Background: The incidence of bacteremia has increased over the last decade due to the aging of the population and the wide implementation of invasive nosocomial procedures and wide-spectrum antimicrobial treatments. Aim: To assess the epidemiology of hospital and ambulatory bacteremias in a public hospital in Santiago. Material and Methods: A prospective longitudinal cohort study of laboratory-confirmed adult patients with bacteremia was undertaken at a public hospital in Santiago, between June 1, 2007 and April 30, 2008. Demographic, clinical, and laboratory data was collected into a standardized study-specific form. Results: In the study period, 253 cases of true bacteremia were identified, with an incidence of 11 per 1,000 patient discharges (63.6% nosocomial, 36.7% fatal). Only 2/3 of the cohort was alive at day 28 of hospitalization. Variables associated with mortality were age with a hazard ratio (HR) of 2.31; (95% confidence intervals (CI) 1.42- 3.77); female gender (HR, 1.70; 95% CI 1.06- 2.71); shock (HR, 3.24; 95%CI 2.01- 5.22); and C reactive protein (HR, 2.10; 95% CI 1.17- 3.78). The variable associated with lower mortality was surgery (HR, 0.43; 95% CI 0.25- 0.75). Selective empiric treatment did not improve survival. Conclusions: Besides age and gender, survival can be influenced by modifiable variables such as presence of shock and surgical intervention, which may provide an opportunity to improve outcomes.