1498. Etiology of Community-acquired Pneumonia in Adults: A Systematic Review

BACKGROUND: Background: Recent guidelines recommend immediate empiric antibiotic treatment for patients (pts) with community-acquired pneumonia (CAP). Concerns about treatment recommendations and antibiotic stewardship motivated a systematic literature review of the etiology of CAP. METHODS: We reviewed English-language literature using PRISMA guidelines. Data were stratified into diagnostic categories according to the microbiologic studies that were done (Table1). Fig.1. Flowchart of systematic literature review and study selection [Image: see text] Table 1. Characteristics of studies reporting the etiology of community-acquired pneumonia [Image: see text] RESULTS: 146 articles with 82,674 CAP pts met criteria for inclusion; 63,938 (77.3%) were inpatients, 16,532 (20.0%) were in- or outpatients, and 2,204 (2.7%) were outpatients. Pneumococcus was the most common cause of CAP without regard to which microbiological techniques were used (33-50% of all cases). The proportion due to this organism declined with time, much more strikingly in the US than in Europe. Haemophilus influenzae was the second most common cause (7-16% of cases), followed by Staphylococcus aureus and Enterobacteriaceae each in 4–10%. Pseudomonas (0.8-4.5%) and Moraxella (1.2-3.5%) were less common; all other bacteria were isolated far less frequently. Mycoplasma caused 4-11% of CAP, Legionella 3-8%, Chlamydophila 2-7%, and Coxiella < 2%; some studies showed a much higher frequency of Mycoplasma. With routine use of viral PCR, a virus was identified in 30-40% of pts; bacterial/viral coinfection was found in 25-35% of these cases. In a separate study of CAP pts in whom viral PCR was positive, 40% had bacterial coinfection. Influenza viruses were identified in 6.2-13.7% of cases and rhinoviruses in 4.1-11.5%. RSV and human metapneumovirus were less common (0.4-4.7%), followed more distantly by other viruses. Even with the use of the most sophisticated diagnostic techniques, no etiologic agent for CAP was identified in > 50% of cases. Trends of identification of S. pneumoniae and H. influenzae as the etiology of CAP (above); and the proportion of S. pneumoniae as the causes of CAP in different geographic regions (below). [Image: see text] CONCLUSION: Our results justify current guidelines for initial empiric antibiotic treatment of all pts with CAP. With pneumococcus and Haemophilus continuing to predominate, efforts at antibiotic stewardship might be enhanced by greater attention to routine use of sputum Gram stain and culture. Because viral/bacterial coinfection is relatively common, the identification of a virus by PCR does not, by itself, permit the non-use of an antibiotic. DISCLOSURES: All Authors: No reported disclosures