ABSTRACT

Objective

Single-stage posterior reconstruction is used for spinal tuberculosis, but outcome data vary. To report outcomes after single-stage posterior debridement/decompression, titanium mesh cage fusion, and posterior instrumentation.

Materials and Methods

Retrospective series of 18 patients (2014-2019) treated with a posterior single-stage approach and ≥22 months’ follow-up. Indications included neurological deficit; instability, progressive collapse, or deformity; compressive epidural or paravertebral abscess; and/or failure of anti-tuberculosis therapy. Outcomes included regional kyphosis/lordosis, visual analog scale (VAS), Frankel grade, Oswestry disability index (ODI) at final follow-up, and C-reactive protein (CRP)/erythrocyte sedimentation rate (ESR).

Results

Mean age was 36.3±11.7 years; mean follow-up 37.9±20.1 months. Involvement was thoracolumbar in 10 cases, thoracic in 4, and lumbar in 4. In thoracic/thoracolumbar cases (n=14), kyphosis improved from approximately 30° to 15.2° (p=0.005); lumbar lordosis showed no significant change (p=0.655). VAS decreased from 7.81 to 3.15 (p<0.002). Frankel increased from 4.75±0.52 to 4.92±0.28 (p>0.05); two patients improved (D→E) with no deterioration. Final ODI was 27.33±17.40 (median 22; 10-64). CRP rose early and returned toward baseline by ~3 months; ESR showed no significant change at ~3 months.

Conclusion

Posterior single-stage reconstruction was associated with maintained regional kyphosis correction and significant pain reduction; functional improvement could not be quantified without baseline ODI.

INTRODUCTION

Pott disease is a granulomatous infection by Mycobacterium tuberculosis that primarily involves the vertebral bodies, spreads subligamentously across discs, and can lead to collapse, kyphosis, and neurologic deficit⁽¹⁾. Extrapulmonary involvement accounts for ~15-20% of all tuberculosis (TB) cases in recent cohorts; half of these involve the spine, and 10-45% develop neurological deficits and vertebral damage^(2-5). The spine is the most frequently affected skeletal site and the one most prone to serious complications^(2-5). Pathological fractures and dislocations produce mechanical instability, while compression from bony fragments or abscesses leads to neurological compromise-scenarios that often render surgery unavoidable^{(6, 7)}. Surgical goals in spinal TB are to eradicate infection, restore spinal stability, and address neurological dysfunction. To this end, a spectrum of procedures has been described: abscess drainage⁽⁸⁾, anterior⁽⁹⁾ and posterior^{(6, 7)} debridement, postero-lateral^{(10, 11)} approaches, anterior or posterior grafting⁽¹²⁾ and kyphosis correction with preservation of fusion to prevent late deformity^{(7, 13)}. Because anterior and postero-lateral strategies can demand broader multidisciplinary resources and may carry higher complication profiles, single-stage posterior surgery is increasingly favored. In this study, we evaluate outcomes of single-stage posterior transpedicular debridement and decompression with titanium mesh cage fusion and posterior instrumentation, and we outline its advantages and limitations relative to alternative methods.

MATERIALS AND METHODS

Between 2014 and 2019, 18 patients (12 females, 6 males) with vertebral TB, who underwent single-stage posterior debridement, decompression, fusion and instrumentation, were included in the study. The mean age was 36.3±11.7 years (range: 18-72). Eligible patients were those with vertebral TB treated with single-stage posterior debridement, decompression, titanium mesh cage fusion, and posterior instrumentation, with a minimum follow-up of 22 months. Patients with follow-up <22 months and/or incomplete clinical-radiographic data were excluded. Surgical indications included neurological deficit due to compression, mechanical instability/progressive collapse or deformity, compressive epidural/paravertebral abscess, and/or failure of appropriate anti-TB chemotherapy with clinical or radiological progression. The level of involvement at presentation was thoracolumbar (TL) in 10, thoracic (T) in 4, and lumbar (L) in 4 patients. Ethical approval was obtained from the University of Health Sciences Türkiye, Gazi Yaşargil Training and Research Hospital Ethics Committee for Non-Interventional Studies (approval no: 694, date: 07.11.2025). In addition, a written informed voluntary consent form was obtained from all patients participating in the study. For final follow-up assessment, patients were contacted by phone/e-mail and invited for clinical and radiographic evaluation. Demographic data, medical history, treatment modalities, physical examination findings, functional outcomes, and laboratory results were evaluated. Imaging studies, including plain radiographs, magnetic resonance imaging, and computed tomography, were also reviewed. In addition, preoperative and postoperative physical examination, laboratory, and radiography findings of the patients were retrieved via the hospital information system. Neurological status was graded preoperatively and at final follow-up using the Frankel classification (A-E)⁽¹⁴⁾. Pain intensity was assessed preoperatively and postoperatively using the visual analog scale (VAS)⁽¹⁵⁾. Functional status was evaluated at final follow-up using the Oswestry disability index (ODI)⁽¹⁶⁾, with higher scores indicating greater disability.

Surgical Procedure

A midline incision was performed, hemostasis was achieved, and the paraspinal muscles were subperiosteally elevated to allow pedicle screw placement. Depending on the level of involvement, laminectomy was performed to expose the spinal cord, and tuberculous debris was circumferentially debrided with specimens sent for microbiology and pathology. The cleaned cavity was irrigated and filled with an appropriately sized titanium cage, and rods were contoured for kyphosis/lordosis correction and fixed. Facet and pedicle sites were grafted and vancomycin powder was applied, a Hemovac drain was placed, and the wound was closed in layers (Figure 1A-G).

Postoperative management included thoracolumbosacral orthosis measurement on the first postoperative day; if there was no contraindication, the patient was mobilized and instructed to wear the corset during mobilization. Corset use for 6-12 months was recommended depending on the level of involvement. Anti-TB chemotherapy was initiated postoperatively under infectious disease supervision using a daily 4-drug regimen (isoniazid, rifampicin, pyrazinamide, and ethambutol) for 2 months, followed by daily isoniazid plus rifampicin for 10 months with weight-based dosing. Microbiological specimens were routinely submitted for smear/culture and drug-resistance assessment (drug-susceptibility testing when culture yielded growth and rapid molecular testing when available), and the regimen was modified according to susceptibility results and tolerability. A 12-month duration was selected for this spinal osteoarticular TB cohort treated with instrumentation, consistent with expert recommendations and contemporary World Health Organization/national guidance that commonly allow extended treatment durations for bone/joint (including spinal) disease^{(17, 18)}. Routine controls were scheduled at post-op 2 weeks, 6 weeks, 3 months, 6 months, and annually thereafter. Radiological fusion was defined as trabecular bridging across the cage/graft-endplate interfaces without progressive radiolucency or hardware failure. Loss of correction was defined as ≥5° deterioration in regional kyphosis/lordosis versus immediate postoperative values. Implant loosening was defined as progressive ≥1-mm peri-screw radiolucency and/or hardware migration/breakage, and cage subsidence as ≥3-mm loss of segmental height or cage sinking compared with immediate postoperative films. Infection recurrence was defined as renewed disease activity after initial improvement, supported by rising erythrocyte sedimentation rate (ESR)/C-reactive protein (CRP) and radiological progression (progressive abscess or bone destruction). Regional kyphosis/lordosis angles were measured by specialist surgeons in our clinic and were re-evaluated by a senior surgeon; measurements were revised when necessary after repeat assessment.

Statistical Analysis

Statistical analysis was performed using IBM SPSS Statistics 21.0. Continuous variables are presented as mean ± standard deviation and categorical variables as n (%). Normality was assessed with Kolmogorov-Smirnov/Shapiro-Wilk tests. Paired comparisons used the paired t-test or Wilcoxon signed-ranks test, as appropriate. All tests were two-tailed; p≤0.05 was considered significant.

RESULTS

Eighteen patients (12 female, 6 male) were analyzed. Mean age was 36.3±11.7 years (range: 18-72). Mean hospital stay was 8.25±3.41 days and mean follow-up was 37.92±20.1 months (range: 22-94). The involved levels were TL (n=10), T (n=4), and L (n=4) (Table 1). In the L subgroup (n=4), lordosis showed no significant change (preoperative mean 27.5°; p=0.655). Pain decreased from a mean VAS of 7.81 preoperatively to 3.15 postoperatively (p<0.002) (Figure 2). In the T/TL subgroup (n=14), regional kyphosis improved from approximately 30° preoperatively to approximately 15.2° at final follow-up (p=0.005) (Figure 3). Mean Frankel grade increased from 4.75±0.52 (median 5; range: 4-5) preoperatively to 4.92±0.28 (median 5; range: 4-5) at final follow-up (p>0.05). Two patients improved from Frankel D to E, and no neurological deterioration occurred. At final follow-up, functional status was assessed using the ODI. The mean ODI was 27.33±17.40 (median 22; range: 10-64), as summarized in Table 1. Regarding inflammatory markers, CRP increased in the early postoperative period (1.39→13.49 mg/L) and returned toward baseline by 3 months (2.08 mg/L), while ESR showed no significant change at 3 months (32.83→30.08 mm/h). No recurrence or loss of correction was observed during follow-up according to the predefined criteria.

DISCUSSION

Defining a single, universally applicable treatment algorithm for spinal TB is challenging because acceptable outcomes can be achieved with different surgical strategies. That said, contemporary implants and posterior techniques have made single-stage posterior surgery a commonly preferred option in many centers, as it can provide decompression/debridement, regional deformity correction, and stable reconstruction through one approach with less soft-tissue morbidity than combined procedures. Anterior or combined anterior-posterior approaches may still be appropriate in selected scenarios (e.g., extensive anterior column destruction, specific abscess patterns, or cases requiring anterior structural support), but our study was not designed to compare strategies and therefore does not support claims of superiority.

Pain reduction in our cohort was comparable to prior posterior-only series, where VAS reductions of approximately 4-6 points have been reported^{(9, 10, 13)}. In this study, mean VAS decreased from 7.81 to 3.15 (p<0.002), indicating a clinically relevant reduction in pain after single-stage posterior reconstruction.

Neurological change in our cohort was modest. Although neurological improvement after posterior surgery is frequently reported^{(11, 13)}, only two patients improved (Frankel D→E) and the overall change was not statistically significant. Mean Frankel grade increased from 4.75±0.52 (median 5; range: 4-5) preoperatively to 4.92±0.28 (median 5; range: 4-5) at final follow-up (p>0.05), with no neurological deterioration. This likely reflects the small sample size and the relatively preserved baseline neurological status, limiting the ability to detect measurable neurological improvement.

Radiographic findings in our study should be interpreted as regional correction rather than global sagittal realignment. Prior posterior-only series have reported meaningful reductions in T kyphosis after debridement, instrumentation, and fusion^{(11, 19)}. In our T/TL subgroup, kyphosis decreased from approximately 30° preoperatively to 15.2° at final follow-up, indicating maintained regional correction. In contrast, L lordosis did not change significantly, and this neutral finding is reported separately. Overall, these data are consistent with posterior single-stage reconstruction providing durable regional kyphosis correction in T/TL disease, while effects on L alignment appear limited in this small cohort.

In our cohort, T/TL involvement predominated and patients were relatively young (mean age 36.3 years), consistent with other published series. The mean length of hospital stay was 8.25±3.41 days, which lies at the lower end of ranges reported in prior posterior-only and combined approach cohorts (reported ~10-18 days)⁽²⁰⁾. This comparison should be interpreted cautiously because length of stay is strongly influenced by local healthcare systems, perioperative pathways, and discharge practices.

Inflammatory markers showed a postoperative pattern that is commonly observed after major spinal procedures. CRP increased early postoperatively and returned toward baseline by approximately 3 months in our cohort, whereas ESR demonstrated a slower and less specific trajectory, as reported in prior studies^{(21, 22)}. Given that CRP/ESR were obtained at limited time points, these measures should be interpreted as supportive laboratory trends rather than standalone indicators of TB control or recurrence.

Functional interpretation is limited by measurement design. ODI was available only at final follow-up (mean 27.33±17.40; median 22; range: 10-64), and baseline ODI was not recorded; therefore, functional improvement cannot be quantified in this series. Accordingly, ODI is reported strictly as final functional status rather than pre-post change. Final ODI values were broadly comparable to those reported in prior posterior-only cohorts⁽²³⁾.

Study Limitations

This study is limited by its retrospective design, small and heterogeneous sample (n=18), absence of a comparator group, and limited power for subgroup analyses, particularly for L disease. Follow-up imaging was not fully standardized, and radiographic assessment was restricted to regional kyphosis/lordosis angles; global sagittal alignment and pelvic parameters were not evaluated, precluding conclusions regarding global alignment. Patient-reported outcomes were limited to VAS and ODI, with ODI available only at final follow-up and without broader health-related quality of life instruments,  which restricts comparability and prevents robust assessment of functional change. CRP/ESR measurements were obtained at limited time points, and rehabilitation adherence and provider-related variability were not systematically captured. Finally, formal interobserver reliability statistics were not calculated. These limitations reduce generalizability and underscore the need for prospective comparative studies with standardized PROMs, comprehensive alignment assessment, and predefined reliability analysis.

CONCLUSION

In this small retrospective series, single-stage posterior debridement/decompression with titanium mesh cage-assisted fusion and posterior instrumentation was associated with durable regional kyphosis correction in T/TL disease and significant postoperative pain reduction. Final functional status was favorable (mean ODI 27.33±17.40; median 22; range: 10-64); however, functional improvement cannot be quantified because baseline ODI was unavailable. No recurrence or loss of correction was observed during follow-up according to the predefined radiological criteria.