1487. Pneumonia due to Co-Infection in the ICU: Detection and Clinical Significance

Background Pneumonia is a significant cause of morbidity and mortality, with increasing interest in the detection and clinical significance of co-infection. However, the impact of methodology to obtain lower respiratory samples along with the utility of various microbiological diagnostic testing remains unclear. Methods A single-center retrospective analysis was performed on bronchoalveolar lavage (BAL) samples obtained from mechanically ventilated adults treated in critical care units from August 2012 to December 2017. BAL methodology (bronchoscopic vs blinded), microbiological diagnostic testing, and outcomes measures were obtained. Associations between categorical variables were assessed using Chi-Square or Fisher’s exact tests. Kruskal Wallace tests analyzed differences in distributions of measures between categories based on number of organism types detected. SAS software version 9.4 (SAS Institute Inc., Cary, NC). Results Analysis of the 803 samples that met inclusion criteria found a significant linear association between mortality and number of organism types detected by BAL, with 30 day mortality rates of 43.0%, 47.8%, and 58.3% among those with zero, one, and two or more organisms respectively (p = 0.003). Comparing BALs with at least one organism isolated, the detection of viruses specifically was associated with increased mortality, with the presence and absence of viral organisms corresponding to 56.3% and 46.5% mortality at thirty days (p = 0.03). No association was found between mortality and isolation of acid-fast bacilli, bacteria, or fungi. Co-infection was detected more frequently among bronchoscopic BALs than blinded BALs (26.3% vs 8.6%, p < 0.0001), with more viruses detected bronchoscopic BALs (41.9% vs 13.1%, p < 0.0001), and more bacteria in blinded BALs (41.8% vs 33.0%, p = 0.01). 30 Day Mortality vs Isolation of Specific Organism Types from BAL Number of Organism Types Isolated from BAL Compared to BAL Methodology BAL Methodology vs Isolation of Specific Organism Types Conclusion Co-infection in mechanically ventilated adult patients with pneumonia appears to be a significant risk factor for mortality, with the detection of viral organisms potentially playing an independent role. Within this population, bronchoscopic BALs may have a valuable diagnostic and prognostic methodology. Disclosures All Authors: No reported disclosures