We compared 3 rapid assays for Clostridium difficile with a cell culture cytotoxicity neutralization assay (CCNA). Of 600 stool samples, 46 were positive for toxigenic C difficile. Both rapid common antigen assays were highly sensitive (91.3%–100%) and, therefore, were appropriate screening tests. The rapid toxin assay had poor sensitivity (61%) but excellent specificity (99.3%). Testing stools for glutamate dehydrogenase (step 1) and those positive with a rapid toxin assay (step 2) would correctly classify 81% of submitted specimens within 2 hours, including during periods of limited staffing (evenings, nights, and weekends). CCNA could then be used as a third step to test rapid toxin–negative samples, thereby providing a final result for the remaining 19% of samples by 48 to 72 hours. The use of rapid assays as outlined could enhance timely diagnosis of C difficile.
Clostridium difficile is the most common cause of antibiotic-associated colitis and of hospital-associated diarrhea. The cell culture cytotoxicity neutralization assay (CCNA) led to the identification of C difficile in 1978, and remains the diagnostic “gold standard” for toxigenic C difficile.1 However, CCNA testing requires technical expertise and up to 72 hours to a final result2,3; therefore, many laboratories still rely on rapid enzyme immunoassays for toxins A and/or B despite sensitivities of 40% to 70%.4 The recent appearance of clinically severe and treatment-refractory epidemic C difficile (also called BI, NAP1, or ribotype 027),5–8 in addition to ongoing disease due to toxin A– and toxin B–producing strains, has made reevaluation of rapid diagnostic tools imperative.
Materials and Methods
To improve rapid diagnosis of toxigenic C difficile, during a 10-month period we tested sequential weekday stool samples submitted to the Johns Hopkins Hospital Clinical Microbiology Laboratory, Baltimore, MD, for suspected C difficile by 3 rapid tests and by CCNA. The consistency of specimens was noted. Trained medical technologists, masked to other results, performed each test and recorded each result. Clinical reporting followed our standard 2-step algorithm, ie, testing stools for C difficile common antigen with confirmation of toxin by CCNA.4
Our current rapid common antigen assay is the C. DIFF CHEK-60 (TechLab, Blacksburg, VA). All stools were tested for C difficile common antigen (glutamate dehydrogenase [GDH]) with the C. DIFF CHEK-60 (TechLab, distributed by Inverness Medical Innovations, Princeton, NJ) per the manufacturer’s instructions. Samples were diluted 1:4 and manually washed using 1× wash solution. Results were quantified by optical density (>0.120 positive) at 450 nm on an ELx800 microplate reader (Bio-Tek Instruments, Winooski, VT).
New rapid common antigen and toxin assays include C. DIFF QUIK CHEK and TOX A/B QUIK CHEK (TechLab). Both assays are membrane-bound lateral flow devices and are self-contained enzyme immunoassays. The C. DIFF QUIK CHEK uses an antibody recognizing GDH, and TOX A/B QUIK CHEK uses antibodies specific for toxins A and B. Assays were performed according to the manufacturer’s instructions. Feces were added to the diluent (same for both assays) and conjugate and the mixture transferred to the sample well. After a 15-minute incubation, the reaction window was washed and substrate added. A blue line (seen by the naked eye) on the control side of the reaction window indicated a valid test, and a blue line on the test side indicated a positive result (captured antigen).
Cell Culture Cytotoxicity Neutralization Assay
All stools were tested by CCNA as described previously.4 Human foreskin fibroblasts (Diagnostic Hybrids, Athens, OH) were seeded onto 96-well tissue culture microtiter plates (5 × 104 cells per well) in fresh 10% minimum essential medium (0.2 mL per well). The plates were incubated at 37°C in 5% carbon dioxide until a confluent monolayer formed. Dilutions of stool (1:2) and positive control samples were prepared with sterile phosphate-buffered saline and antitoxin and incubated for 1 hour at room temperature. Thereafter, 20 μL of each dilution of control toxin, control toxin plus antitoxin, stool filtrate, and stool filtrate plus antitoxin were added to duplicate wells of the plate and incubated overnight at 37°C in 5% carbon dioxide. Plates were read at 24 and 48 hours and at 72 hours if equivocal at 48 hours. Wells with 50% or more cell rounding were considered positive if the cytotoxicity was neutralized by specific antitoxin.
The gold standard for toxigenic C difficile was CCNA. Confidence intervals were calculated using exact methods. Dichotomous variables were compared using the Fisher exact test. Only the first specimen per patient was included to avoid bias related to repeated (correlated) sampling.
Of 600 stool samples tested, 46 (7.7%) were positive for toxigenic C difficile by CCNA. All were liquid or soft. The sensitivity and specificity of our current common antigen assay (C. DIFF CHEK-60) were 91.3% (42/46) and 89.9% (498/554), respectively, and those of the new common antigen assay (C. DIFF QUIK CHEK) were 100% (46/46) and 82.7% (458/554), respectively. The sensitivity of the rapid toxin assay (TOX A/B QUIK CHEK) was only 60.9% (28/46); however, it was highly (99.3%) specific (550/554). The assay’s positive predictive value was high (88%) despite low (7.7%) prevalence Table 1. Using the TOX A/B QUIK CHEK to confirm toxigenic C difficile identified by a C. DIFF QUIK CHEK screen would correctly classify 81.0% (486/600) of specimens within 2 hours and misclassify only 0.7% (4/600); using C. DIFF CHEK-60 to screen would correctly classify 87.7% (526/600) and misclassify 1.3% (8/600).
In comparing 3 rapid assays for C difficile with the gold standard (CCNA), we found that both common antigen assays were highly sensitive and had high negative predictive values, which makes them both appropriate screening tests. The new common antigen assay could replace our current common antigen screening test because it is easier to perform and faster. The poor sensitivity of the rapid toxin assay makes it an inappropriate standalone test; however, its high specificity (99.3%) makes CCNA confirmation of positive results unnecessary. It is important to note that it is much quicker than CCNA (2 hours vs 2–3 days to a positive result) and can, therefore, guide early therapy and aid infection control efforts.
Unique strengths of this study include its large size, clinically relevant and reproducible inclusion criteria (suspected C difficile), and that all specimens were tested by the gold standard (CCNA) rather than only specimens positive by a common antigen assay. We posit that this study design is essential to avoid verification bias. A limitation is that CCNA, although generally considered the diagnostic standard, is not 100% sensitive for C difficile. We recently evaluated stool culture for C difficile compared with CCNA and found that neither culture nor stool CCNA is 100% sensitive compared with a composite standard that includes culture with CCNA testing of isolates as well as stool CCNA.9 Similarly, some GDH assays may miss cases of toxigenic C difficile detected by culture or polymerase chain reaction.10 However, we would argue that comparing the new rapid assays for GDH and toxin with stool CCNA was the appropriate (clinically relevant) comparison because culture with assay for cytotoxin production is generally considered an epidemiologic rather than a clinical tool. Indeed, culture for C difficile is rarely done in the United States because it is labor-intensive and time-consuming and requires expertise not available in most laboratories.
Therefore, we assert that sequential testing with a rapid common antigen assay followed by a rapid toxin assay could allow accurate, speedy (<2 hours) reporting for more than 80% of specimens received in clinical laboratories. In laboratories with limited staffing and equipment, this algorithm could largely obviate the need for laborious, technically difficult CCNA testing. Use of CCNA as a third step would classify the remaining approximately 20% of samples within 48 to 72 hours. Hence, we propose a 3-step algorithm: (1) Screen for C difficile with a common antigen assay (C. DIFF QUIK CHEK, which is preferable to C. DIFF CHEK-60 for speed). (2) Test with a rapid toxin A/B test if the screen is positive. (3) Test with CCNA if the rapid toxin A/B test is negative. The 3-step algorithm could quickly identify 60% of cases of C difficile colitis, thereby supporting early treatment and swift implementation of infection control measures at a reasonable additional cost (~$9.00). Confirmation of toxin production by stool polymerase chain reaction rather than a rapid toxin A/B test or CCNA may ultimately prove best for rapid, accurate diagnosis of C difficile.11 However, molecular assays require technical expertise not available in many smaller laboratories and pose challenges when staffing is limited for all laboratories because manual DNA extraction is still required. Furthermore, they are costly (current list price for a Food and Drug Administration–cleared Prodesse [Gen-Probe Prodesse, Waukesha, WI] or GeneOhm [Becton Dickinson, Franklin Lakes, NJ] kit, $2,375 to $2,500 with per test cost listed as $49.00), especially if batch size is small because clinical requests are few. Hence, they will likely remain more expensive and outside the technical reach of many clinical laboratories for some time.
We thank the virology staff of the clinical microbiology laboratory at Johns Hopkins Hospital for testing the specimens and Inverness Medical Innovations for supplying kits.