RESUMO
The aim of the present study is to evaluate the usefulness of two biomarkers-procalcitonin (PCT) and C-reactive protein (CRP)-in addition to the CURB-65 score for assessing the site of care and the etiology of non-severe community-acquired pneumonia (CAP). We conducted a prospective observational study from April 1, 2006, to June 30, 2007, in a single teaching hospital in northern Spain among patients with non-severe CAP. In addition to collecting data needed to determine the CURB-65 score, microbial cultures were taken and levels of PCT and CRP were measured. We compared the prognostic accuracy of these biomarkers with the CURB-65 score to predict hospitalization and microbial etiology using receiver operating characteristic (ROC) curves. A total of 344 patients with non-severe CAP were enrolled; 73 were admitted to the hospital and 271 were treated on an outpatient basis. An etiologic diagnostic was made for 44 %, with atypical pathogens predominating. Levels of PCT and CRP increased with increasing CURB-65 scores. Patients admitted to the hospital had higher PCT and CRP levels than outpatients (p < 0.001). For predicting hospitalization, PCT had a better area under the ROC curve (AUC) (0.81) than the CURB-65 score alone (0.77). For PCT plus the CURB-65 score, the AUC increased significantly from 0.77 to 0.83. In patients with bacterial CAP, the biomarker levels were significantly higher than among patients with atypical or viral etiology (p < 0.001). PCT with a cut-off point of 0.15 ng/mL was the best predictor for bacterial etiology and for select patients eligible for outpatient care. In conclusion, levels of PCT and CRP positively correlate with increasing severity of CAP and may have a role in predicting both patients who can safely receive outpatient care and the microbial etiology in patients with low CURB-65 scores.