Clostridium difficile infection diagnosis in a paediatric population: comparison of methodologies

J. Hart, P. Putsathit, D. R. Knight, L. Sammels, T. V. Riley, A. Keil

The increasing incidence of Clostridium difficile infection (CDI) in paediatric hospitalised populations, combined with the emergence of hypervirulent strains, community-acquired CDI and the need for prompt treatment and infection control, makes the rapid, accurate diagnosis of CDI crucial. We validated commonly used C. difficile diagnostic tests in a paediatric hospital population. From October 2011 to January 2012, 150 consecutive stools were collected from 75 patients at a tertiary paediatric hospital in Perth, Western Australia. Stools were tested using: C. Diff Quik Chek Complete, Illumigene C. difficile, GeneOhm Cdiff, cycloserine cefoxitin fructose agar (CCFA) culture, and cell culture cytotoxin neutralisation assay (CCNA). The reference standard was growth on CCFA or Cdiff Chromagar and PCR on isolates to detect tcdA, tcdB, cdtA, and cdtB. Isolates were PCR ribotyped. The prevalence of CDI was high (43 % of patients). Quik Chek Complete glutamate dehydrogenase (GDH) demonstrated a low negative predictive value (NPV) (93 %). Both CCNA and Quik Chek Complete toxin A/B had poor sensitivity (33 % and 29 % respectively). Molecular methods both had 89 % sensitivity. Algorithms using GDH + Illumigene or GeneOhm reduced the sensitivity to 85 % and 83 % respectively. Ribotype UK014/20 predominated. GDH NPV and GeneOhm and Illumigene sensitivities were reduced compared with adult studies. Quik Chek Complete and CCNA cannot reliably detect toxigenic CDI. A GDH first algorithm showed reduced sensitivity. In a high prevalence paediatric population, molecular methods alone are recommended over the use of GDH algorithm or culture and CCNA, as they demonstrate the best test performance characteristics.