Systemic immune-inflammation index predicts neurological recovery in degenerative cervical myelopathy: a comparative analysis of OPLL and non-OPLL patients

Why clinicians should know about it

• Picked for Spine Surgery (top studies of the week, 8 November 2026): Postoperative recovery predictors in surgically treated cervical myelopathy

Abstract

PURPOSE: To compare systemic inflammatory characteristics between patients with ossification of the posterior longitudinal ligament (OPLL) and non-OPLL degenerative cervical myelopathy (DCM) and to evaluate the prognostic value of the systemic immune-inflammation index (SII), alone and in combination with radiological parameters, for postoperative neurological recovery. METHODS: This retrospective prognostic cohort study included 317 surgically treated patients with DCM (98 with OPLL and 219 without OPLL) with complete 2-year follow-up data. Neurological and clinical outcomes were assessed using the visual analog scale (VAS), Neck Disability Index (NDI), modified Japanese Orthopaedic Association (mJOA) score, and mJOA recovery rate; patients were categorized as having favorable recovery (mJOA recovery rate ≥ 50%; n = 231) or unfavorable recovery (< 50%; n = 86). Preoperative hematologic inflammatory indices, including the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and SII, serum biochemical profiles, surgical characteristics, and cervical radiological parameters were collected. Magnetic resonance imaging (MRI) was used to assess maximum spinal cord compression (MSCC), maximum canal compromise (MCC), transverse area (TA), and compression ratio (CR). Full multivariable logistic regression models identified factors independently associated with favorable recovery. Sensitivity analyses separately adjusted for surgical approach, surgical procedure, and the numbers of decompressed and fused levels. Receiver operating characteristic (ROC) analyses assessed discriminative performance and optimal cutoffs. Predictors that remained independently associated with favorable recovery in the full multivariable models were retained, and parsimonious logistic regression models were refitted for ROC analysis, nomogram construction, and internal validation. The combined scores represented weighted linear predictors from the corresponding refitted SII-TA models and were not predicted probabilities. Internal validation was performed using 1,000 bootstrap resamples; calibration was assessed using calibration curves and the Hosmer-Lemeshow test, and clinical net benefit was evaluated using decision curve analysis. RESULTS: Patients with OPLL had significantly higher white blood cell (WBC) counts and SII values than those without OPLL. In the overall cohort, lower SII and larger TA were independently associated with favorable neurological recovery. A combined predictive score was derived from the refitted parsimonious model as follows: Score = - 0.202 + 5.697 × TA - 0.294 × (SII/100). Score ≥ 0.50 was identified as the optimal cutoff for predicting favorable neurological recovery. ROC analysis further identified cutoff values of SII ≤ 556.5 and TA ≥ 0.405 cm² as predictors of favorable outcomes. Subgroup analyses suggested differing predictor patterns between the OPLL and non-OPLL groups. In the OPLL subgroup, lower SII remained independently associated with favorable neurological recovery, with an optimal cutoff value of SII ≤ 558.4. In contrast, in patients with non-OPLL DCM, a distinct combined predictive score was derived from the refitted parsimonious model: Score = - 0.453 + 5.522 × TA - 0.210 × (SII/100). Score ≥ 0.99 was identified as the optimal cutoff for predicting favorable neurological recovery. Corresponding ROC-derived thresholds of SII ≤ 390.6 and TA ≥ 0.445 cm² were identified for predicting favorable neurological recovery in this subgroup. The association between lower SII and favorable recovery remained significant after separate adjustment for each surgical characteristic. The optimism-corrected areas under the ROC curve were 0.749, 0.711, and 0.722 for the overall DCM, OPLL, and non-OPLL DCM models, respectively; goodness-of-fit tests did not indicate lack of fit, and decision curve analysis suggested potential clinical net benefit. CONCLUSION: Systemic inflammatory burden, quantified by SII, was independently associated with long-term neurological recovery in patients with DCM. Subgroup analyses yielded different predictor patterns and SII cutoff estimates in the OPLL and non-OPLL groups; however, these findings should be considered exploratory and require external confirmation. Models integrating systemic inflammatory and radiological parameters may assist risk stratification, but their clinical utility should be established in independent prospective cohorts before implementation.

Abstract as published, via PubMed.