Clinical features not sufficiently pathognomonic in pneumonia

By Joanna Lyford, Senior medwireNews Reporter

Clinical signs and symptoms are not sufficiently reliable to distinguish the different etiologies of pneumonia, Dutch researchers report.

Thus, patients with community-acquired pneumonia (CAP) should be routinely tested for all common respiratory pathogens to enable appropriate treatment.

At present, patients with CAP are typically evaluated using a relatively limited range of bacterial and viral culture or antigen assays – a costly strategy that is not always informative.

In this study, Elisabeth Huijskens (Albert Schweitzer Hospital, Dordrecht) and team used extensive molecular testing of 263 patients with CAP to classify the infection as purely bacterial (n=146), purely viral (n=52), or a mixture of viral and bacterial (n=65).

All patients were also assessed for a comprehensive range of signs and symptoms and for disease severity using the pneumonia severity index (PSI).

The most prevalent pathogens were Streptococcus pneumoniae, Coxiella burnetii, human rhinovirus, and influenza viruses A, the researchers report. In multivariate analysis, most clinical indicators did not differ significantly among the three CAP etiologies, although there were some exceptions.

Specifically, the probability of a viral etiology increased with advancing patient age (odds ratio [OR]=1.007 per 1-year increment) and with the presence of cough (OR=5.536). Meanwhile, patients with immunodeficiency were more likely to have a mixed etiology (OR=2.323) and less likely to have a bacterial or viral etiology (OR=0.595 and 0.816, respectively).

Also, patients with a purely bacterial etiology tended to have higher levels of C-reactive protein than those with purely viral or mixed etiologies (219 vs 78 and 103 mg/dL, respectively).

The PSI also differed according to etiology, with each 1-point increase being associated with an increased likelihood of finding a viral (OR=1.011) or mixed etiology (OR=1.012). As might be expected, mean PSI was significantly higher among patients with any form of infection than in the 145 patients with no detectable pathogen (105 vs 97 points).

Additional analyses found no difference in mortality between patients with bacterial versus viral infection. Furthermore, there was no interdependence of viral and bacteriologic etiology, meaning that the risk for bacterial infection did not change with the absence versus presence of viral infection, and vice versa.

Writing in the Journal of Medical Microbiology, the researchers say that they found few clinical characteristics that could accurately discern the etiology of CAP. While age, cough, PSI, and immunodeficiency were significant independent predictors of etiology, there was substantial overlap in clinical presentation.

“In summary, in this study no reliable clinical predictors could be identified to guide microbiological evaluation,” they conclude. “Our findings suggest that routine testing for common respiratory pathogens is warranted for all adults with CAP.”

Licensed from medwireNews with permission from Springer Healthcare Ltd. ©Springer Healthcare Ltd. All rights reserved. Neither of these parties endorse or recommend any commercial products, services, or equipment.