Abstract
BACKGROUND: Community acquired bloodstream infection (CABSI) in low- and middle income countries is associated with a high mortality. This study describes the clinical manifestations, laboratory findings and correlation of SOFA and qSOFA with mortality in patients with CABSI in northern Vietnam. METHODS: This was a retrospective study of 393 patients with at least one positive blood culture with not more than one bacterium taken within 48 h of hospitalisation. Clinical characteristic and laboratory results from the first 24 h in hospital were collected. SOFA and qSOFA scores were calculated and their validity in this setting was evaluated. RESULTS: Among 393 patients with bacterial CABSI, approximately 80% (307/393) of patients had dysfunction of one or more organ on admission to the study hospital with the most common being that of coagulation (57.1% or 226/393). SOFA performed well in prediction of mortality in those patients initially admitted to the critical care unit (AUC 0.858, 95%CI 0.793-0.922) but poor in those admitted to medical wards (AUC 0.667, 95%CI 0.577-0.758). In contrast qSOFA had poor predictive validity in both settings (AUC 0.692, 95%CI 0.605-0.780 and AUC 0.527, 95%CI 0.424-0.630, respectively). The overall case fatality rate was 28%. HIV infection (HR = 3.145, p = 0.001), neutropenia (HR = 2.442, p = 0.002), SOFA score 1-point increment (HR = 1.19, p < 0.001) and infection with Enterobacteriaceae (HR = 1.722, p = 0.037) were independent risk factors for in-hospital mortality. CONCLUSIONS: Organ dysfunction was common among Vietnamese patients with CABSI and associated with high case fatality. SOFA and qSOFA both need to be further validated in this setting.