Bacterial infections are a frequent complication in liver cirrhosis with a bad prognosis. However, the clinical outcome of cirrhotic patients with serious infections is significantly improved over the last 30 years due to early diagnosis and to the use of a more appropriate antibiotic therapy.

As in the general population, empirical treatment should be initiated soon after diagnosis, after making the necessary sampling and should be based on the use of an antibiotic with low toxicity and broad spectrum antibacterial efficacy, taking into account the local epidemiology and prevalence of antibiotic resistance.

The third generation cephalosporins are considered the gold standard in the treatment of most infections in cirrhotics due to their effectiveness against enterobacteriaceae and against non-enterococcal streptococci and due to their low toxicity.

However, the recommendations for the antibiotic therapy are currently based on results of trials of '80s and '90s, when the proportion of resistant pathogens was lower. Similarly to nosocomial infections, the increasing rate of infections due to multidrug resistant (MDR) bacteria represents the rational for a different choice of empirical antibiotic therapy with a higher resistance barrier. This change in the epidemiology of community acquired infections is mainly due to the increased contacts with healthcare system of these patients and for the larger use of antibiotic prophylaxis. With this regard, it was recently proposed to introduce a third epidemiological class of infection "Health care-associated": Infections occurring in community in patients who have been in contact with the health system shortly before the infection.

This epidemiological class of infection should be distinguished from "community-acquired" because they are more similar to"nosocomial" in terms of their sensitivity to antibiotics. Therefore for this class should be taken into consideration the use of a different empirical antibiotic therapy.

The investigators aim was to evaluate prospectively the effectiveness of a broad spectrum antibiotic treatment in a cohort of cirrhotic patients with "healthcare-Associated"infections

Cirrhotic patients with "Healthcare Associated" admitted to the gastroenterology department of the Policlinico Umberto I and in the Department of Hepatology of the Hospital of Marino will be consecutively enrolled.

Randomized controlled trial with randomisation stratified by epidemiological class of infection.

Unblinded randomized controlled trial with randomisation stratified by epidemiological class of infection.

After obtaining informed consent at enrollment, participants will undergo medical examination with a detailed anamnesis and will begin collecting data on the following parameters:

• Demographic characteristics (age, sex, etiology of cirrhosis, when the diagnosis was made).
• Severity of liver disease(defined by the Child-Pugh classification, MELD score), indices of renal function and serum electrolytes.
• APACHE score in the case of development of sepsis
• Compensated cirrhosis / decompensated cirrhosis
• If decompensated, will be listed complications (ascites, esophageal-gastric varices, hepatic encephalopathy, coagulopathy, Hepato-Renal Syndrome, variceal bleeding).
• Treatment and dosage (eg β-blockers, diuretics)
• Hemodynamic parameters (heart rate / blood pressure)

Patients enrolled will be randomized and treated with conventional therapy provided for community-acquired infections or with the "2nd line antibiotic therapy", based on use of of imipenem-cilastatin.

The protocol of antibiotic therapy chosen for this category of patients will be based on a antibiotic class recommended for a specific site of infection in international guidelines for the treatment of nosocomial infections and / or healthcare associated with a high safety profile and is already used in the investigators hospital. The cost of empirical treatment with imipenem / cilastatin is roughly comparable to that with the currently used first-line antibiotics for such infections, and frequently used drugs in the forefront of these patients are ineffective is therefore necessary to embark on a new antibiotic with higher costs also in terms of days of hospitalization.

During the hospitalization will be recorded the type of infection, empirical therapy, its duration, side effects, isolated pathogens and their resistance, when available, the occurrence of complications related to infection and mortality.

The efficacy of therapy will be assessed based on the resolution of infection attested by normalization of clinical , biochemical and microbiological parameters and the instrumental tools.

Inclusion Criteria:

• cirrhosis
• healthcare infection
• older than 18 years
• agreement to participate

Exclusion Criteria:

• HIV
• patients underwent to liver transplantation