LAUSR

Abstract Background Procalcitonin (PCT) is a biomarker that is finding increasing diagnostic and prognostic utility in lower respiratory infections. It remains unclear, however, whether it can be helpful in predicting the bacterial etiology of pneumonia, with a view to informing antibiotic choice and duration. This study examines the relationship between serial PCT measurements and microbial etiology in patients hospitalized for pneumonia to determine whether changes in PCT levels provide discriminatory information on microbial etiology. Methods We performed a subgroup analysis of data from a prospective cohort study of 505 patients admitted to a tertiary care center with findings concerning for pneumonia. Microbial etiology of pneumonia was determined from high quality respiratory samples, blood cultures or other relevant diagnostic tests according to standard protocols. Procalcitonin levels were measured serially during the first four days of hospitalization. We compared procalcitonin levels between different bacterial etiologies over the first four days of admission, using the Mann–Whitney-U test to assess for statistical significance. Results Out of 505 patients, the diagnosis of pneumonia was adjudicated in 317, and bacterial etiology determined in 62 cases. The predominant pathogens were Staphylococcus aureus (N = 18), Streptococcus pneumoniae (N = 6), Pseudomonas aeruginosa (N = 11) and Haemophilus influenza (N = 5). Admission levels of PCT were lowest in Pseudomonas infections and highest in pneumococcal infections, though not reaching statistical significance. On hospital days two and three, pneumococcal procalcitonin levels were significantly higher than all other etiologies, but on day four, there was no statistically significant difference in PCT values for different microbial etiologies. Conclusion Serial procalcitonin levels during the early course of bacterial pneumonia reveal a difference between pneumococcal and other bacterial etiologies, and may have an adjunct role in guiding antibiotic choice and duration. Disclosures All authors: No reported disclosures.