Algorithm Combining Toxin Immunoassay and Stool Culture for Diagnosis of Clostridium difficile Infection

Enzyme immunoassays (EIA) to detect glutamate dehydrogenase or toxins A (TcdA) and B (TcdB), a cytotoxicity assay, and bacteriologic culture have disadvantages when applied individually to diagnosis of Clostridium difficile infections. Stool specimens (n = 1,596) were subjected to toxin detection via an enzyme-linked fluorescent immunoassay (ELFA; Vidas CDAB assay) and bacteriologic culture for toxigenic C. difficile in a three-step algorithm with additional toxigenic culture. Isolates (n = 163) from ELFA-negative stool specimens were examined via ELFA for toxin production. We amplified tcdA and tcdB from C. difficile isolates and tcdB from stool specimens that were ELFA positive or equivocal and culture negative, and we compared the results to those obtained with the three-step algorithm. More than 26% of stool specimens (419/1,596) were culture positive, yielding 248 isolates (59.2%) with both toxin genes (tcdA- and tcdB-positive isolates), 88 isolates (21.0%) with either tcdA or tcdB, and 83 (19.8%) that had no toxin genes (tcdA- and tcdB-negative isolates). Among 49 (culture-negative/ELFA-positive or -equivocal) stool specimens, 53.1% (26/49) represented tcdB-positive isolates. Therefore, the total number of PCR-positive cases was 362, and 27.1% (98/362) of these were detected through toxigenic culture. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 63.3%, 96.7%, 90.5%, and 92.4% (ELFA alone); 92.8%, 93.3%, 80.2%, and 97.8% (culture); and 70.7%, 91.4%, 95.5%, and 100% (three-step algorithm ELFA and bacterial culture with toxigenic culture), respectively, with culture and PCR for tcdA and tcdB as the standards. Thus, sensitivity and specificity were highest using culture and ELFA, respectively, but we recommend the three-step algorithm comprising EIA to detect both toxins and toxigenic culture for C. difficile as a practical method for achieving better PPV and NPV.Clostridium difficile is an important nosocomial pathogen, causing antimicrobial-associated diarrhea and pseudomembranous colitis. Toxins A (TcdA) and B (TcdB) mediate the pathogenesis of C. difficile infection (CDI), and toxin detection is an important part of diagnosis. A cytotoxicity neutralization assay (CNA) is the reference method for toxin detection, but it is expensive and time-consuming and requires tissue culture facilities (34, 35). Most laboratories now use a commercial enzyme immunoassay (EIA) to detect TcdA and/or TcdB, with the benefits of rapid turnaround time and ease of use (3, 21, 22, 23, 26, 27, 33, 35). The putative >90% sensitivity of toxin EIAs is not often realized in practice, but EIA is the only toxin detection method available to many routine medical laboratories. The demand for EIA kits detecting both TcdA and TcdB has increased due to increased worldwide prevalence of TcdA-negative, TcdB-positive (TcdA− TcdB+) strains (1, 12, 24, 29, 32).A two-step algorithm, based upon EIA-based detection of species-specific antigen glutamate dehydrogenase (GDH-Ag) and toxin detection via CNA, was suggested to have improved sensitivity and specificity in the detection of toxigenic C. difficile (34). However, the GDH-Ag assay detects both nontoxigenic and toxigenic strains, and the aforementioned shortcomings of the CNA assay make it unavailable to many routine laboratories.Bacteriologic culture can be time-consuming, but it is more straightforward and sensitive than CNA for the detection of toxigenic C. difficile. Furthermore, it provides isolates for characterization, yielding information about CDI epidemiology and antimicrobial susceptibility (11, 28, 36). We evaluated the combination of bacteriologic culture and EIA-based detection of TcdA and TcdB as a new strategy for diagnosis of CDI, especially in areas where TcdA− TcdB+ strains are prevalent.