Abstract
Four commercial enzyme immunoassays (EIAs) for the detection of Clostridium difficile toxin A have recently been developed and marketed (Premier, Meridian Diagnostics, Cincinnati, Ohio; VIDAS, bioMerierux Vitek, Inc., Hazelwood, Mo.; Tox-A-Test, TechLab, Blacksburg, Va.; and Bartels, Baxter Diagnostics, McGaw Park, Ill.). The performances of these EIAs were compared with those of the tissue culture cytotoxicity assay and a definition of C. difficile-associated disease based on both laboratory and clinical criteria for 329 clinical specimens. Two EIAs (Premier and VIDAS) showed good overall agreement (96 and 95%, respectively) with the cytotoxicity assay. However, they were less sensitive (84 and 71%, respectively) than the Bartels (94%) or Tox-A-Test (93%) EIAs. The Bartels and Tox-A-Test assays were much less specific, resulting in poor positive predictive values (56%) of the two assays when compared with that of the cytotoxicity assay. Tox-A-Test had the added drawback of having a significant number of indeterminate results (6.4%). These data indicate that the four EIAs all have specific shortcomings. When using these EIAs, testing strategies that take these shortcomings into consideration should be developed.