Correlation of the bacterial aetiology with initial procalcitonin levels and duration of antibiotic therapy in lower respiratory tract infections

Abstract number: O564

Künzli E., Woitzek K., Dusemund F., Schuetz P., Müller B., Albrich W.

Objectives: Procalcitonin (PCT) has been established as a reliable and efficient marker for the differentiation between viral and bacterial lower respiratory tract infections (LRTI) and to guide antibiotic therapy. There is conflicting data about a correlation of higher PCT levels with Gram-negative bacteraemia. In the current analysis we correlate the bacterial aetiology with initial PCT-levels and the duration of antibiotic therapy in LRTI.

Methods: Initial PCT levels of patients hospitalized with LRTIs as part of the Swiss multicenter randomized controlled ProHOSP Study were correlated with the identified bacterial aetiology and total antibiotic duration in survivors. PCT was measured using the highly sensitive Kryptor® (BRAHMS) assay. Bacteria were identified using standard microbiological cultures from blood and sputum and urinary antigen tests for pneumococcal and legionella (Binax®). Statistical analysis was performed using Mann–Whitney U-Test and Spearman correlation coefficient.

Results: The current analysis includes 388 patients with LRTIs (age 70.5 + 16.7 years) from 2 of the 6 ProHOSP hospitals. Streptococcus pneumoniae (45; 23 of those bacteraemic) was the most common known aetiology, followed by legionella (8), and Enterobacteriaceae (5). Other Gram-positive organisms were identified in 12, other Gram-negative organisms in 6 and in 306 patients no causative organisms could be detected. LRTI caused by legionella and Enterobacteriaceae showed similar initial PCT values (median 1.46, p = 0.31; and 4.09, p = 0.80, respectively) as pneumococcal infections. PCT values were higher for pneumococcal LRTI (median 9.10mg/L) than for other Gram-negative organisms (PCT median 0.30, p = 0.04), other Gram-positive organisms (PCT median 0.40, p = 0.01) and if no organisms were identified (PCT median 0.25, p < 0.001). Total antibiotic treatment duration was guided according to PCT kinetics based on predefined cut-off ranges.

Conclusion: Initial PCT values differed according to the aetiology of LRTIs with more virulent organisms including pneumococci being associated with higher levels. Considerable overlap in mean PCT values precluded prediction of the aetiology based on PCT values alone. However, combining knowledge of initial PCT values and causative organisms might predict the required antibiotic duration. At the meeting, we will present the data of the entire 1359 patients of all 6 ProHOSP hospitals.

Session Details

Date: 10/04/2010
Time: 00:00-00:00
Session name: Abstracts 20th European Congress of Clinical Microbiology and Infectious Diseases
Location: Vienna, Austria, 10 - 13 April 2010
Presentation type:
Back to top