Comparison of antigen and two molecular methods for the detection of Clostridium difficile toxins

Abstract

Clostridium difficile (CD) is considered an important cause of diarrhoea associated with the antimicrobial treatment of infections. The pathogenicity of CD is due to toxins A and B, produced by toxigenic CD strains. We evaluated 3 methods for detecting CD toxins: the RIDASCREEN® enzyme immunoassay (EIA) (R-Biopharm)--one detecting toxins directly in the stool specimens and another detecting toxins from isolated CD strains--and 2 molecular methods, the illumigene™ loop-mediated isothermal amplification (LAMP) assay (Meridian) and RIDA®GENE polymerase chain reaction (PCR) assay (R-Biopharm), as direct identification methods from stool specimens. Toxigenic culture (TC) was used as the reference method. Altogether 884 stool samples were analyzed, of which 253 (29%) were positive by TC. Six hundred and seventy-two specimens were tested by RIDASCREEN EIA, 430 were tested with the illumigene LAMP assay, and 212 were tested with the RIDA GENE PCR assay. CD toxin A and B antigen tests by EIA were very insensitive, both directly from stool specimens (2 series; 57-61%) and in isolated CD strains (53%); consequently the negative predictive value remained low (84-93% and 91%, respectively). Specificity, however, was very good at 98-100%. The 2 molecular methods detected CD toxin genes excellently and equally, resulting in sensitivities, specificities, and positive and negative predictive values of 98%, 100%, 100%, and 98%, respectively. Both molecular assays were easy to use, rapid, sensitive, and specific for the detection of toxigenic CD strains.