Background: Procalcitonin (PCT) it is a new marker of bacterial infection. Because of its shorter half-life and earlier ascent it offers advantages over C-reactive protein (CRP).
Objective: To compare the diagnostic performance of PCT in the early detection of invasive bacterial infection in infants with that of CPR.
Material and methods: Between January of 1998 and February of 2000 we performed a prospective observational study in the emergency department of infants aged between 1 and 36 months who had been treated for fever and for whom PCT and CRP plasmatic values had been obtained. Plasmatic PCT and PCR values were evaluated and correlated with the final diagnosis. ROC curves for both markers were calculated.
Results: One hundred infants with a mean age of 8.8 months (SD 7.59) were included in four groups of 25 patients each (viral infection, localized bacterial infection, invasive bacterial infection and control group). The mean PCT and CRP values in invasive bacterial infections [PCT: 14.45 ng/mL (SD 27.95) and CRP: 95.10 mg/L (SD 7 2.77)] were significantly higher than in non-invasive infections [PCT: 0.27 ng/mL (SD 0.19) and CRP: 25.67 mg/L (SD 33.04)] but the diagnostic performance of PCT was better. The area under the curve for PCT was 0.95 (SD 0.03), which was significantly higher (p < 0.001) than that obtained for CRP [0.81 (SD 0.05)]. The optimal cut-off for PCT was > 0.4 ng/mL (sensitivity: 95.5 %; specificity: 86.4 %) and that for CRP was > 42.9 mg/L (sensitivity: 75 %; specificity: 81.8 %). In infants who had fever for less than 12 hours (n 30) the area under the curve for PCT was 0.90 (SD 0.06), which was higher (p < 0.001) than that for PCR [0.64 (SD 0.11)]. The optimal cut-off for PCT in this group was > 0.4 ng/mL (sensitivity: 90 %; specificity: 94 %) and that for CRP was > 26.6 mg/L (sensitivity: 60 %; specificity: 77.8 %).
Conclusions: The diagnostic performance of PCT was higher than that of CRP in the early detection of invasive infection in febrile infants, even when evolution was less than 12 hours.