Pyogenic versus tuberculous and brucellar spondylodiscitis in a Tunisian population: a bicentric comparative study

Abstract Background Infectious spondylodiscitis (ISD) represents a serious spinal infection with significant diagnostic and therapeutic challenges. In Tunisia, where tuberculosis and brucellosis are endemic, accurate discrimination between pyogenic and specific etiologies is essential to guide appropriate antimicrobial. This study aims to identify clinical, laboratory, and radiological features that enable accurate pathogen differentiation in cases lacking microbiological confirmation. Methods A bicentric retrospective study conducted in Rheumatology and Infectious Diseases departments in Mahdia and Monastir, Tunisia (2012–2024). In total, 131 patients with ISD, diagnosed on the basis of integrated clinical, biological, radiological, and microbiological criteria, were included in the study: 86 with pyogenic spondylodiscitis (PSD) and 45 with specific spondylodiscitis (SSD). Results PSD predominantly affected older patients (median age 64 vs 50 years, p < 0.001), presenting with acute onset (58.1% vs 24.4%, p < 0.001) and shorter consultation delays (30 vs 90 days, p < 0.001). PSD patients had significantly elevated inflammatory markers (median CRP 90.36 vs 28 mg/L, p = 0.004; median ESR 92.7 vs 59.3 mm/h, p < 0.001) and higher prevalence of diabetes (44.7% vs 13.3%, p < 0.001) and renal failure (14% vs 2.2%, p = 0.034). SSD showed higher prevalence in rural populations (68.9% vs 50%, p = 0.033) and lower socioeconomic status ( p = 0.001), with more frequent night sweats (48.9% vs 29.1%, p = 0.025), vertebral cysts (33.3% vs 4.9%, p < 0.001), and soft tissue abscesses (72% vs 46.3%, p = 0.031). In microbiologically confirmed PSD cases (50.0%; 43/86), Staphylococcus aureus was the most frequently isolated pathogen (62.8%; 27/43). The duration of antimicrobial therapy differed significantly between groups (PSD: 104 days vs SSD: 270 days). Favorable clinical outcomes were observed in 83.7% of PSD patients and 93.3% of SSD patients. Overall cohort mortality was 3.8% (5/131) and occurred exclusively in the PSD group (5/86; 5.8%). Conclusions In this bicentric study, we delineated distinct clinical, demographic, and radiological profiles associated with PSD and SSD. In multivariable models, diabetes mellitus, advanced age, and elevated ESR remained independently associated with PSD, whereas the presence of vertebral cystic lesions on plain radiography was independently associated with SSD. However, in light of the retrospective design and the considerable proportion of cases lacking microbiological confirmation, these results should be interpreted as clinically informative associations rather than as a validated diagnostic algorithm and they require confirmation in larger prospective cohorts. Recognition of these differential presentations may refine pre-test clinical suspicion for pyogenic versus specific etiologies in endemic settings while awaiting microbiological confirmation and may facilitate earlier initiation of appropriate empirical therapy.