Clinical and hematological factors associated with neurological involvement in spinal tuberculosis: a retrospective study

Background The association between peripheral blood immune-inflammatory markers and neurological involvement in patients with spinal tuberculosis, and to develop a predictive model for early risk assessment. Spinal tuberculosis with neurological impairment is a major cause of disability, yet simple, cost-effective early warning indicators remain lacking. Peripheral blood immune-inflammatory markers are valuable in infectious disease prognostication, but their predictive role in spinal tuberculosis neurological involvement remains unclear. Methods A retrospective study was conducted on 294 patients with spinal tuberculosis who underwent surgical treatment in the Department of Orthopedics, General Hospital of Ningxia Medical University, between December 2019 and December 2024. Patients were stratified into two groups: the uncomplicated spinal tuberculosis group (n=194, ASIA grade E) and the neurological involvement group (n=100, ASIA grades A–D). Fasting venous blood test results collected at the initial presentation were analyzed. Immune-inflammatory markers, including lymphocyte percentage (LYM), mixed cell percentage (MXD), and platelet-to-lymphocyte ratio (PLR), were compared between the two groups. Univariate and multivariate logistic regression analyses were performed to identify influencing factors, followed by the establishment of a predictive model. Receiver operating characteristic (ROC) curves and a nomogram were constructed to evaluate the diagnostic efficacy of the model. Results The neurological involvement group exhibited a significantly lower LYM [23.150% (16.900, 29.525) vs. 26.150% (19.750, 32.900)], and significantly higher MXD [9.000% (7.375, 10.200) vs. 8.100% (6.725, 9.900)] and PLR compared with the uncomplicated group ( P < 0.05). Multivariate analysis revealed that LYM (odds ratio [OR]=0.961, 95% confidence interval [CI]: 0.935–0.989) and MXD (OR = 1.107, 95% CI: 1.013–1.209) were independent predictors of neurological involvement in spinal tuberculosis. Specifically, each 1% increase in LYM was associated with a 3.9% reduction in the risk of neurological impairment, whereas each 1% increase in MXD correlated with a 10.7% increase in risk. The combined predictive model achieved an AUC of 0.803 (95% CI: 0.749–0.857), with a sensitivity of 70.0% and specificity of 80.9%. The calibration curve confirmed good model fit (χ²=4.215, P = 0.837). In addition, the Brier score was 0.185, indicating favorable overall accuracy of the probabilistic predictions. Conclusion Decreased peripheral blood LYM and increased MXD are independent risk factors for neurological involvement in spinal tuberculosis. The combined predictive model shows favorable diagnostic efficacy and calibration. The nomogram is a simple, economical clinical tool for early high-risk patient identification, guiding individualized treatment decisions.