• Background: Ventilator-acquired pneumonia (VAP) is the most prevalent nosocomial infection in intensive care units (ICUs). Early microbiological diagnosis and initial administration of appropriate antimicrobial therapy are associated with a better outcome. Broad-spectrum antibiotics should therefore be administered initially. However, inconsiderate antibiotic use can increase the prevalence of multi-resistant bacteria.

• Purpose: A rapid antimicrobial susceptibility method is required to decrease the unnecessary use of empirical broad-spectrum antibiotics. The aim of this study is to compare the efficiency of a rapid antibiogram, provided by E-test strips directly applied to bronchoalveolar lavage (BAL) samples and analysed at 24 h, to that obtained with standard methods of culture which provide a later result.

• Study design: This will be an open-label, prospective cohort study of consecutive patients with VAP, conducted in a medical ICU. In addition to standard culture methods, an E-test will be performed directly on BAL samples and analysed at 24 h. Each standard BAL culture will be used as a control for the E-test method.

• Primary outcome: The occurrence of major errors, defined as isolates determined to be susceptible by the E-test but resistant by standard culture methods.

• Secondary outcomes: The occurrence of minor errors (defined as isolates determined to be resistant by the E-test and susceptible by the standard method), and a comparison of two methods of seeding BAL samples on Mueller Hinton agar plates (swabbing method, flooding method).

• Eligibility criteria:

  • Inclusion criteria: all patients with suspected VAP (defined by a Clinical Pulmonary Infection Score ≥5) undergoing BAL will be eligible.
  • Exclusion criteria: contraindications for BAL (PaO2/FIO2 <100, risk of bronchoscopy-related haemorrhagic complications), secondary exclusion of patients with negative cultures, defined by a threshold of bacteria <104 CFU/ml.

• Interventions:

BAL samples will be cultured by standard methods and the minimal inhibitory concentration (MIC) of bacteria to the usual antibiotics will be determined using standard procedures. At the time of BAL collection, a rapid antibiogram will be performed by placing E-test antibiotic strips (AB Biodisk) directly onto Mueller-Hinton agar plates seeded with the BAL specimen (both by flooding and swabbing). E-test strips will be impregnated with cefoxitin, piperacillin-tazobactam, cefepime, imipenem, ciprofloxacin and amikacin. At 24 h, the E-test plates will be photographed and then examined separately by both a bacteriologist and a medical ICU physician following a consensus method. The final E-test results will be compared with the standard MIC cultures.

Inclusion Criteria:

• all patients with suspected VAP (defined by a Clinical Pulmonary Infection Score ≥5) undergoing BAL will be eligible.

Exclusion Criteria:

• contraindications for BAL (PaO2/FIO2 <100, risk of bronchoscopy-related haemorrhagic complications), secondary exclusion of patients with negative cultures, defined by a threshold of bacteria <104 CFU/ml.