2140. Utility of Respiratory Specimen Gram Stain for Predicting Final Culture Result in Patients with Clinically Diagnosed Pneumonia

BACKGROUND: Obtaining a high-quality respiratory tract specimen for Gram stain and culture in patients with suspected lower respiratory tract infections is recommended by the IDSA guidelines. However, conflicting results correlating Gram stain with final culture growth has led to questions about the utility of a respiratory specimen Gram stain. The purpose of this study was to assess the correlation of Gram stain with final culture in patients with pneumonia. METHODS: A retrospective chart review was conducted to evaluate adult inpatients with a diagnosis of pneumonia (based on the CDC surveillance definition) who had a respiratory specimen submitted for Gram stain and culture. A specimen was considered acceptable if less than ten epithelial cells were visualized under low power field. Each Gram stain was compared with the corresponding final culture. The primary outcome was to evaluate the correlation of Gram stain with final culture using positive predictive value (PPV), negative predictive value (NPV), sensitivity, and specificity. A culture was considered negative if no bacteria were isolated or if only normal flora grew. Secondary outcomes were PPV and NPV based on antibiotic exposure prior to specimen collection, semi-quantitative number of bacteria on Gram stain, and collection method. Additionally, discordance between Gram stain and final culture morphology was evaluated. RESULTS: A total of 269 acceptable specimens were assessed. Of the 72 specimens with a positive Gram stain, 41 yielded bacteria in final culture (PPV: 56.9%). In contrast, 154 of the 197 specimens with a negative Gram stain were associated with negative final culture (NPV: 76.7%). The NPV of Gram stain was decreased when antibiotics were given for > 24 hours pre-specimen. The PPV of Gram stain improved as an increasing amount of bacteria were reported. Less invasive collection methods had a lower PPV but a higher NPV in comparison to invasive collection methods. Finally, the discordance rate between Gram stain and final culture morphology was low. CONCLUSION: This study shows inconsistent results regarding the ability of Gram stain to predict final culture. Pneumonia should continue to be managed clinically and caution taken prior to adjusting empiric antimicrobial regimens based solely on the Gram stain. DISCLOSURES: All authors: No reported disclosures.