Prognostic value of the identified genus of anaerobic bacteria in bloodstream infection beyond host-related risk factors: a five-year retrospective cohort study.

BACKGROUND: Anaerobic bacteraemia is an uncommon but clinically significant bloodstream infection associated with substantial mortality. The respective roles of pathogen-specific and host-related factors in determining outcomes are not fully elucidated. AIM: To identify predictors of mortality in anaerobic bacteraemia and to evaluate whether genus-based microbiological classification offers prognostic value beyond established clinical risk factors. METHODS: A retrospective cohort study was performed involving patients with anaerobic bacteraemia (n = 96) treated at a tertiary-care centre in Hungary from January 2021 to December 2025. Survival was analysed using Kaplan-Meier estimates and Cox proportional hazards regression. The incremental prognostic value of anaerobic genus was assessed using nested multivariable models, with Firth-penalised logistic regression applied as a sensitivity analysis. RESULTS: Among 96 patients (median age 68 years), the most common isolates were Bacteroides spp. (33%), Clostridium spp. (17%), Actinomyces spp. (16%), and Fusobacterium spp. (9%). Mortality rates were 43.8% at 30 days, and 54.2% at 180 days. Survival varied significantly by genus, with Clostridium spp. bacteraemia exhibiting the highest mortality (75.0%). Independent predictors of 180-day mortality were increasing age (HR 1.03, p = 0.037), acute kidney injury (HR 2.27, p = 0.007), liver cirrhosis (HR 2.83, p = 0.022), hypertension (HR 1.95, p = 0.048), and Clostridium spp. bacteraemia (HR 3.00, p = 0.006). Microbiological classification significantly improved model performance (likelihood-ratio p = 0.016). The final model had an apparent 180-day AUC of 0.81. CONCLUSIONS: Bacterial genus provides prognostic information beyond host-related risk factors in anaerobic bacteraemia. Clostridium spp. bacteraemia was associated with the highest mortality and independently predicted 180-day mortality.