ABSTRACT
OBJECTIVE: Febrile infants younger than 60 days are at risk for serious bacterial infections (SBIs) and often undergo extensive laboratory investigation and hospitalization. We aim to determine the diagnostic value of serum procalcitonin (PCT) concentration for identification of febrile infants at low risk for SBI in comparison to the Rochester Criteria (RC). METHODS: Infants 2 to 60 days of age with rectal temperature 38°C were enrolled between May 2004 and March 2007. Blood was obtained from each, and PCT was assessed using BRAHMS PCT LIA method. Information for identification of low-risk infants using RC was obtained. Negative predictive value, sensitivity, specificity, and likelihood ratio of PCT were compared with the RC. In univariate analysis, the components of RC and PCT were considered. Variables holding a significant association with the absence of SBI were included in a backward stepwise logistic regression model with SBI as the dependent variable, creating new low risk criteria. RESULTS: One hundred fifty-five patients were enrolled. Thirteen (8.4%) had an SBI. Procalcitonin concentration at a cutoff value of 0.26 ng/mL is similar in sensitivity (92%) and better in specificity (64%) than RC. A combination of urine white blood cell and PCT was the best model in the regression analysis. CONCLUSIONS: Procalcitonin concentration is a serological marker for identification of or exclusion of SBI in infants aged 2 to 60 days. The predictive value of PCT in combination with urinary white blood cell count may be clinically useful. A validation study is indicated.
