Early Antibiotic Discontinuation in Patients With Clinically Suspected Ventilator-Associated Pneumonia and Negative Quantitative Bronchoscopy Cultures

Objectives: Preliminary data suggest that antibiotic discontinuation in patients with negative quantitative bronchoscopy and symptom resolution will not increase mortality. Because our hospital algorithm for antibiotic discontinuation rules out ventilator-associated pneumonia in the setting of negative quantitative bronchoscopy cultures, we compared antibiotic utilization and mortality in empirically treated, culture-negative ventilator-associated pneumonia patients whose antibiotic discontinuation was early versus late. Design: Retrospective, observational cohort study. Setting: Eight hundred sixty-seven bed, tertiary care, teaching hospital in Hartford, CT. Patients: Eighty-nine patients with clinically suspected ventilator-associated pneumonia and a negative (<10(4) colony forming units/mL) quantitative bronchoscopy culture between January 2009 and March 2012. Early discontinuation patients (n = 40) were defined as those who had all antibiotic therapy stopped within one day of final negative culture report, whereas late discontinuation patients (n = 49) had antibiotics stopped later than one day. Measurements: Univariate analyses assessed mortality, antibiotic duration, and frequency of superinfections. Multivariate logistic regression was performed to assess the effect of early discontinuation on hospital mortality. Results: Patients had a mean +/- sd Acute Physiology and Chronic Health Evaluation II score of 26.0 +/- 6.0. Mortality was not different between early discontinuation (25.0%) and late discontinuation (30.6%) patients (p = 0.642). Antibiotic duration (days) was also not different for patients who died vs. those who survived (Median [interquartile range]: 3 [1-7.5] vs. 3 [1.75-6.25], respectively, p = 0.87), and when controlling for baseline characteristics and symptom resolution, only Acute Physiology and Chronic Health Evaluation II score was associated with hospital mortality on multivariate analyses. There were fewer superinfections (22.5% vs. 42.9%, p = 0.008), respiratory superinfections (10.0% vs. 28.6%, p = 0.036), and multidrug resistant superinfections (7.5% vs. 35.7%, p = 0.003), in early discontinuation compared with late discontinuation patients. Conclusions: In this severely ill population with clinically suspected ventilator-associated pneumonia and negative quantitative bronchoalveolar lavage cultures, early discontinuation of antibiotics did not affect mortality and was associated with a lower frequency of MDR superinfections.