C-reactive protein is the most useful biomarker for diagnosis of outpatient community-acquired pneumonia

Référence

Ebell MH, Bentivegna M, Cai X, Hulme C, Kearney M. Accuracy of biomarkers for the diagnosis of adult community-acquired pneumonia: A meta-analysis. Acad Emerg Med 2020;27(3):195-206.

Sommaire

These investigators thoroughly searched PubMed and reference lists of pertinent articles for studies that evaluated the accuracy of readily available biomarkers in adult patients who present with symptoms of acute respiratory infection and those with clinically suspected pneumonia. The authors included only studies with at least one biomarker test and an acceptable reference standard (chest radiography or computed tomography) performed on all participants. They did not exclude any studies based on country, year, or language. Two or more individuals independently reviewed potential studies for inclusion and risk of bias using a standard evaluation tool. Discrepancies were resolved through consensus discussion. The authors ended up with 14 studies with a total of 6599 patients. Half of the studies occurred in the emergency department and half occurred in the primary care setting. Of the 14 studies, 8 were scored at low risk of bias, with the remainder considered at moderate risk of bias. Overall, CRP was the most accurate test (CRP < 10 mg/L; LR- = 0.27 and CRP > 20 mg/L, 50 mg/L, and 100 mg/L; LR+ = 2.08, 3.68, and 5.79, respectively). Procalcitonin had a good LR+ (> 0.25 ug/L and 0.50 ug/L; LR+ = 5.43 and 8.25, respectively), but a minimally useful LR-. Leukocytosis was only modestly, if at all, accurate (> 9.5 - 10.5; LR+ = 3.15). Based on existing clinical decision rules using signs and symptoms, if the probability of CAP is 25% or less, a CRP of less than 20 mg/L decreases the probability of CAP to less than 10%, making imaging unlikely to be clinically useful and thus unnecessary.