We included participants with a diagnosis of SLSS undergoing their first spine surgery. The primary outcomes collected 12 months after surgery included back and leg pain, disability (Oswestry Disability Index (ODI)), walking capacity (ODI item 4) and quality of life ( EQ-5D-5L). Recovery was identified using a composite score of leg and back pain (NRS 0-3) and ODI less than 30. Predictors included demographics, physical activity level, smoking status, previous rehabilitation, medication intake, depression, expectations, and number of comorbidities. A multivariate partial least squares (PLS) model was used to identify predictors of outcomes. One of the strengths of the PLS is that it allows for the inclusion of multiple outcomes in one model. Analysis was conducted separately for men and women.   

Results:  Patient data collected between January 2015 to September 2019 were included. A total of 1868 participants were eligible for inclusion, however; after excluding participants with missing data for the outcomes or 20% of the predictors (less than 7 out of 27), our final sample size included 1068 participants (654 men and 414 women) after multiple imputation for missing predictors. Of the total sample, 466 participants (44%) (311 men and 155 women) were identified as recovered (a per our composite score) at 1-year follow-up. The percentage of variance in the outcomes explained by the PLS model ranged between 0.16-0.24 with the lowest value corresponding to the women in the non-imputed dataset and the highest being related to the men in the imputed dataset. The variance importance in projection (VIP) scores for men identified higher depression (PHQ9), higher number of comorbidities and high disability (ODI) at baseline as important predictors of worse outcomes in the PLS models. The variables that predicted worse outcomes for women were higher BMI, higher depression, and higher disability baseline. Results without imputation included the same predictors with the addition of BMI for men, symptom duration and previous treatment with a physical trainer for women. For both men and women in imputed and non-imputed data ODI at baseline (with the VIP score of more than 3) was the most significant contributor in the PLS model to predict all outcomes.

Discussion:  The results demonstrate that higher depression and disability at baseline are associated with worst outcomes post-surgery for men and women. Higher number of comorbidities for men and high BMI for women were also associated with worse outcomes. The majority of predictors are modifiable so future studies should evaluate whether modifying these parameters prior to surgery can improve outcomes.

#ISSLS2022