The study is a prospective, randomized, controlled intervention trial conducted in 9 centers, comparing a conventional strategy versus a PCT-guided strategy to start or to discontinue antibiotics, in patients with suspected community or hospital- acquired infection.

Clinical and laboratory signs are neither specific nor sensitive for diagnosis of sepsis in critically-ill patients. Because delaying antimicrobial therapy may be deleterious, broad-spectrum antibiotics are widely used in ICU -patients, even when they are not needed. In addition, only few well-designed studies concerning the duration of antibiotic treatment have been so far published. Consequently, many patients received antibiotics during the ICU stay. Many studies have shown that exposure to antibiotics, the so called "selection pressure" is an independent risk factor for acquisition of resistance in individual patients. Therefore, reducing antibiotic use is probably necessary to control antibiotic resistance. Many clinical studies have shown that procalcitonin (PCT) is able to distinguish the inflammatory response to infection from other types of inflammation and to distinguish bacterial from viral infections. Recent studies have shown that PCT guidance substantially and safely reduced antibiotic overuse in patients with lower respiratory tract infections. We aimed to evaluate the role of PCT in reducing the use of antibiotics in ICU adult patients. The study is a prospective, randomized, controlled intervention trial conducted in 9 centers, comparing a conventional strategy versus a PCT-guided strategy to start or to discontinue antibiotics, in patients with suspected community or hospital- acquired infection.

Inclusion Criteria:

• Patient hospitalised in the ICU
• Bacterial infection suspected
• At ICU admission the patient do not receive antibiotics or receive antibiotics for less than 24hours and an interval between admission and inclusion < 12 hours
• During ICU stay, provided that the interval between the start of suspected infection and inclusion is < 12hours
• Written inform consent from the patient or relatives. The consent may be obtained after the enrollment if the patient is not able to give consent and if there is no relatives

Exclusion Criteria:

• Age < 18 years
• Pregnancy
• Patient expected to remain hospitalised in the ICU for less than 3 days
• Neutropenia
• Infection or presumed infection requiring prolonged antibiotic therapy (endocarditis,osteo-articular infection, mediastinitis, deep abscess, tuberculosis, pneumocystis pneumonia, toxoplasmosis).
• Simplified Acute Physiology Score II at ICU admission (calculated during the first 12h)
• Attending physician declining to use full life support.