OBJECTIVE: Performance of a combined approach for the detection of toxigenic strains in patients suspected of having Clostridium difficile-associated disease was evaluated. METHODS: In this approach, stools were cultured for 24 h on a selective medium supplemented with sodium taurocholate (TCCFA), in anaerobic conditions created with the Martreg Anoxomat system, and toxin A detection was performed directly on C. difficile colonies, by enzyme immunoassay (EIA). This method was compared with three others: cytotoxigenic culture consisting of a 48-h culture on selective medium followed by detection of in vitro cytotoxin production on cell monolayers, fecal cytotoxin detection and fecal toxin A detection by EIA. RESULTS: From 548 stools, 108 yielded a positive culture by at least one of the methods, and 81 isolates were cytotoxin producers. Cultures for 24 h on TCCFA were positive in 106 cases and EIA performed on colonies gave 73 positive results, giving a sensitivity of 90.1% and a specificity of 100%. By comparison, the sensitivity and specificity of cytotoxigenic culture, stool cytotoxin and stool toxin A were respectively 96.2% and 100%, 61.7% and 100%, and 66.7% and 95.9%. CONCLUSIONS: Performing EIA on colonies recovered after 24 h culture allows us to improve the detection of toxigenic strains in patients suspected of having C. difficile-associated disease.