Two-week antibiotic courses have been considered standard care for most patients with bacteremia who do not have sepsis or an untreated primary source (e.g. endocarditis). No good evidence ever supported the practice, which was supported mainly by retrospective data in patients with sepsis. A new study suggests that treating gram-negative bacteremia for seven days is equally effective as 14 days of treatment, if resolution of sepsis and source control are both present.