Procalcitonin to Predict Bacterial Coinfection in Infants With Acute Bronchiolitis

Abstract

Objective The aim of this study was to conduct a preliminary analysis of serum procalcitonin (PCT) to predict bacterial coinfection in infants with acute bronchiolitis. Methods Retrospective cohort chart review of 40 infants admitted with acute bronchiolitis to the pediatric intensive care unit. Logistic regression models were used to determine the association of PCT and white blood count with presence of bacterial coinfection defined by either positive culture or chest radiograph result. Results Fifteen (38%) of 40 patients had a diagnosis of bacterial coinfection by positive culture (9/15) or chest radiograph (6/15). Procalcitonin (P < 0.0001) was significantly associated with bacterial coinfection. A cutoff value of 1.5 ng/mL had sensitivity of 0.80, specificity of 1.00, and area under the operating curve of 0.88. White blood count (P = 0.06) was borderline significant with sensitivity of 0.33, specificity of 0.96, and area under the operating curve of 0.67. Three of 15 patients were later found to have bacterial coinfection with initial PCT of less than 1.5 ng/mL. None had follow-up PCT measurements taken. Thirty-five of 40 were prescribed empiric antibiotic therapy, including 20 of 25 patients without evidence of bacterial coinfection. None had a PCT of greater than 1.5 ng/mL. If a PCT cutoff of greater than 1.5 ng/mL had been used, 57% fewer patients would have received antibiotics with a 45% reduction in antimicrobial charges. Conclusions An elevated PCT may assist clinicians in determining presence of bacterial coinfection at admission in infants with acute bronchiolitis. Implementation of a PCT cutoff of 1.5 ng/mL at admission may prevent unnecessary antibiotic use with associated cost savings. Serial PCT levels may increase sensitivity. Further validation is warranted.