The Role of Serological and Tuberculosis Tests in the Diagnosis of Vertebral Osteomyelitis and the Medical Treatment of Paravertebral Involvement

Introduction

Vertebral osteomyelitis (VO) accounts for approximately 1–7% of all osteomyelitis cases and represents the third most common type of osteomyelitis, particularly in individuals over 50 years of age1-3. Its incidence has increased in recent years, likely due to an aging population, the increased use of invasive procedures, and advances in diagnostic imaging4.

The clinical presentation of VO is often nonspecific. Although axial back pain is the most common symptom, diagnosis may be delayed for weeks or even months. Diagnostic delays of up to four months have been reported5. In the absence of standardized diagnostic and therapeutic ainorithms, management remains challenging. Tissue sampling is performed in relatively few centers, and many physicians avoid these interventions due to the risk of complications.

Contrast-enhanced magnetic resonance imaging (MRI) is the most used imaging modality for the diagnosis of VO. imaging modality of choice, with diagnostic accuracy exceeding 90%. Although histopathological examination and microbiological culture are considered the gold standard for identifying the etiological agent, tissue sampling is performed in relatively few centers and many physicians avoid these interventions due to the risk of complications. Etiological diagnosis, invasive sampling is not always feasible, and the causative organism may remain unidentified5. Consequently, empirical treatment decisions are frequently based on clinical findings and serological tests, including the Brucella standard tube agglutination test (SAT), the tuberculin skin test (PPD), and interferon-gamma release assays (IGRAs).

This condition, which develops mostly spontaneously and partly secondary to surgical or invasive procedures, is more common in elderly and immunosuppressed patients.6 Due to its rarity and nonspecific presentation, VO is frequently overlooked.

This study aimed to evaluate the role of serological and TB tests in the etiological classification of VO, to determine treatment durations according to the causative microorganism, and to assess the clinical significance of paravertebral involvement.

Materials and Methods

Study Design and Population

This retrospective study was conducted in the Infectious Diseases and Clinical Bacteriology Department. Adult patients (≥18 years) diagnosed with VO between February 2017 and February 2022 who completed treatment were included. Patients with prior spinal surgery, hematological malignancy, or incomplete treatment were excluded. A total of 127 patients met the inclusion criteria. We would like to clarify that none of the patients included in this study underwent open or percutaneous biopsy.

The study was approved by the Non-Interventional Clinical Research Ethics Committee (25 April 2022; Decision No: 2022/283) and was conducted in accordance with the Declaration of Helsinki.

Data Collection

Demographic data, comorbidities, laboratory parameters (CRP, ESR, WBC, procalcitonin, and vitamin D), pain severity (Numeric Rating Scale), vertebral level of involvement, paravertebral involvement, microbiological findings, treatment regimens, duration of therapy, and outcomes were analyzed.

Statistical Analysis

Normality was assessed using the Kolmogorov–Smirnov and Shapiro–Wilk tests. Continuous variables were expressed as mean ± standard deviation or median (interquartile range), as appropriate. Normally distributed variables were expressed as mean ± standard deviation, while non-normally distributed variables. The Mann–Whitney U test was used to compare two independent groups for non-normally distributed variables. The Kruskal–Wallis test was applied for comparisons of more than two groups, and Conover post hoc analysis was performed to identify the source of statistically significant differences. A p-value of <0.05 was considered statistically significant.

Demographic data and laboratory findings

The anatomical regions involved, effective treatment times, mean treatment times, MR recovery time, and PVI distribution

Results

Of the 127 patients 69.3% were female. And the mean age was 56.86 ± 13.19 years. Demographic characteristics, sites of involvement, etiological factors, and mean treatment durations are presented in Table 1.

VO was more frequent among females and individuals younger than 65 years. No significant association was observed between vitamin D levels and either pyogenic or granulomatous VO. Among comorbid conditions, hypertension was the most common.

Pyogenic VO was diagnosed in 91 patients (71.6%), tuberculous VO in 31 patients (24.4%), and brucellar VO in 5 patients (3.9%).

The most frequently affected vertebral regions were identified, in descending order, as the lumbar (78%), cervical, and thoracic vertebral segments (Table 2)

Blood cultures were positive in 14 patients (11%). Among these, Brucella spp. was isolated in 2 patients, methicillin-resistant Staphylococcus aureus (MRSA) in 5 patients, and methicillin-sensitive coagulase-negative Staphylococcus (MSCNS) in 7 patients.

Identified pathogens included methicillin-resistant Staphylococcus aureus, methicillin-sensitive coagulase-negative Staphylococcus, and Brucella spp.

PPD and IGRA results did not reliably differentiate etiological groups (p = 0.245). Among patients with tuberculous VO, only 58.1% had positive PPD and/or IGRA results.

Even in patients with positive PPD or IGRA results, pyogenic VO treatment was administered for at least one month. Anti-tuberculosis(anti-TB) therapy was initiated in cases without regression on contrast-enhanced MRI performed after one month of treatment and in whom clinical findings did not improve. In four patients with thoracic and paravertebral involvement and strong PPD or IGRA positivity, anti-TB therapy was initiated directly.

In patients with negative PPD and IGRA tests, anti-TB therapy was also started if no clinical or radiological improvement was observed after three months of pyogenic VO treatment and Brucella agglutination tests remained negative.

Patients with SAT positivity at any titer had a significantly higher likelihood of Brucella spp. or Mycobacterium tuberculosis etiology (p < 0.001). Among six patients initially treated for brucellosis but later switched to anti-TB therapy due to treatment failure, four had SAT titers >1/160 and two had titers <1/160.

CRP levels were significantly higher in tuberculous VO, whereas ESR values were significantly higher in pyogenic VO. WBC count and procalcitonin levels demonstrated limited diagnostic value. No significant association was found between vitamin D levels and VO subtype.

Paravertebral involvement was detected in 26 patients (20.5 was significantly more common in tuberculous and brucellar VO than pyogenic VO (p < 0.001). Twelve patients had associated abscess formation; all responded to medical therapy alone. The lumbar spine was the most frequently affected region across all etiological groups. Mean treatment durations, MRI resolution times, and PVI distributions are shown in Table 2.

Among all patients, 112 (88.1%) were initially treated as pyogenic VO, 11 (8.7%) as brucellar VO, and 4 (3.1%) as TB VO. Following treatment modifications based on clinical examination and MRI findings, complete cure was achieved in 91 pyogenic VO patients, 31 TB VO patients, and 5 brucellar VO patients (Table 2).

All four patients who received direct anti-TB therapy had thoracic and paravertebral involvement. Two of these patients were IGRA-positive, one had a PPD induration of 15 mm, and one was PPD-anergic.

Initial and final treatments in VOM patients

Treatment Protocols

Patients with pyogenic VO received initial therapy with teicoplanin plus ciprofloxacin. Those who did not demonstrate clinical or radiological improvement after one month were switched to daptomycin plus a carbapenem.

Patients with brucellar VO were treated with streptomycin–rifampicin–doxycycline for the three weeks, followed by rifampicin–doxycycline.

Patients with tuberculous VO received isoniazid–rifampicin–ethambutol–pyrazinamide for the first two months, followed by isoniazid–rifampicin.

Treatment modifications were guided by clinical findings and serial MRI evaluations. None of the patients required surgical intervention. Treatment modification was required in 27 patients (21.3%). Of these patients, 22 (81.5%) had lumbar VO and 4 (14.8%) had thoracic VO. In all of these cases, treatment was switched to anti-tuberculosis therapy.

The mean duration of treatment was 3 months (1.5–9 months) for pyogenic VO, 12 months (6–12 months) for tuberculous VO, and 4 months (3–9 months) for brucellar VO.

Treatment modification was required in 27 patients (21.3%), most commonly involving a switch from pyogenic to anti-TB therapy.

Complete clinical and radiological cure was achieved in all patients.

Initial treatment regimens and final treatments after therapy modifications are summarized in Table 3.

Data analysis revealed that PPD and IGRA tests were not useful in distinguishing disease type (p = 0.245). Among patients who underwent PPD testing, 8 were positive, 13 negative, and 10 anergic. Of the 13 patients tested with IGRA, 10 were positive and 3 negatives.

Among the 31 patients treated for TB VO, only 58.1% (n = 18) had positive PPD and/or IGRA results (PPD anergy or ≥15 mm induration was considered positive).

Number and Percentages of PPD and Quantiferon gold test groups in patients who recovered with anti-tuberculosis therapy in the final treatment

*LR

Cross-Table of PPD and Quantiferon results

*LR

Discussion

This study evaluated etiological distribution, diagnostic testing, and treatment strategies in a relatively large cohort of patients with VO. Pyogenic infections were the most common cause, consistent with contemporary epidemiological data, although regional differences may explain variations reported in other studies. The mean patient age was 56.8 years, consistent with the literature. Female predominance (69.3%) was observed; while some studies report higher prevalence among females4, others report male predominance3,7.

In the literature, the most commonly involved regions are reported as lumbar, thoracic, and cervical vertebrae, respectively8. In contrast, our study identified lumbar, cervical, and thoracic vertebrae as the most frequent sites of involvement. Despite evaluating etiological factors, trauma history, and intravenous drug use, no clear explanation for the relatively high cervical involvement was identified, indicating the need for larger, well-designed studies. Lumbar involvement was the most frequent overall, whereas paravertebral involvement was significantly associated with tuberculous and brucellar VO. Importantly, all paravertebral abscesses were successfully managed with medical therapy alone, suggesting that close monitoring may eliminate the need for surgical drainage in selected patients.

Historically, granulomatous infections were the predominant cause of VO; however, with improvements in animal vaccination, sanitation, and surgical techniques, pyogenic infections and TB have become more prevalent9,10. Our findings are consistent with the literature11, although some studies report brucellosis as the leading cause10, likely due to regional epidemiological differences.

Consistent with prior reports, TB VO predominantly involved the thoracolumbar region, whereas brucellar and pyogenic VO were more common in the lumbar and cervical regions8. Differences compared with Gökmen et al.10 may be attributable to regional livestock exposure.

Vitamin D deficiency has been associated with increased susceptibility to infections, including TB, malignancies, multiple sclerosis, and diabetes12,13, and may predispose to osteomyelitis14. However, no prior study has evaluated vitamin D levels in VO. In our cohort, no significant association between vitamin D levels and VO was identified.

VO often presents with nonspecific symptoms, leading to diagnostic delays. While diagnosis is usually prompted by imaging performed for pain, rare asymptomatic cases have been reported15. The mean diagnostic delay in our study was 18 weeks, consistent with reported delays ranging from 6 weeks to 7 months16.

MRI has over 90% diagnostic accuracy in VO17 and served as the primary diagnostic and follow-up modality in our study. Although tissue biopsy remains the diagnostic gold standard, its invasive nature limits its use, and diagnosis often relies on serological testing18, which may be inconclusive.

CRP and ESR were more informative than WBC count or procalcitonin levels; however, none were disease specific. Serial contrast-enhanced MRI played a central role in evaluating treatment response and guiding therapy modification.

Serological and TB tests demonstrated limited discriminatory value. Negative PPD or IGRA results did not exclude tuberculous VO, and false-positive SAT results were observed in patients ultimately diagnosed with TB. Cross-reactivity and false-positive SAT results may occur in endemic regions due to exposure to other pathogens such as Francisella, Yersinia, Legionella, and Mycoplasma19. SAT may yield false-positive results not only in acute brucellosis and miliary TB but also in TB osteomyelitis, and this possibility should be considered during management. Only 7 of 21 patients who were switched from pyogenic to TB therapy had positive PPD or IGRA results, indicating that negative tests do not exclude TB VO. PPD anergy should not be interpreted as absence of disease.

Although current guideline recommends 6–9 months of therapy for extrapulmonary TB, the mean treatment duration for tuberculous VO in this cohort was 12 months. We therefore suggest continuing therapy until complete radiological resolution is achieved.

A major limitation of this study is the absence of histopathological confirmation, as no patient underwent biopsy. Nevertheless, the findings reflect real-world clinical practice in which invasive procedures are often avoided.

Conclusion

Vertebral osteomyelitis is a rare but clinically challenging disease with diagnostic difficulties due to nonspecific early symptoms and potential morbidity in chronic stages. There is a need for standardized diagnostic and therapeutic algorithms supported by well-designed clinical studies. Our findings indicate that serological tests alone are insufficient for diagnosis and that treatment response should be assessed through close clinical follow-up and MRI. Treatment duration should be individualized and extended until complete radiological recovery. Larger studies are required to support guideline-level recommendations.

Statement of ethics

This study was conducted in accordance with 25 April 2022; Decision No: 2022/283, which was granted by the Clinical Research Ethics Committee of Mersin Training and Research Hospital. The study adhered to the Helsinki Declaration.

genAI

No artificial intelligence-based tools or generative AI technologies were used in this study. The entire content of the manuscript was originally prepared, reviewed, and approved by both authors.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest statement

The authors declare that they have no conflict of interest.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.