Comparison of four laboratory tests for diagnosis of Clostridium difficile-associated diarrhea.

Four different laboratory tests for diagnosis of Clostridium difficile-associated diarrhea were compared to determine the optimal one for management of patients with hospital-acquired diarrhea. Stool samples from 231 patients with diarrhea were tested by the following methods: culture for Clostridium difficile with subsequent determination of exotoxin production, with a toxigenic Clostridium difficile positive (TCP) result considered truly positive; enzyme immunoassay (EIA); latex agglutination test; and an immunobinding blot assay. The rates of positive results were as follows: EIA 5.5%, TCP 7.3%, latex agglutination 16.7%, and immunobinding blot assay 26.1%. Compared to the TCP results, the sensitivity and specificity were, respectively, 61 and 98% for EIA, 47 and 85% for latex agglutination, and 60 and 76% for the immunobinding blot assay. Samples from patients with > or = 6 stools/day were TCP and EIA positive in 27 and 17% of cases, respectively, whereas in patients with < 6 stools/day, these percentages decreased to 2 and 3%, respectively (p < 0.001). In hospitalized patients with > or = 6 stools/day, EIA appears to be the optimal test for diagnosis of Clostridium difficile-associated diarrhea, with a 73% positive predictive value and a 97% negative predictive value. However, in patients with < 6 stools/day, the prevalence of Clostridium difficile is low, and laboratory detection of this organism remains problematic.