Procalcitonin in prediction of bacteremia in patients with community acquired fever

Abstract

Objective To explore the value of procalcitonin (PCT) in predicting bacteremia in patients with community acquired fever. Methods The adult patients with fever who visited our emergency department and were hospitalized were retrospectively enrolled from January, 2015 to January, 2016; the basic data, white cell count (WBC), neutrophil percentage (NEU %), PCT, and blood culture results were recorded. The area under the ROC curve (AUC) was constructed to evaluate the discriminative power of variables to predict bacteremia. Results During the study, a total of 223 patients were enrolled. Of these patients, respiratory tract infection (54.7%) was the most common fever cause, and virus infection (11.7%) came next. Blood culture were obtained from 78.0% (174/223) patients, including bacteremia (n=13) and non-bacteremia (n=161); the PCT level was significantly higher in the bacteremia group than in the non-bacteremia group (P 0.01). The area under the receiver operating characteristic curve was 0.793, and 95% CI 0.793-0.631 for PCT, which was significantly higher than that for WBC (AUC = 0.691, 95% CI 0.510-0.872) and NEU% (0.620, 95% CI 0.391-0.849). While the area under the curve of the WBC and NEU% had no statistical difference (P>0.01). When the PCT was less than 0.45 ng/ml, bacteremia can be excluded, with a negative predictive value of 95.5%. Conclusions At our emergency department, respiratory infection is the most common fever cause. PCT concentration is more useful in predicting bacteremia than WBC and NEU%. Also, the determination time of PCT is less than 1 h, which may be effectively used to exclude bacteremia and establish early antibiotic treatment plans. Key words: Emergency department; Fever; Bacteremia; Procalcitonin; ROC curve