The SATURN ICU-trial studies the effect of antibiotic rotation on the prevalence of antibiotic resistant Gram-negative colonisation.

Antibiotic rotation has been previously studied with varied results. The theory behind antibiotic rotation is that intermittently changing antibiotic classes will reduce the ecological selective pressure that drives the emergence of antibiotic resistance.

This study compares the effect of 2 types of antibiotic rotation on Gram-negative colonisation in the ICU and also compares both interventions with standard care.

The two interventions apply to the empiric treatment and are: 1) "fast" rotation, i.e. every other patient another class and 2) "slow" rotation, i.e. every other 1.5month another preferred class for empiric Gram-negative antibiotic therapy.

Inclusion Criteria:

• There are at least 8 beds, with an average bed-occupancy of 80%; all of which have capacity for mechanical ventilation.
• The ICU can adhere to the selected antibiotics for empiric treatment of infections.
• There is an operational digital patient-information system, from which data can be extracted and delivered in a pre-defined format. Specifically an automated process for digital data-collection regarding microbiological culture-results (from swabs and bacteraemias), antibiotic prescription and patient demographics and illness severity-scores.
• Colonization with ESBL or resistance for any of the antibiotic groups is endemic, with proportions of ICU-acquired bacteraemias used as a proxy. Therefore, the investigators prefer proportions of AMRB infection in the period 2008-2009 to be: ESBL resistance among GNB 1 to 10% Piperacillin/Tazobactam among GNB 1 to10% Carbapenem resistance among Klebsiella Pneumoniae less than 5%
• Have the ability of at least one dedicated Infection Control HCW available for 0,2fte, for patient monitoring, compliance monitoring and instruction of HCWs regarding interventions. In the following this person will be called "Research-Nurse" or "RN".
• Can store screening-cultures at -70ºC
• Can facilitate transport through a UMCU courier.
• There is written approval for the study from the institution's IRB with a waiver for patient informed consent.
• A signature page is signed by the daily management of the candidate-ICU by both ICU physician and director and the ICU nursing-director and presented to the UMCU, indicating willingness to enroll the candidate-ICU in the study.

Exclusion Criteria:

ICUs planning to introduce, during the SATURN trial period, any major diagnostic- or intervention program that will affect AMRB ecology*

• Burn units; due to the specific nature of the care provided and the patients admitted.
• Cardiothoracic surgery units; because of the expected small number of patients admitted for three days or more.
• Paediatric and neonatal ICUs.