Introduction: Extracorporeal membrane oxygenation (ECMO) is associated with complications, including nosocomial infection. Fungemia during pediatric ECMO is associated with increased mortality in single-center studies, but there is little multicenter data of its epidemiology and outcomes. We aim to describe prevalence, treatments, and outcomes of fungemia in a large multicenter database. Methods: With IRB approval, we identified PICU patients < 18yo in the Pediatric Health Information Systems database (PHIS) who received ECMO from 2018-2024. We collected demographics, ICD diagnosis codes, comorbidity “flags” and pharmacy billing data. Pediatric Clinical Classification System (PECCS) primary diagnoses were organized into six groups (e.g. cardiac surgical, infectious etc). We defined invasive fungemia with previously published ICD codes. Patient characteristics and outcomes (90d ECMO-free, PICU-free PFDs, and ventilator-free days VFDs; and discharge mortality) were compared with chi-square (categorical data) and Wilcoxon rank-sum (continuous data, shown as median IQR) tests), followed by multivariate models, with p< 0.05 as statistical significance. Results: Among 8537 subjects from 48 centers, the 496 (6%) subjects with fungemia were older (52 7.8-155 vs 6.67 0.03-74 mo.) and more likely to have immunologic (41% vs 17%), oncologic (17% vs 4%) and transplant (21% vs 9%) (all p< 0.001) comorbidities. Fungemia patients more commonly received amphotericin (44% vs 4%), fluconazole (63% vs 31%), and micafungin (63% vs 15%) (all p< 0.001). Fungemia patients had fewer ECMO-free days (48 0–81 vs 79 0–85d), PFDs (0 0–53 vs. 36 0–73d), VFDs (0 0-59 vs 45 0-76d), and higher mortality (48% vs 38%) (all p< 0.001). In multivariate models adjusted for age, race, comorbidities, and PECCS diagnosis groups, fungemia was associated with 9.2, 7.6, and 9.8 fewer ECMO-free days, PFDs, and VFDs, respectively (all p< 0.001) and increased odds of mortality (OR: 1.57 95% CI: 1.29-1.92). Conclusions: Among thousands of pediatric ECMO patients in an administrative database, fungemia occurred in approximately 1 of 20 cases and was associated with unfavorable outcomes even after adjusting for confounders. Further research is needed for treatment and prevention strategies in this vulnerable cohort.
Srinivasan et al. (Sun,) studied this question.