Development and validation of a clinico-biological score for Brucella spondylitis: a two-center study

Brucellosis is a zoonotic disease distributed across numerous countries and regions worldwide, presenting with diverse clinical manifestations. The most common complication is spondylitis, which is diagnosed primarily through imaging studies. However, in resource-limited areas, the imaging examinations necessary for diagnosing brucellosis-related spondylitis are often inaccessible. The objective of this study is to establish a simple and readily predictive score within brucellosis patients to screen for spinal involvement.

We retrospectively collected patient data with brucellosis admitted to the Kashi Affiliated Hospital of Sun Yat-sen University from January 2019 to December 2023, and randomly assigned them into a training cohort and an internal validation cohort. Data of brucellosis patients admitted to The third Affiliated Hospital of Sun Yat-sen University from January 2014 to December 2023 were collected for an external validation cohort. A diagnostic model was constructed by using a nomogram. Calibration plots, receiver operating characteristic curve, and decision curve analyses were employed to evaluate the model's calibration, accuracy, and clinical utility.

This study included data from total of 784 patients, of which 210 were diagnosed with Brucella spondylitis. The data was divided into a training cohort (460 patients), an internal validation cohort (198 patients), and an external validation cohort (126 patients). The diagnostic model was formulated using six diagnostic factors: course of disease, age, back pain, joint pain, white blood cell count, and levels of C-reactive proteins. In our study, the AUC values of 0.93 (training), 0.87 (internal validation), and 0.77 (external validation) indicate that the model maintained excellent discriminative ability in the training and internal validation cohorts, and acceptable performance in the external cohort. Decision curve analyses graphically display the significant clinical utility and net benefit of a nomogram.

The diagnostic model for Brucella spondylitis developed in this study has the potential to assist clinicians in resource-limited settings in achieving a rapid and effective diagnosis of the disease. Our model facilitates the identification of brucellosis-related spondylitis patients requiring extended treatment courses, thereby reducing misdiagnosis and missed diagnosis in resource-constrained areas where imaging examinations are difficult to access, and improving the efficiency of healthcare resource utilization.