SESSION TITLE: ICU Infections Posters
SESSION TYPE: Original Investigation Poster
PRESENTED ON: Wednesday, October 30, 2013 at 01:30 PM - 02:30 PM
PURPOSE: The focus of sepsis therapy is early initiation of broad-spectrum antibiotics. However, attempting to balance between adequate coverage and prevention of antibiotic overuse in the ICU is an ongoing challenge. Current recommendations focus on early antibiotics to cover resistant organisms and then narrowing based on culture results. This often leads to 48-72 hours of broad antibiotic coverage for all patients with suspected sepsis. We looked at the feasibility of using Gram stain results from either endotracheal aspirate (ETA) or bronchoalveolar lavage (BAL) to facilitate rapid narrowing of antibiotic coverage.
METHODS: This study is a retrospective cohort of all patients admitted to the ICU between 12/25/09 and 10/20/11 with both an ETA and BAL bacterial culture processed within 48 hours of each other. In the case of multiple specimens processed in the same 48-hour period, only the first paired sample was considered. Cultures growing only oral flora at <10,000 cfu were considered negative. ETA and BAL gram stain were then compared by McNemar’s test based on with final BAL culture to determine feasibility of antibiotic de-escalation.
RESULTS: A total of 324-paired samples were considered. 238 (73.4%) patients had at least one positive specimen (Gram stain or culture). ETA and BAL Gram stains were positive in 108 patients (33.3%) and 115 (35.4%) of patients respectively. BAL was culture positive on 109 patients (33.3%). Looking at early de-escalation, an ETA Gram stain had a negative predictive value (NPV) of 84.2% and a BAL Gram stain had a NPV of 87.3%. For ETA, de-escalation based on Gram stain would have missed 38% of MRSA and 31% of aerobic Gram negatives grown on BAL culture. BAL Gram stain detected 92% of MRSA isolates, but only 75% of potentially resistant Gram negatives.
CONCLUSIONS: BAL Gram stain was significantly better than ETA for detection of MRSA (p=0.01). However, Gram-negative coverage cannot be safely de-escalation based on Gram stain results.
CLINICAL IMPLICATIONS: De-escalation of MRSA coverage can be done safely based on BAL Gram Stain.
DISCLOSURE: The following authors have nothing to disclose: Edward Haas, Jeremy Taylor, Naeem Ali, Matthew Exline
No Product/Research Disclosure Information