A novel clinical classification of spinal tuberculosis for decision-making guidelines from chinese spine surgeons: a multicenter retrospective cohort study

BACKGROUND Spinal tuberculosis (TB) presents substantial therapeutic challenges as it can lead to severe structural and neurological complications. However, evidence-based therapeutic decision-making is hindered by the limitations of existing classification systems, which lack anatomical specificity, quantifiable parameters, and reliability. METHODS The classification system proposed herein categorizes spinal TB into four types based on clinical manifestations and radiologic features: type I (mild): single-segment involvement, with <50% vertebral collapse and preserved stability; type II (moderate): paravertebral abscesses extending beyond the height of two vertebrae (IIA: prevertebral and IIB: intraspinal); type III (severe): single vertebral collapse ≥50%, kyphosis of 30°–60°, and spinal instability; and type IV (extremely severe): kyphosis >60° with sagittal imbalance. Cohen’s kappa was used to assess the interrater and intrarater reliability. RESULTS The proposed classification system demonstrated robust clinical applicability in a cohort of 2,520 patients with spinal TB (type I, 160; type II, 242; type III, 1,992; and type IV, 126), with excellent interrater reliability (intraclass correlation coefficient [ICC] = 0.925 initial, 0.948 on reassessment) and intrarater consistency (ICC = 0.943). Type III spinal TB was the most common in this classification system, comprising 79.05% of cases. Patients with type III spinal TB exhibited moderate kyphotic deformity (mean Cobb angle: 41.5° ± 10.8°), and 38.4% of these patients had neurological deficits. Patients with type IV spinal TB demonstrated a severe sagittal imbalance (mean Cobb angle: 94.3° ± 23.8°), and 51.6% of these patients had neurological deficits. Multidrug antitubercular therapy achieved a 94.6% success rate in patients with type I spinal TB. For patients with type II spinal TB, computed tomography-guided abscess drainage (mean Hounsfield Units threshold: 22.3) or surgical debridement was selected based on abscess density, yielding a 93.3% clinical cure rate. Management of type III spinal TB involved anterior/posterior debridement with spinal reconstruction tailored to lesion topography, neurological deficit severity, and surgeon’s expertise, achieving disease resolution or fusion in 95.5% of patients. Patients with type IV spinal TB underwent corrective osteotomy via posterior-only or combined anteroposterior approach depending on deformity rigidity, achieving a mean kyphosis correction rate of 72.0%. CONCLUSIONS The proposed classification system demonstrated high reproducibility and reliability for spinal TB. The classification algorithms and corresponding treatment protocols are based on distinct clinical manifestations and radiographic characteristics. Moreover, this system shows good potential to facilitate standardized therapeutic management of spinal TB.