OBJECTIVES: To explore the diagnostic performance of WBC count, C-reactive protein (CRP), and procalcitonin in identifying nosocomial infections in pediatric patients supported with extracorporeal membrane oxygenation (ECMO). DESIGN: Single-center retrospective cohort of pediatric patients supported with ECMO, from November 2017 to December 2022. SETTING: PICU, neonatal ICU, and cardiac ICU in a Quaternary Children's Hospital. PATIENTS: Critically ill patients age younger than 18 years supported with ECMO for greater than 24 hours. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: We identified 154 ECMO runs involving 151 pediatric patients. New infections were detected in 63 of 154 ECMO runs (41%). There were eight of 1442 positive blood cultures (0.5%), nine of 617 positive urine cultures (1.4%), and 60 of 722 positive respiratory cultures (8.3%). Antibiotics were administered on 1217 ECMO days (84.7%), and 98.7% of the patients received more than 48 hours of antibiotics. On evaluating the diagnostic performance of tests for distinguishing new infection, using the area under the receiver operating characteristic curve (AUC, with 95% CI), we found that both CRP and procalcitonin had poor diagnostic accuracy, respectively (AUC, 0.66; 95% CI, 0.60-0.72 and AUC, 0.62; 95% CI, 0.53-0.72). The WBC showed no discrimination with AUC 0.51 (95% CI, 0.44-0.57). Notably, a procalcitonin value less than 1.77 ng/mL had a 95% negative predictive value (NPV), CRP less than 36.5 mg/L had a 97% NPV, and WBC less than 11.02 K/µL had a 95% NPV. CONCLUSIONS: In our retrospective, exploratory evaluation of nosocomial infection in ECMO patients (2017-2022), no laboratory marker demonstrated adequate discriminative ability for indicating the presence of infection. However, normal values for these markers may help clinicians decide whether unnecessary antibiotic treatment can be reduced and whether surveillance cultures can be eliminated in pediatric patients on ECMO.
Gillett et al. (Wed,) studied this question.