Enzyme immunoassays (EIAs) based on monoclonal antibodies for the detection of Clostridium difficile toxins have recently been developed for clinical use. The aim of this study was to compare three commercially available EIAs, two for toxin A (Premier C. difficile Toxin A; Meridian, Osi, Elancourt, France; and Vidas C. difficile Toxin A; bioMérieux, Marcy l'Etoile, France) and one for toxins A and B (Cytoclone A + B EIA; Cambridge Biotech Corp., Codiapharm, Evian, France), with a cytotoxicity assay and toxigenic culture for the diagnosis of C. difficile-associated diarrhea (CDAD). The study was performed with 285 fresh stools from 285 patients with suspected CDAD. In case of disagreement, the tests were repeated on a frozen aliquot of the same stool sample, and the patient's chart was reviewed. CDAD diagnosis was established in 55 cases (incidence, 19.3%). The sensitivities and specificities of the methods were, respectively, 92.7 and 100% for the cytotoxicity assay, 96.4 and 99.1% for toxigenic culture, 75.5 and 97.8% for Cytoclone, 65.4 and 99.6% for Premier, and 65.4 and 100% for Vidas. The results were uninterpretable in 3.2% of cases with Cytoclone, 0.3% with Premier, and 2.5% with Vidas. We conclude that the cytotoxicity assay and toxigenic culture remain the best methods for the diagnosis of CDAD even though they lack standardization and require 48 to 96 h to obtain the result. Despite their rapidity and simplicity, EIAs are not sensitive enough to be relied on as the sole laboratory test.