Can clinical findings at admission allow withholding of antibiotics in patients hospitalized for community acquired pneumonia when a test for a respiratory virus is positive?

Background: Current guidelines recommend empiric antibiotic therapy for patients who require hospitalization for community-acquired pneumonia (CAP). We sought to determine whether clinical, imaging or laboratory features in patients hospitalized for CAP in whom PCR is positive for a respiratory virus enable exclusion of bacterial coinfection so that antibiotics can be withheld.

Methods: For this prospective study, we selected patients in whom an etiologic diagnosis was likely to be reached, namely those who provided a high-quality sputum sample at or shortly after admission, and in whom PCR was done to test for a respiratory virus. We performed quantitative bacteriologic studies on sputum to determine the presence of bacterial infection or coinfection and reviewed all clinical, imaging and laboratory studies.

Results: Of 122 CAP patients studied, 77 (63.1%) had bacterial infection, 16 (13.1%) viral infection, and 29 (23.8%) bacterial/viral coinfection. Underlying pulmonary disease and a history of smoking were more common in bacterial pneumonia. Upper respiratory symptoms were more common, and mean white blood cell (WBC) counts were lower viral pneumonia. Nevertheless, no clinical, laboratory or imaging findings allowed exclusion of bacterial coinfection in patients who tested positive for a respiratory virus. In fact, patients with bacterial/viral coinfection were sicker than those with bacterial or viral pneumonia; 30% were admitted required transfer to the ICU during their hospital course, compared to 17% and 19% of patients with bacterial or viral infection, respectively (p < .05). In this study, 64.4% of patients who tested positive for a respiratory virus had a bacterial coinfection.

Conclusions: If a test for a respiratory virus test is positive in a patient hospitalized for CAP, no sufficiently differentiating features exclude bacterial coinfection, thereby supporting the recommendation that empiric antibiotics be administered to all patients who are sufficiently ill to require hospitalization for CAP.