ABSTRACTObjectives:Recent Infectious Disease Society of America guidelines recommend multistep testing algorithms to diagnose Clostridioides difficile infection (CDI), including a combination of nucleic acid amplification‐based testing (NAAT) and toxin enzyme immunoassay (EIA). The use of these algorithms in children, including the ability to differentiate between C. difficile colonization and CDI, however, has not been evaluated.Methods:We prospectively enrolled asymptomatic pediatric patients with cancer, cystic fibrosis (CF), or inflammatory bowel disease (IBD) and obtained a stool sample for NAAT testing. If positive by NAAT (colonized), EIA was performed. In addition, children with symptomatic CDI who tested positive by NAAT via the clinical laboratory were enrolled, and EIA was performed on residual stool. A functional cell cytotoxicity neutralization assay (CCNA) was also applied to stool samples from both the colonized and symptomatic cohorts.Results:Of the 225 asymptomatic children enrolled in the study, 47 (21%) were colonized with C. difficile including 9/59 (15.5%) with cancer, 30/92 (32.6%) with CF, and 8/74 (10.8%) with IBD. An additional 41 children with symptomatic CDI were enrolled. When symptomatic and colonized children were compared, neither EIA positivity (44% vs 26%, P = 0.07) nor CCNA positivity (49% vs 45%, P = 0.70) differed significantly or were able to predict disease severity in the symptomatic cohort.Conclusions:Use of a multistep testing algorithm with NAAT followed by EIA failed to differentiate symptomatic CDI from asymptomatic colonization in our pediatric cohort. As multistep algorithms are moved into clinical care, the pediatric provider will need to be aware of their limitations.