Prognostic Factors and Models predicting Unfavorable Outcome After lumbar Discectomy: A Systematic Review.

Abstract

This review aimed to identify 1) preoperative prognostic factors and 2) prognostic models predicting unfavorable outcomes one year post lumbar discectomy. We searched Cochrane, Embase, Medline, and CINAHL (inception-April 24, 2025) for prospective cohort studies of adults with MRI-confirmed lumbar disc herniation undergoing elective discectomy. We searched for preoperative prognostic factors and (developed/validated) prognostic models. Eligible studies reported predefined patient-reported outcomes (e.g., disability, pain, quality of life) at one year. Data extraction (CHARMS), risk of bias (QUIPS/PROBAST), and certainty of evidence (GRADE) were performed by 2 independent reviewers. We reported according to the TRIPOD guidelines and summarized the results narratively due to interstudy heterogeneity. We included 15 studies comprising 9 prognostic factor studies and 6 prognostic model studies. The Oswestry Disability Index was the most common outcome measure in the reviewed studies. Moderate-certainty evidence linked several prognostic factors with post operative disability: older age, female sex, smoking, high BMI, comorbidities, and psychological distress. Studies predicting outcomes like pain, health related quality of life or satisfaction provided less robust findings. Only one machine-learning-based model study demonstrated robust internal-external validation with good performance. Several preoperative demographic and clinical factors are associated with higher disability after lumbar discectomy. Prognostic models show potential, but most lack external validation and carry high risk of bias, limiting clinical applicability.