Introduction: Although lumbar discectomy for radiculopathy effectively reduces pain and disability for most patients, some report continued pain following surgery. Information to predict patient outcomes following discectomy could assist surgeons with patient selection. This study aimed to i) describe the perioperative trajectories of leg pain and overall clinical outcome following lumbar disc surgery for radiculopathy, ii) identify the preoperative prognostic factors that predict trajectories representing poor clinical outcomes, and iii) develop and internally validate multivariable prognostic models.
Methods: This cohort study included patients enrolled in the Canadian Spine Outcomes and Research Network diagnosed with lumbar disc pathology and radiculopathy who underwent lumbar discectomy at one of 18 spine centers. Potential outcome predictors included preoperative demographic, health-related, and clinical prognostic factors. Clinical outcomes were univariable latent-class trajectories of leg pain intensity (numeric pain rating scale) and overall outcomes modelled with multivariable trajectories of leg and back pain intensity and pain-related disability (Oswestry index). Multivariable risk model performance and internal validity were evaluated with discrimination and calibration statistics based on bootstrap shrinkage with 500 resamplings.
Results: We included data from 1,142 patients (47.6% female) operated on by one of 66 surgeons. The trajectory models identified 3 subgroups, with 11.4% of patients in the leg pain model and 28.2% in the overall outcome model experiencing a poor clinical outcome. Eleven demographic, health, and clinical factors predicted patients’ leg pain and overall outcomes. The performance of the leg pain risk model was inadequate. The overall outcome model had acceptable discrimination, calibration, and evidence of internal validity in predicting patients at risk of experiencing a poor outcome following discectomy.
Discussion: Patients experienced heterogeneous outcomes following lumbar discectomy that were associated with numerous preoperative prognostic factors. A multivariable risk model adequately predicted the overall outcomes experienced by patients. This tool can assist with patient selection for lumbar discectomy but requires additional replication and validation before confident clinical implementation.