The continuous increase in the bacterial resistance rate and the slow arrival of new therapeutic options have turned into an antibiotic crisis. One of the strategies proposed by stewardship programs to try to change this situation described worldwide is the use of antibiotics with the lowest possible antimicrobial spectrum.

Enterobacteriaceae bacteremia is a good example of how this strategy would be applied. The empirical treatment of nosocomial bacteremia by Enterobacteriaceae comprises in several cases one or two antibiotics with antipseudomonal activity, being much less common than desirable a subsequent change to narrower spectrum antibiotics based on susceptibility data ("de escalation"). This is because the safety of de escalation is based only on expert advice and some observational studies, so their efficacy and safety is questioned by many clinicians and therefore its use is lower than desired. In fact, a recent systematic review of the Cochrane Library concluded that randomized studies to support this practice are needed. Investigators propose a "real clinical practice-based" randomized trial to compare the efficacy and safety of continuing with an antipseudomonal agents vs. de-escalation according to a pre-specified rule, in patients with bacteraemia due to Enterobacteriaceae.

Inclusion Criteria:

1. ≥18 years old hospitalized patients with bacteremia from any source with isolation of an enterobacteria in blood cultures.
2. Active empiric treatment with antipseudomonal betalactamic at 48 hours from the symptoms of sepsis and the blood culture.The patient could have received any other type of antibiotic therapy up to 24 hours after blood extraction.
3. Microorganism susceptible at least one treatment from the experimental arm.
4. Patients with intravenous treatment at least 3 days from the randomization o 5 days from the initial blood culture.
5. Patients to sign the informed consent form.

Exclusion Criteria:

1. Palliative care or life expectance < 90 days.
2. Pregnancy or lactation period.
3. To isolate the Extended-spectrum β-lactamases producing Enterobacteriaceae
4. Late randomization >48 hours after the enterobacteriaceae blood culture´s identification
5. Severe neutropenic (< 500 céls/mm3) at the randomization.
6. Treatment of infection > 28 days (endocarditis and osteomyelitis) or meningitis.