A total of 320 stool specimens obtained from 262 patients suspected of having Clostridium difficile-associated gastrointestinal disease were examined with two cytotoxicity assays (CTAs) and two commercially available enzyme immunoassays (EIAs). The CTAs were an in-house-developed procedure (University of Massachusetts Medical Center [UMMC], Worcester, Mass.) and a commercial test (Bartels CTA; Baxter Healthcare Corp., West Sacramento, Calif.). One EIA was a monoclonal antibody-based assay for C. difficile toxins A and B (Cambridge Biotech Corp. [CBC], Worcester, Mass.). The other EIA employed monoclonal antibodies directed against only toxin A (Meridian Diagnostics, Cincinnati, Ohio). True-positive and true-negative results were defined on the basis of the results of the four assays, clinical assessments of patients, and the results of other laboratory tests. The sensitivities of the four assays were as follows: Bartels CTA, 100%; UMMC CTA, 97.2%; CBC EIA, 84.5%; and Meridian EIA, 69.0%. The Bartels CTA demonstrated a specificity of 99.2%. The other three assays had a specificity of 100%.