<p><strong>OBJECTIVES: </strong>To determine the frequency of false-positive Clostridium difficile toxin enzyme immunoassay (EIA) results in hospitalized children and to examine potential reasons for this false positivity.</p>

<p><strong>DESIGN: </strong>Nested case-control.</p>

<p><strong>SETTING: </strong>Two tertiary care pediatric hospitals.</p>

<p><strong>METHODS: </strong>As part of a natural history study, prospectively collected EIA-positive stools were cultured for toxigenic C. difficile, and characteristics of children with false-positive and true-positive EIA results were compared. EIA-positive/culture-negative samples were recultured after dilution and enrichment steps, were evaluated for presence of the tcdB gene by polymerase chain reaction (PCR), and were further cultured for Clostridium sordellii, a cause of false-positive EIA toxin assays.</p>

<p><strong>RESULTS: </strong>Of 112 EIA-positive stools cultured, 72 grew toxigenic C. difficile and 40 did not, indicating a positive predictive value of 64% in this population. The estimated prevalence of C. difficile infection (CDI) in the study sites among children tested for this pathogen was 5%-7%. Children with false-positive EIA results were significantly younger than those with true-positive tests but did not differ in other characteristics. No false-positive specimens yielded C. difficile when cultured after enrichment or serial dilution, 1 specimen was positive for tcdB by PCR, and none grew C. sordellii.</p>

<p><strong>CONCLUSIONS: </strong>Approximately one-third of EIA tests used to evaluate pediatric inpatients for CDI were falsely positive. This finding was likely due to the low prevalence of CDI in pediatric hospitals, which diminishes the test's positive predictive value. These data raise concerns about the use of EIA assays to diagnosis CDI in children.</p>

<p><strong>BACKGROUND: </strong>The incidence and severity of Clostridium difficile infection (CDI) is increasing among adults; however, little is known about the epidemiology of CDI among children.</p>

<p><strong>METHODS: </strong>We conducted a nested case-control study to identify the risk factors for and a prospective cohort study to determine the outcomes associated with severe CDI at 2 children's hospitals. Severe CDI was defined as CDI and at least 1 complication or ≥2 laboratory or clinical indicators consistent with severe disease. Studied outcomes included relapse, treatment failure, and CDI-related complications. Isolates were tested to determine North American pulsed-field gel electrophoresis type 1 lineage.</p>

<p><strong>RESULTS: </strong>We analyzed 82 patients with CDI, of whom 48 had severe disease. Median age in years was 5.93 (1.78-12.16) and 1.83 (0.67-8.1) in subjects with severe and nonsevere CDI, respectively (P = 0.012). All patients with malignancy and CDI had severe disease. Nine subjects (11%) had North American pulsed-field gel electrophoresis type 1 isolates. Risk factors for severe disease included age (adjusted odds ratio [95% confidence interval]: 1.12 [1.02, 1.24]) and receipt of 3 antibiotic classes in the 30 days before infection (3.95 [1.19, 13.11]). If infants less than 1 year of age were excluded, only receipt of 3 antibiotic classes remained significantly associated with severe disease. Neither the rate of relapse nor treatment failure differed significantly between patients with severe and nonsevere CDI. There was 1 death.</p>

<p><strong>CONCLUSIONS: </strong>Increasing age and exposure to multiple antibiotic classes were risk factors for severe CDI. Although most patients studied had severe disease, complications were infrequent. Relapse rates were similar to those reported in adults.</p>