1361. Laboratory diagnosis overestimates the burden of pediatric community-onset Clostridium difficile infection: Impact of symptom review, culture, and typing on case validation
Session: Poster Abstract Session: Clostridium difficile
Saturday, October 5, 2013
Room: The Moscone Center: Poster Hall C
  • ID Week 2013 Poster Final.pdf (2.5 MB)
  • Background: Increased incidence of laboratory-identified (labID) pediatric community-onset (CO) Clostridium difficile infection (CDI), particularly among children with few risk factors, is reported from several population-based studies. These studies are limited by lack of clinical data to confirm infection. Since C. difficile colonization is common among young children, misdiagnosis impacts accuracy of epidemiologic data.  

    Methods: All labID cases of CDI at an academic children’s hospital are reviewed, and saved stool specimens positive by toxin gene PCR undergo culture and restriction endonuclease analysis (REA) typing. We present data from all cases Dec. 2012 – Feb. 2013.

    Results: 32 patients had onset of CDI symptoms within the community: 7 CO-healthcare facility-associated; 17 community-associated; 4 indeterminate; and 4 recurrent. Mean age was 6 years (range 11 months – 18 years; 7 patients ≤ 24 months-old). A probable alternate etiology of diarrhea was identified in 20 (63%): 14 had vomiting suggesting an upper gastrointestinal (UGI) illness; 3 were receiving diarrhea-causing medications; and 3 with inflammatory bowel disease did not improve with oral vancomycin. 17 (53%) patients had not received antibiotics associated with CDI, proton-pump inhibitors, or immunosuppressives in the past month. Stool of sufficient quantity was saved from 26 (81%) patients, and C. difficile was isolated by culture from 23 specimens (88%). REA identified 18 (78%) toxigenic and 3 (9%) non-toxigenic isolates from a wide variety of REA groups; 2 (9%) were inconclusive. No binary toxin-producing strains, including BI/NAP1/027, were identified. Compared to toxigenic culture, positive predictive value (PPV) of PCR for detection of toxigenic C. difficile was 69-77%.

    Conclusion: Among children with labID CO-CDI, preliminary data demonstrate high frequency of UGI symptoms, low frequency of CDI risk factors, lack of group BI (endemic in Chicago adult healthcare facilities), and suboptimal PPV of PCR for detection of toxigenic C. difficile. These data suggest a significant proportion of labID CO-CDI may represent colonized children with an alternate cause of diarrhea or patients with a false positive toxin gene PCR. The true burden of CO-CDI may be overestimated in population-based studies.

    Larry Kociolek, MD1, Stanford Shulman, MD1 and Dale Gerding, MD2, (1)Infectious Diseases, Ann and Robert H. Lurie Children's Hospital of Chicago, Chicago, IL, (2)Loyola University and Hines VA Hospital, Hines, IL


    L. Kociolek, Merck: Grant Investigator, Grant recipient
    Optimer: Investigator, Research support

    S. Shulman, None

    D. Gerding, None

    Findings in the abstracts are embargoed until 12:01 a.m. PST, Oct. 2nd with the exception of research findings presented at the IDWeek press conferences.