Effect of Viral Illness on Procalcitonin as a Predictor of Bacterial Infection in Febrile Infants

Abstract

OBJECTIVE The impact of confirmed viral infections (CVI) on procalcitonin (PCT) levels in febrile infants aged 8–60 days with a bacterial illness (BI) is unknown. The objectives of the study were to (1) examine the association of CVI with PCT levels in patients with/without a concurrent BI, defined as bacteremia, meningitis, or urinary tract infection, and (2) assess PCT as a predictor of BI in infants with a concurrent CVI. METHODS In this single-center, retrospective cohort study, we examined febrile infants aged 8–60 days presenting between January 1, 2018 and December 31, 2020. PCT levels were compared between groups, according to results of bacterial cultures and viral tests, using the Wilcoxon rank test. The prediction ability of PCT to detect BI with/without concurrent CVI was assessed by using area under the curve from logistic regression. RESULTS Patients included: 404 BI−/CVI+, 73 BI+/CVI−, 48 BI+/CVI+, and 138 BI−/CVI−. Median PCT level in the BI+/CVI+ group was significantly lower when compared to BI+/CVI− (0.36 ng/mL vs 0.89 ng/mL), but significantly higher than the BI−/CVI− group (0.36 ng/mL vs 0.1 ng/mL). The presence of a CVI reduced the sensitivity of PCT in BI detection (68% vs 44%), with minimal impact specificity (93% vs 96%). CONCLUSIONS In previously healthy febrile infants 8–60 days old, the presence of a CVI reduces the sensitivity of PCT BI detection without impacting its specificity. The impact of a CVI on PCT levels in febrile infants has implications for how this marker of infection should be considered when assessing risk of BI in infants.