Accuracy of C-reactive protein, procalcitonin, and mid-regional pro-atrial natriuretic peptide to guide site of care of community-acquired pneumonia

The use at bedside of C-reactive protein (CRP), procalcitonin (PCT) or mid-regional pro-atrial natriuretic peptide (ANP) has been advocated to help management of patients with community-acquired pneumonia (CAP) in emergency medicine. To assess the effectiveness of CRP, PCT, and ANP measures in assisting emergency physicians deciding hospital admission for CAP with low risk of complication. Multicenter, prospective, observational study with blind evaluation. Emergency departments of 12 French hospitals. Five hundred forty-nine consecutive, immunocompetent adult patients with mild CAP. Centralized and blind measure of baseline CRP, PCT, and ANP; sensitivity, specificity, and positive and negative likelihood ratios for determining hospital admission. Gold standard for admission was defined by experts' advice combined with admission requirement or death at 28 days. Optimal threshold values were determined by receiver operating characteristic (ROC) curves, and area under curve (AUC) of the three biomarkers was compared. According to gold standard, 310 (56%) patients required admission and 239 (44%) needed to be discharged. PCT and ANP levels increased with Pneumonia Severity Index risk categories. ANP (AUC 0.76 [95% CI 0.72-0.80]) more accurately predicted admission requirement than did PCT (AUC 0.65 [95% CI 0.61-0.70]) or CRP (AUC 0.59 [95% CI 0.54-0.64]) (both p values < 0.01). We determined that 135 pmol/L was a threshold for ANP level to discriminate admission requirement (positive likelihood ratio 7.45 [95% CI 4.22-8.16]). In a selected population of CAP with low risk of complication, a single ANP measurement was more accurate than CRP and PCT to predict appropriate admission. These results should be confirmed by additional studies.