C-reactive protein levels in community-acquired pneumonia

This is the authors' abstract. We don't add key points for this paper.

The diagnostic value of C-reactive protein (CRP) admission serum levels as an indicator of the aetiology of community-acquired pneumonia (CAP) was evaluated. A cohort of 1,222 patients with CAP was assessed. CRP levels were analysed in 258 patients with a single aetiological diagnosis. The mean CRP values in patients with pyogenic, atypical, viral and Legionella pneumophila pneumonia were: 16 mg x dL(-1), 13 mg x dL(-1), 14 mg x dL(-1) and 25 mg x dL(-1), respectively. CRP levels were not significantly different among patients outcome research team (PORT) groups (19 mg x dL(-1) in groups I-II, 16 mg x dL(-1) in group III and 16 mg x dL(-1) in groups IV-V). A cut-off point of 25 mg x dL(-1) had a sensibility, specificity, positive predictive value and negative predictive value of 0.6, 0.83, 0.3, and 0.94, respectively. After controlling for age and PORT score, the odds of having a CRP level >25 mg x dL(-1) was 6.9 times higher in patients with L. pneumophila pneumonia than in those with non-L. pneumophila pneumonia. Patients with Legionella pneumophila pneumonia had higher C-reactive protein levels than those with pneumonia of any other aetiology, independently of severity of infection. Being a cheap and readily available test, C-reactive protein may be a useful adjunctive procedure in the diagnosis of community-acquired pneumonia.