Send to

Choose Destination
See comment in PubMed Commons below
Infection. 2007 Jun;35(3):143-9.

A prediction model for bacterial etiology in acute exacerbations of COPD.

Author information

Helios Klinikum Emil von Behring, affiliated Free University Berlin, Chest Hospital Heckeshorn (Infectious Disease and Immunology), Zum Heckeshorn 33, 14109, Berlin, Germany.



Bacteria play a leading role in acute exacerbations of chronic obstructive pulmonary disease (COPD), but we lack predictors of bacterial etiology. We developed a prediction model for infection with gram-negative enteric bacteria (GNEB) and Pseudomonas aeruginosa.


Clinical presentation, sputum characteristics, microbial sputum patterns, lung function and previous and concomitant medication were prospectively recorded in patients with moderate to severe exacerbation of COPD. Risk factors for a specific bacterial etiology were calculated and a prediction model developed.


A total of 193 patients with acute exacerbation were included. In 121 (62.6%) of them a microbial etiology could be identified, most frequently Haemophilus influenzae (32 strains), Streptococcus pneumoniae (22 strains) and P. aeruginosa (12 strains). Multivariate analysis identified severe airflow obstruction and use of systemic steroids as predictors for exacerbation due to gram-negative enteric bacilli and P. aeruginosa. A prediction model including FEV1 < 35% of predicted value, systemic steroid use and prior antibiotic therapy within preceeding 3 months had a negative predictive of 89%, being a helpful tool in excluding patients at risk of exacerbation due to gram-negative enteric bacilli and P. aeruginosa when all criteria are absent.


A simple prediction model based on three factors may identify COPD patients at low risk for exacerbations with gram-negative enteric bacilli and P. aeruginosa. Bacterial Etiology in COPD Exacerbations.

[Indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Springer
    Loading ...
    Support Center