Abstract
Clostridium difficile is the leading cause of healthcare-associated infections worldwide. The diagnosis of C. difficile infection (CDI) in pediatric oncology patients is complex as diarrhea is common, and there is a high rate of colonization in infants and young children. This study was conducted to assess the accuracy of the surveillance definitions of healthcare-associated CDI (HA-CDI) and to determine the prevalence of toxigenic C. difficile colonization among pediatric oncology and stem cell transplant patients. Methods A prospective cohort study was conducted over a three-year period in an inpatient pediatric oncology and stem cell transplant setting. Baseline stool samples were collected within three days of admission and were genotypically compared with clinically indicated samples submitted after three days of admission. Results A total of 175 patients were recruited with a total of 536 admissions. The adjusted prevalence of baseline toxigenic C. difficile colonization among admissions was 32.8%. Seventy-eight percent of positive admissions did not have history of CDI. Colonization with a toxigenic strain on admission was predictive of CDI (OR = 28.6; 95% CI, 6.58-124.39; P < 0.001). Nearly all clinical isolates (8/9) shared identical pulsed-field gel electrophoresis patterns with baseline isolates or were closely related (1/9). Only one of the 11 cases that were considered HA-CDI was potentially nosocomially acquired. Conclusion The prevalence of colonization with toxigenic C. difficile in our cohort is high. Unfortunately, the current CDI surveillance definitions overestimate the incidence of HA-CDI in pediatric oncology and stem cell transplantation settings.