|| Checking for direct PDF access through Ovid
To assess the rate of appropriateness of empirical antimicrobial therapy for ventilator-associated pneumonia, to evaluate de-escalation in patients with ventilator-associated pneumonia treated according to local pathway, and to identify the bacteria responsible for recurrence of ventilator-associated pneumonia.Prospective observational study during a 36-month period.Medical-surgical intensive care unit of a university hospital.One hundred and fifteen patients hospitalized in an intensive care unit developing ventilator-associated pneumonia with positive cultures. The patients with ventilator-associated pneumonia were treated with limited-spectrum antibiotics (i.e., without activity against Pseudomonas aeruginosa) if they had no prior hospitalization (within 21 days) or prior administration of antibiotics (within 10 days). Quantitative cultures obtained by bronchoscopy or tracheal aspiration were used to reassess empirical therapy.None.A limited-spectrum therapy was used in 79 patients (69%). Empirical antimicrobial therapy was appropriate in 100 patients (85%). The mortality rate was significantly higher in the patients in whom empirical therapy was inappropriate than in those in whom treatment was appropriate (47 vs. 20%, p = .04). De-escalation was done in respectively 26% and 72% of patients with early- and late-onset ventilator-associated pneumonia, whereas treatment was escalated in 27 patients (23%). Ventilator-associated pneumonia episodes were recurrent in 22 cases, including eight episodes due to high-risk bacteria.A rational empirical antimicrobial therapy for ventilator-associated pneumonia using limited-spectrum antibiotics is possible if local ecology and patient medical history and clinical status are considered. In addition, de-escalation is feasible in 42% of patients. This integrative approach may reduce the emergence of resistant bacteria, which in turns reduces the need for broad-spectrum antibiotics, breaking the vicious circle of antibiotic overuse.