Abstract
BACKGROUND: Melioidosis, caused by Burkholderia pseudomallei, is an underdiagnosed cause of community-acquired pneumonia (CAP) in India. Due to overlapping features with other bacterial pneumonias and limited access to culture facilities, early diagnosis and treatment remain challenging. This study aimed to develop a clinical scoring system to distinguish melioidosis from other bacterial causes of CAP in an endemic setting. METHODS: We conducted a retrospective cohort study of 337 patients with radiologically confirmed blood or respiratory culture-positive CAP cases at a tertiary care hospital in South India from 2017 to 2023. This included 55 melioidosis cases and 282 controls with other documented bacterial etiologies. Demographic, clinical, laboratory, and radiological variables were compared. Multivariable logistic regression identified independent predictors of melioidosis. A scoring system was developed using the natural logarithms of adjusted odds ratios (aORs). RESULTS: Four independent predictors were retained in the final model: monsoon season exposure (aOR = 9.0, 95% CI: 3.6-22.6), diabetes mellitus (aOR = 10.1, 95% CI: 3.6-28.5), shock at presentation (aOR = 17.2, 95% CI: 5.9-49.9), and extrapulmonary focal involvement (aOR = 36.5, 95% CI: 11.0-121.4). The model showed excellent discrimination. A score of ≥ 4 out of 11 yielded a sensitivity of 87.3% and specificity of 83.6%, while a score of ≥ 5 yielded a sensitivity and specificity of 67.3% and 95.4%, respectively. CONCLUSION: We propose a simple four-point clinical scoring tool to identify melioidosis in patients with CAP. This score can guide early suspicion and appropriate therapy in endemic resource-limited settings. Prospective validation in other endemic regions is warranted.