The investigators plan an open label randomized controlled trial to compare short-course antibiotic therapy (<=7 days) versus longer treatment (>7 days). The investigators will include hospitalized patients with gram-negative bacteremia. The investigators primary objective is to investigate the safety and efficacy of short-course antibiotics.

Inclusion Criteria:

• patients with gram-negative aerobic bacilli bacteremia, defined as growth of a single gram-negative microorganism in one or more blood cultures, associated with evidence of infection (hyper- or hypothermia, a localized infection, sepsis or septic shock).
• We will include patients receiving appropriate antibiotic treatment for 7 days and are afebrile / not hypothermic for the last 48 hours. Both community and hospital acquired gram-negative bacteremias will be included, regardless of antibiotic susceptibility patterns. We will allow the inclusion of patients receiving less than 7 days if clinically stable and discharge from hospital is considered. We will then recruit the patient before discharge, if stable at least for 48 hours before randomization.

We will include the following sources of bacteremia:

1. Primary bacteremia / unknown source
2. Urinary tract
3. Abdominal
4. Respiratory tract
5. Central venous catheter(CVC), when the catheter was removed before randomization
6. Skin and soft tissue, including surgical site infection

Exclusion Criteria:

1. Gram-negative bacteremia due to specific infections as detailed here:

   1. Endocarditis / endovascular infections
   2. Necrotizing fasciitis
   3. Osteomyelitis
   4. Abdominal abscesses and other unresolved abdominal sources requiring surgical intervention (e.g., cholecystitis)
   5. Central nervous system infections
   6. Empyema
   7. CVC- related or CVC-associated bloodstream infections when the catheter is retained. We will permit the inclusion of patients with retained CVCs in whom the source of the bacteremia is not the CVC.

2. Polymicrobial growth in blood cultures involving gram-positive or anaerobes in addition to gram-negatives (defined as either growth of two or more different species of microorganisms in the same blood culture, or growth of different species in two or more separate blood cultures within the same episode (< 48 h) and with clinical or microbiological evidence of the same source).

3. Specific pathogens including:

   1. Salmonella spp.
   2. Brucella spp.

4. Immunosuppression, including:

   1. HIV infection
   2. Hematopoietic stem-cell transplantation
   3. Neutropenia on day of randomization or in the 48 hours prior to randomization. Patients with neutropenic fever at presentation that are afebrile and non-neutropenic in the 48 hours before randomization will be included.

5. Clinical instability during the 48 hours before randomization, defined as mean blood pressure<60 mmHg despite adequate fluid resuscitation or vasopressors support.

6. Repeated positive blood cultures for the same organism separated by at least 24 hours, regardless of antibiotic treatment. Patients with repeated isolates on the first 24 hours will be included.

7. Uncontrolled focus of infection: e.g. an abscess that was not drained sufficiently; non-drained moderate to severe hydronephrosis in a patient with bacteremia of urinary source; deep seated intra-abdominal infections that were not drained properly.

8. Fever > 38.0C measured at least twice in the 48 h prior to recruitment; or > 38.5C once during the 48 h; or hypothermia <35.5C measured once during the 48 h.

9. Previous enrollment in this trial

10. Concurrent participation in another clinical trial