Key points are not available for this paper at this time.
Context.— Laboratory testing practices for diagnosis of Clostridioides difficile infection (CDI) have evolved in response to published guidelines, availability of highly sensitive nucleic acid amplification tests (NAATs), perceived problems with the specificity of NAATs, and CDI reporting requirements. Objective.— To assess the current state of laboratory practice for diagnostic CDI testing. Design.— An optional 8-item supplemental questionnaire was distributed in December 2019 to the 1374 laboratories participating in the College of American Pathologists C difficile Detection (CDF) proficiency testing program challenge CDF-C. Results.— Of 1374 CDF-C participants, 1160 (84.4%) responded, predominantly representing laboratories based in the United States (1077 of 1160; 92.8%). The majority reported using a multistep testing algorithm (684 of 1159; 59.0%). Initial testing with a glutamate dehydrogenase and toxin A/B combination test followed by NAAT for discrepant results was the most common testing method (360 of 1146; 31.4%). NAAT alone (299 of 1146; 26.1%) was next, then NAAT followed by an assay that included toxin A/B enzyme immunoassay if NAAT is positive (258 of 1146; 22.5%). Only 5.4% (62 of 1146) reported using toxin A/B immunoassay alone. Most respondents (1093 of 1131; 96.6%) reported rejecting CDI tests on formed stool, but rejection of CDI testing in pediatric patients was uncommon (211 of 1131; 18.7%). Rejection of CDI testing in patients using laxatives was reported more often by US-based respondents (379 of 1054 36.0% versus 9 of 77 11.7%, P .001). Conclusions.— Multistep algorithms for CDI diagnosis are widely used in line with published recommendations. Most respondents reported rejection of formed stool for CDI testing, but few reported rejection of testing in infants and patients taking laxatives, suggesting these may be areas of opportunity for laboratories to pursue in improving CDI testing practices.
Sullivan et al. (Mon,) studied this question.