Abstract
Asymptomatic Clostridium difficile colonization is common in hospitalized patients. Existing C. difficile assay comparisons lack data on severity of diarrhea or patient outcomes, limiting the ability to interpret their results in regard to the diagnosis of C. difficile infection (CDI). The objective of this study was to measure how including patient presentation with the C. difficile assay result impacted assay performance to diagnose CDI. Stool specimens from 150 patients that met inclusion and exclusion criteria were selected. Nine methods to detect C. difficile in stool were evaluated. All patients were interviewed prospectively to assess diarrhea severity. We then assessed how different reference standards, with and without the inclusion of patient presentation, impact the sensitivity, specificity, and positive and negative predictive values of the assays to diagnose CDI. There were minimal changes in sensitivity; however, specificity was significantly lower for the assays Tox A/B II, C. diff Chek-60, BD GeneOhm Cdiff, Xpert C. difficile, and Illumigene C. difficile and for toxigenic culture (P was <0.01 for all except Tox A/B II from fresh stool, for which the P value was 0.016) when the reference standard was recovery of toxigenic C. difficile from stool plus the presence of clinically significant diarrhea compared to when the reference standard was having at least four assays positive while ignoring diarrhea severity. There were 15 patients whose assay result was reported as negative but subsequently found to be positive by at least four assays in the comparison. None suffered from any CDI-related adverse events. In conclusion, clinical presentation is important when interpreting C. difficile diagnostic assays.