Clinical Characteristics of Patients with Pyogenic Vertebral Osteomyelitis and Concurrent Infections and Their Clinical Outcomes.

Abstract

Patients with pyogenic vertebral osteomyelitis (PVO) often develop concurrent infections, and a significant number of these patients show rapid deterioration in their medical condition, leading to mortality without PVO-related structural instability or neurological deficits. To improve clinical outcomes, we investigated the clinical presentation and treatment outcomes of patients with PVO and concurrent infections. This study included 695 patients with PVO, of which 175 (25%) had concurrent infections and 520 (75%) did not. The clinical characteristics of the two groups were compared, and multivariable analysis was performed to identify the association between concurrent infections and clinical outcomes. Patients with concurrent infections were older and had more comorbidities than those without. Moreover, there were significant intergroup differences in the anatomical involvement of PVO, and patients with concurrent infections had a higher number of regions involved more frequently than those without concurrent infections (15% vs. 6%). In contrast, patients with concurrent infections showed a lower degree of focal invasiveness, including a lower incidence of posterior abscess (47% vs. 59%; p = 0.008) and fewer neurological impairments according to the American Spinal Injury Association grade (p < 0.001) than those without concurrent infections. The causative organisms also differed significantly between the two groups, and patients with concurrent infections had a greater proportion of Gram-negative infections (31% vs. 16%, respectively) and a smaller proportion of methicillin-resistant S. aureus infections than those without concurrent infections (6% vs. 24%). Consequently, their clinical outcomes were significantly different, and patients with concurrent infections showed lower recurrence and higher mortality rates. We investigated the 1-year recurrence and mortality rates and their 95% confidence intervals according to the types of concurrent infections and their time of diagnosis and found variations in these parameters. Our results, based on a large number of patients, can be practically used as a reasonable reference to warn clinicians of the clinical risks of concurrent infections in patients with PVO and to help predict their clinical outcomes.