Ventilator associated Pseudomonas aeruginosa pneumonia: The clinical and microbiological response to antibiotic therapy

Objective: To assess the effects of antibiotic therapy on the symptomatology, microbiology and outcome from Pseudomonas aeruginosa ventilator associated pneumonia (VAP). Design: A prospective observational study Setting: A 12-bedded medical/surgical, adult Intensive Care Unit within a University based, tertiary referral hospital. Subjects: Fifty consecutive, intubated, ventilated patients with clinical and microbiological characteristics of P. aeruginosa VAP. Interventions: All patients received 5 days of conventional systemic, anti-pseudomonal antibiotics including aminoglycosides. Fifteen patients also received inhaled colistin. Endpoints: The diagnosis of pneumonia was based on achieving threshold levels of a clinical pulmonary infection score (CPIS) combined with a threshold colony count (≥ 10⁴ colony-forming units [CFU] per ml of lavage fluid) of P. aeruginosa from NBL fluid. ICU outcome, prevalence of sepsis, septic shock, organ failure and length of stay in ICU were recorded for each patient. Measurements: The admission APACHE II, predicted mortality and Charlson Comorbidity score was calculated for each subject on admission to ICU. Subjects underwent alternate day, non-directed bronchial lavage (NBL) with quantitative culture and daily estimation of CPIS. Clinical and biochemical variables were recorded to identify sepsis, septic shock and organ failure based on consensus criteria. Main results: The overall mortality for the 50 patients was 54%. Patients who died had a significantly higher APACHE II predicted mortality (p<0.02 ANOVA). Forty-one patients (82%) had sepsis syndrome and 31 of those were shocked. Twenty-eight patients (56%) had multiple organ failure. CPIS changed significantly with time, falling with a linear trend. There was no effect of CPIS either overall or as an interaction with time on outcome. P. aeruginosa colony counts in NBL fluid also changed significantly with time, though with a quadratic polynomial trend, and again there was no significant effect on eventual outcome. Conclusions: The trends in CPIS or NBL colony counts in response to therapy cannot be used to predict survival. The reappearance of P. aeruginosa in lung lavage during the course of antibiotic therapy does not mitigate against recovery.