Research article

Diagnostic and prognostic accuracy of clinical and laboratory parameters in community-acquired pneumonia

Beat Müller¹ , Stephan Harbarth² , Daiana Stolz³ , Roland Bingisser⁴ , Christian Mueller¹ , Jörg Leuppi³ , Charly Nusbaumer⁵ , Michael Tamm³ and Mirjam Christ-Crain¹

¹  From the Departments of Internal Medicine, University Hospital, Petersgraben 4, CH-4031, Basel, Switzerland

²  Division of Hospital Epidemiology, University Hospital, CH-1211, Geneva, Switzerland

³  Department of Pneumology, University Hospital, Petersgraben 4, CH-4031, Basel, Switzerland

⁴  Emergency Department, University Hospital, Petersgraben 4, CH-4031, Basel, Switzerland

⁵  Department of Cinical Chemistry, University Hospital, Petersgraben 4, CH-4031, Basel, Switzerland

author email corresponding author email

BMC Infectious Diseases 2007, 7:10doi:10.1186/1471-2334-7-10

Published:	2 March 2007

Abstract

Background

Community-acquired pneumonia (CAP) is the most frequent infection-related cause of death. The reference standard to diagnose CAP is a new infiltrate on chest radiograph in the presence of recently acquired respiratory signs and symptoms. This study aims to evaluate the diagnostic and prognostic accuracy of clinical signs and symptoms and laboratory biomarkers for CAP.

Methods

545 patients with suspected lower respiratory tract infection, admitted to the emergency department of a university hospital were included in a pre-planned post-hoc analysis of two controlled intervention trials. Baseline assessment included history, clinical examination, radiography and measurements of procalcitonin (PCT), highly sensitive C-reactive protein (hsCRP) and leukocyte count.

Results

Of the 545 patients, 373 had CAP, 132 other respiratory tract infections, and 40 other final diagnoses. The AUC of a clinical model including standard clinical signs and symptoms (i.e. fever, cough, sputum production, abnormal chest auscultation and dyspnea) to diagnose CAP was 0.79 [95% CI, 0.75–0.83]. This AUC was significantly improved by including PCT and hsCRP (0.92 [0.89–0.94]; p < 0.001). PCT had a higher diagnostic accuracy (AUC, 0.88 [0.84–0.93]) in differentiating CAP from other diagnoses, as compared to hsCRP (AUC, 0.76 [0.69–0.83]; p < 0.001) and total leukocyte count (AUC, 0.69 [0.62–0.77]; p < 0.001). To predict bacteremia, PCT had a higher AUC (0.85 [0.80–0.91]) as compared to hsCRP (p = 0.01), leukocyte count (p = 0.002) and elevated body temperature (p < 0.001). PCT, in contrast to hsCRP and leukocyte count, increased with increasing severity of CAP, as assessed by the pneumonia severity index (p < 0.001).

Conclusion

PCT, and to a lesser degree hsCRP, improve the accuracy of currently recommended approaches for the diagnosis of CAP, thereby complementing clinical signs and symptoms. PCT is useful in the severity assessment of CAP.