Abstract
A multicenter clinical trial assessed the performance of the Cepheid Xpert C. difficile assay on stool specimens collected from patients suspected of having Clostridium difficile infection (CDI). A total of 2,296 unformed stool specimens, collected from seven study sites, were tested by Xpert C. difficile enrichment culture followed by cell culture cytotoxicity testing of the isolates (i.e., toxigenic culture with enrichment) and the study sites' standard C. difficile test methods. The methods included enzyme immunoassay (EIA), direct cytotoxin testing, and two- and three-step algorithms using glutamate dehydrogenase (GDH) screening followed by either EIA or EIA and an in-house PCR assay. All C. difficile strains were typed by PCR-ribotyping. Compared to results for toxigenic culture with enrichment, the sensitivity, specificity, and positive and negative predictive values of the Xpert assay were 93.5, 94.0, 73.0, and 98.8%, respectively. The overall sensitivity of the EIAs compared to that of enrichment culture was 60.0%, and the sensitivity of combined GDH algorithms was 72.9%; both were significantly lower than that of Xpert C. difficile (P <0.001 and P = 0.03, respectively). The sensitivity of the EIA was significantly lower than that of the Xpert C. difficile assay for detection of ribotypes 002, 027, and 106 (P <0.0001, P <0.0001, and P = 0.004, respectively, Fisher's exact test), and the sensitivity of GDH algorithms for ribotypes other than 027 was lower than that for Xpert C. difficile (P <0.001). The Xpert C. difficile assay is a simple, rapid, and accurate method for detection of toxigenic C. difficile in unformed stool specimens and is minimally affected by strain type compared to EIA and GDH-based methods.