Evaluation of six commercial assays for the rapid detection of Clostridium difficile toxin and/or antigen in stool specimens

Abstract

To evaluate six commercially available assays for the detection of Clostridium difficile toxin and/or antigen in stool samples: one latex agglutination test (Culturette brand CDT, Becton Dickinson), two ELISAs (Culturette brand Toxin CD, Becton Dickinson, and Ridascreen C. difficile Toxin A/B, R-biopharm), two chromatographic assays (Clearview C. difficile A, Oxoid, and ColorPac Toxin A, Becton Dickinson) and one enzyme immunoassay for the simultaneous detection of C. difficile common antigen and toxin A (Triage C. difficile Panel, Biosite). Over a period of 3 months, 366 liquid or semi-liquid stool samples were tested using cell-culture cytotoxin assay as standard, ethanol shock stool culture and latex agglutination (Culturette brand CDT). Of these, 78 samples, positive with at least one of these three methods, and 98 randomly selected negative samples were further evaluated using the other five kits. PCR was also performed on positive cultures to confirm the presence of toxin A and B genes. Triage C. difficile Panel had the best sensitivity (95%), followed by Clearview C. difficile and ColorPac Toxin A (both 89%), Culturette brand Toxin CD (73%), Ridascreen C. difficile Toxin A/B (57%) and Culturette brand CDT (23%). For Triage, the sensitivity of C. difficile antigen detection was 93%, and the sensitivity of toxin detection was lower (77%). Most false-positive results were obtained with the Triage C. difficile Panel (25 specimens) and Clearview C. difficile A (20 specimens). Culturette brand CDT had the best specificity (99%); followed by Ridascreen C. difficile Toxin A/B (97%), Culturette brand Toxin CD (95%), ColorPac Toxin A (89%), Clearview C. difficile A (83%) and Triage C. difficile Panel (75%). The positive predictive values ranged from 68% to 94%, and the negative predictive values from 83% to 98%. The sensitivity is much higher for Triage and the two new chromatographic assays than for the conventional EIAs. These tests also have a high negative predictive value. For Triage, C. difficile antigen-positive, toxin A-negative results can be obtained; the clinical value of these must be established by additional studies. Overall, the new-generation assays are still less sensitive than the cytotoxin assay; however, they provided same-day results, could be used as a screening test and may be useful in laboratories without tissue-culture facilities. Our results do not allow the recommendation of one single assay for the diagnosis of C. difficile-associated diarrhea. It remains the case that laboratory results must be correlated and interpreted with the clinical presentation of the patient.