Association of procalcitonin values and bacterial infections in pediatric patients receiving extracorporeal membrane oxygenation.

Objective: There is increasing data in pediatrics demonstrating procalcitonin (PCT) is more sensitive and specific than other biomarkers in the setting of bacterial infections. However, the use of PCT in neonatal and pediatric extracorporeal membrane oxygenation (ECMO) is not well described. Therefore, the purpose of this study was to describe the clinical utility of PCT in determining the absence or presence of bacterial infections in neonatal and pediatric patients on ECMO. Methods: This was a retrospective electronic medical record (EMR) review of data between January 1, 2010 to June 30, 2016 at a single, free-standing, children's hospital. All patients on ECMO with ≥1 PCT level obtained while receiving ECMO support were eligible for inclusion. The EMR was searched for chest radiographs (CXR) and bacterial culture results (urine, blood, cerebrospinal fluid (CSF), bronchoalveolar lavage (BAL) and respiratory cultures). All bacterial and viral cultures obtained within 5 days of PCT levels being obtained were analyzed. PCT levels of 0.5, 0.9, 1.0, 1.4 and 2.0 were used as the initial cut-off values for the analysis. The sensitivity, specificity, positive predictive value (PPV), negative predictive values (NPV) and likelihood ratios were calculated for each of the PCT levels. Results: Twenty-seven patients met the inclusion criteria and contributed 193 PCT values for the analysis. The median age was 8 months (range 0 days to 18 years). Linear regression analysis demonstrated that a PCT cut-off of 0.5, 0.9 and 1.4 predicted the presence of a bacterial infection. The PCT value with the most utility was 0.5, with a sensitivity of 92%, a specificity of 43%, a positive predictive value of 60% and a negative predictive value (NPV) of 86%. Conclusion: This is the largest data set evaluating PCT in neonatal and pediatric patients on ECMO. A PCT value of 0.5 ng/mL had the most utility for determining the absence or presence of a bacterial infection in the setting of ECMO with a high sensitivity and NPV. [ABSTRACT FROM AUTHOR]