Low Levels of Procalcitonin Are Related to Decreased Antibiotic Use in Children Hospitalized Due to Influenza.

Abstract

Procalcitonin increases in bacterial infections, which are often suspected (though rarely confirmed) in paediatric influenza. We retrospectively verified procalcitonin’s usefulness in antibiotic guidance in children hospitalized due to laboratory-confirmed influenza. The ROC curve analysis evaluated procalcitonin’s performance in terms of antibiotic implementation or continuation in patients who were naive or had been receiving antibiotic treatment prior to hospital admission. We also assessed the procalcitonin’s usefulness to predict lower-respiratory-tract infections (LRTI), the presence of radiologically confirmed pneumonia, an intensive care unit transfer and a fatal outcome. Multiple regression models were built to verify the previously reported procalcitonin cut-off values. The study enrolled 371 children (median age 33 months). The AUC (area under the curve) for antibiotic implementation reached 0.66 (95%CI: 0.58–0.73) and 0.713 (95%CI: 0.6–0.83) for antibiotic continuation; optimal cut-offs (0.4 and 0.23 ng/mL, respectively) resulted in a negative predictive value (NPV) of 79.7% (95%CI: 76.2–82.9%) and 54.6% (95%CI: 45.8–63%), respectively. The use of 0.25 ng/mL as a reference decreased the odds of antibiotic treatment by 67% (95%CI: 43–81%) and 91% (95%CI: 56–98%), respectively. Procalcitonin showed lower AUC for the prediction of LRTI and pneumonia (0.6, 95%CI: 0.53–0.66, and 0.63, 95%CI: 0.56–0.7, respectively), with a moderately high NPV in the latter case (83%, 95%CI: 79.3–86.1%). Procalcitonin use may decrease the antibiotic frequency in hospitalized influenza cases both in terms of antibiotic administration and continuation. Procalcitonin concentrations may suggest bacterial suprainfections at lower concentrations than in adults, and a focus on its rule-out value is of special interest.