Aetiology of community-acquired pneumonia: a prospective study among adults requiring admission to hospital.

BACKGROUND The prevalence of microorganisms causing community-acquired pneumonia in patients who required admission to hospital was investigated and the percentage of cases whose aetiology remained unknown due to the study design and logistical problems estimated. METHODS Between January 1991 and April 1993 all patients with community-acquired pneumonia admitted to six hospitals were included in the study. Aetiological diagnosis, categorised as definite, probable and possible, was based on the results of routine microbiological and serological tests. RESULTS Three hundred and thirty four patients with a median age of 65 (range 17-92) years were enrolled in the study. The diagnosis of community-acquired pneumonia was definite in 108 cases, and probable or possible in 73 and 27 cases, respectively, including dual infections. Streptococcus pneumoniae was the predominant pathogen (27%) followed by viruses and Haemophilus influenzae (both about 8%) and Mycoplasma pneumoniae (6%). Chlamydia spp (3%) and Legionella pneumophila (2%) were less frequently detected. No diagnosis was made in 45% of the cases. With adjustment for anti-microbial therapy before admission and for other logistical considerations, it is estimated that the aetiology could have been ascertained in 65% of the cases. CONCLUSIONS Streptococcus pneumoniae is the most frequently detected cause of community-acquired pneumonia. The inability to detect a micro-organism results mainly from the use of routine diagnostic tests and, to a lesser extent, from logistical problems or the use of antibiotics before admission.