|Home | About | Journals | Submit | Contact Us | Français|
The diagnosis of Clostridium difficile infection continues to be a challenge for many clinical microbiology laboratories. A new lateral flow assay, the C.Diff Quik Chek Complete assay, which tests for the presence of both glutamate dehydrogenase (GDH) and C. difficile toxins A and B, was evaluated for its ability to diagnose C. difficile disease. The results of this assay were compared to those of both PCR and toxigenic culture. The results showed that this assay allows 88% of specimens to be accurately screened as either positive (both tests positive) or negative (both tests negative) for the presence of toxigenic C. difficile in less than 30 min and with minimal hands-on time. Use of a random-access PCR for the analysis of specimens with discrepant results (one test positive and the other negative) allows the easy, rapid, and highly sensitive (100%; 95% confidence interval [CI], 89.6 to 100%) and specific (99.6%; 95% CI, 97.3 to 99.9%) diagnosis of C. difficile disease. The use of this algorithm would save institutional costs, curtail unnecessary isolation days, reduce the nosocomial transmission of disease, and increase the quality of care for patients.
The laboratory diagnosis of Clostridium difficile disease has evolved significantly over the last several years, and many tests that may be used to assist with the detection of C. difficile infection are now available. These assays include enzyme immunoassays (EIAs), lateral flow tests, PCR assays, tissue culture cytotoxicity neutralization tests, and toxigenic culture. Many recent papers have reported on the use of different algorithms that use the tests mentioned above to allow the better diagnosis of C. difficile disease (4, 5, 8, 18, 19, 20, 23, 26). Many of these approaches incorporate cytotoxicity neutralization (CTN) assays or anaerobic agar culture with identification of the organism, followed by toxin testing. However, many laboratories do not have the technical expertise, facilities, or training to perform CTN assays (which are labor-intensive and somewhat subjective), and an anaerobic agar culture with toxin detection may take several days; both of these methods delay the reporting of results (14, 17, 19, 20, 21, 29). The use of PCR for the diagnosis of this disease has been shown to be very specific and sensitive but often does not allow for random-access (i.e., real-time) results and can be quite costly to perform as a stand-alone testing method (2, 6, 10, 12, 15, 17, 27, 28). There is a real need for a rapid and simple testing strategy for the accurate diagnosis of C. difficile disease. With this in mind, we examined a simple, rapid two-step algorithm for the laboratory diagnosis of Clostridium difficile disease utilizing the new C.Diff Quik Chek Complete assay, a lateral flow kit that uses a combination of glutamate dehydrogenase (GDH) antigen detection plus toxin A and B detection for initial screening and a random-access, highly sensitive Xpert C. difficile PCR assay which detects the toxin B gene for samples with discrepant results.
The study described here was prospectively conducted at the Northwest Kaiser Permanente regional laboratory in Portland, OR. The regional laboratory performs C. difficile diagnostic testing for two medical centers and 20 clinics. Testing was performed with nonformed stool specimens from May to November 2009. The specimens were transported and stored at 4°C until they were assayed, and testing was performed with fresh specimens daily for all assay systems. Of the 261 patients included in the study, only 21 patients provided two samples and 1 patient provided three samples for evaluation.
Briefly, 284 samples (from 261 patients) were received for the diagnosis of C. difficile infection and were tested by four assays, resulting in five results: a GDH antigen-specific EIA (GDH-EIA) that tests for glutamate dehydrogenase (C.Diff Chek-60; Techlab, Blacksburg, VA); a lateral flow assay for toxins A and B (LF-TOX assay; C.Diff Quik Chek; Techlab); the C.Diff Quik Chek Complete lateral flow assay (Techlab), which tests for both GDH (COMP-GDH assay) and toxins A and B (COMP-TOX assay); and the Xpert C. difficile PCR assay (Cepheid, Sunnyvale, CA), a random-access, rapid PCR assay for the detection of the toxin B gene (tcdB), which is associated with toxigenic C. difficile. All assays were performed and the results were interpreted according to the manufacturers' instructions, including the use of appropriate controls, as specified by each company. A sample was determined to be positive for C. difficile disease if it was positive by at least one test for GDH, at least one test for toxin, and the Xpert C. difficile PCR assay. Patients determined to be negative for C. difficile disease tested negative by all five tests.
All discrepant specimens (not determined to be positive or negative by the tests described above) were subjected to enriched toxigenic culture, and a part of the sample was frozen at −20°C on the same day. Toxigenic culture was performed as described previously (20). The specimens were first subjected to alcohol shocking by adding a small portion of the stool specimen to equal amounts of 95% ethanol for 15 min. Three to 4 drops of these treated specimens were added to chopped meat anaerobic broth (Anaerobe Systems, Morgan Hill, CA) and incubated for 24 to 48 h. The inoculated chopped meat broths (CMBs) were filtered through a 0.45-μm-pore-size filter (Fisher Scientific, Pittsburgh, PA) and tested for toxin by the LF-TOX assay after 24 h of incubation, and if the result was negative, they were tested again after 48 h of incubation. If the CMBs were negative by this assay, they were subcultured onto cycloserine-cefoxitin-fructose agar (CCFA) anaerobic plates (Anaerobe Systems), incubated for up to 7 days, and observed daily for growth. C. difficile organisms isolated from CCFA cultures were subcultured into chopped meat broth (C-CMB), incubated for 24 and 48 h, and then tested as described above. In addition, samples that were positive by the Xpert C. difficile PCR assay but negative by the enriched toxigenic cultures described above were sent for referral testing to the R. M. Alden Research Laboratories (Culver City, CA). Briefly, freshly frozen stool samples were thawed, placed directly into TAL broth (cycloserine-cefoxitin-mannitol broth with taurocholate and lysozyme; Anaerobe Systems), and then ethanol treated and plated directly onto CCFA-HT agar plates (cycloserine-cefoxitin-fructose agar with horse blood and taurocholate; Anaerobe Systems). If the specimens directly plated on CCFA-HT were negative for the growth of C. difficile, the TAL broths were subcultured onto CCFA-HT agar plates to look for organisms. If organisms grew at any time, they were directly tested for toxin production by the C. difficile Tox A/B II EIA (Techlab).
Several GDH-positive specimens were frozen at −20°C immediately after they were tested (within ~30 h after collection). After the initial experimentation was concluded, these specimens were then removed from the freezer, thawed at room temperature, and retested by the two assays for GDH.
Two hundred eighty-four stool specimens were tested by all five tests. Twenty-seven specimens (9.5%) tested positive by at least one GDH test, at least one toxin test, and the Xpert C. difficile PCR assay (25 of these 27 specimens were positive by all five tests); and these specimens were considered positive without further testing. In addition, 224 (78.9%) specimens were negative by all five tests and were considered negative without further testing (Table (Table1).1). Thirty-three samples (11.6%) that did not meet the criteria mentioned above, including samples positive only for the GDH antigen, were considered discrepant and were subjected to in-house toxigenic culture. On the basis of toxigenic culture testing, 15 of these discrepant specimens were resolved to be positive, 12 of these samples were resolved by in-house toxigenic culture testing, and 3 were resolved by referral toxigenic culture testing (Table (Table1).1). A total of 18 specimens were resolved to be negative: 15 by in-house toxigenic culture and 3 by referral toxigenic culture testing (no nontoxigenic C. difficile isolates were found) (Table (Table1).1). Thus, of the 284 samples tested, 42 were considered positive (prevalence, 14.8%) and 242 were considered negative for toxigenic C. difficile. On the basis of these data, the assays tested had the following sensitivities and specificities: GDH-EIA, 100% and 94.2%, respectively; LF-TOX assay, 59.5% and 99.2%, respectively; COMP-GDH assay, 97.6% and 94.6%, respectively; COMP-TOX assay, 61.9% and 99.2%, respectively; and the Xpert C. difficile PCR assay, 100% and 99.6%, respectively. When the C.Diff Quik Chek Complete assay was used and two results which were deemed to be indeterminant (GDH negative and toxin positive [these two specimens were PCR negative]) were discarded, as recommended by the manufacturer, the sensitivity and the specificity were 60.0% and 99.6%, respectively. Use of the C.Diff Quik Chek Complete assay with reflex to the Xpert C. difficile PCR assay to test any discrepant samples (GDH positive/toxin negative or GDH negative/toxin positive) (Fig. (Fig.1)1) demonstrated a sensitivity and a specificity of 100% and 99.6%, respectively, for the diagnosis of toxigenic C. difficile infection (Table (Table22).
As reported by others, we found that both toxin assays (the LF-TOX and COMP-TOX assays) had poor sensitivities (59.5% and 61.9%, respectively), although they were both highly specific (1, 3, 4, 5, 16). However, both GDH assays proved to be highly sensitive in our study, having sensitivities of from 97.6 to 100%. This agrees with the findings of other studies that concluded that the test for GDH is excellent for screening (5, 17, 18, 20) but is in contrast to the findings of others who have reported that GDH assays have sensitivities of from 70 to 88% compared to the results of toxigenic culture and/or PCR (1, 3, 8, 10, 12, 16, 21). These different sensitivities might possibly be due to regional/geographical differences in strain ribotypes that affect the GDH assays (indicating the possibility of a limited number of strains in our area) or may be due to the freezing and thawing of the stool samples prior to testing by the GDH assay. Some of the previously cited studies, which exhibited lower sensitivities for the GHD assay, performed this testing after the stool specimens had been frozen and then thawed (8, 12). In order to gain some insight into whether a freeze-thaw cycle might possibly account for the differences in the sensitivities of GDH assays that have been published, we thawed and tested 27 GDH-positive specimens that had been frozen within a few hours after they tested and that were positive by both the GDH-EIA and the COMP-GDH assay. For the GDH-EIA, 26 of the 27 specimens again tested positive, for a loss of 3.7% positivity; for the COMP-GDH assay, 24 of the 27 specimens tested positive, with 3 specimens testing negative (a loss of 11.1% positivity). Although these results are based on the results for a small number of retested samples, if a laboratory is using fresh stool specimens, either assay can be used; however, if frozen samples are to be used for testing for GDH, the GDH-EIA appears to be preferred.
Of the Xpert C. difficile PCR assay-positive discrepant samples in our study, 15/16 were toxigenic culture positive. The one sample with discrepant results by the Xpert C. difficile PCR assay that was not positive by toxigenic culture was from a patient from whom an additional sample was collected on the same day and which did resolve to a positive result by toxigenic culture. In all likelihood, this was also a truly positive sample but was considered to have a negative result for this study. These results might be because the patient received therapy for C. difficile infection, because of sampling discrepancies, or perhaps because of the less than 100% sensitivity of toxigenic culture. Even with this one false-positive result compared to the result of toxigenic culture, this Xpert C. difficile PCR assay could certainly be used as a stand-alone assay for the detection of C. difficile disease, as it has a sensitivity and a specificity of 100% and 99.6%, respectively. All patients in this study would have received the appropriate diagnosis from the results of the Xpert C. difficile PCR assay alone, and the results would have been available in a much more timely fashion compared to current toxigenic culture methods. The results of the Xpert C. difficile PCR assay would allow the more rapid institution of therapy as well as the implementation of the appropriate isolation procedures for patients who test positive, decreasing the spread of disease through nosocomial transmission. These data coincide with those from previously published studies regarding the performance of PCR assays for the detection of C. difficile infections (3, 7, 8, 10, 12, 21, 25).
The two-step algorithm used in this study, which utilizes a highly sensitive lateral flow GDH assay in combination with a highly specific lateral flow toxin assay (the C.Diff Quik Chek Complete assay), allows the detection of virtually all C. difficile-negative specimens within 30 min. Those specimens with discrepant results (GDH positive/toxin negative or GDH negative/toxin positive) would reflex to Xpert C. difficile PCR testing, and the results would be available within an additional 1 h. The C.Diff Quik Chek Complete assay is very simple to perform and permitted the very rapid reporting of final results for up to 88% of our specimens. For those samples with discrepant results that would reflex in this work flow to PCR testing, the Xpert C. difficile PCR assay is a simple and quick, random-access PCR method that provides confirmatory results in an expedient manner. These data show that although the Xpert C. difficile PCR assay method could be used as a stand-alone testing procedure, the price per test would add significant cost to the laboratory, and an initial capital expenditure is necessary for purchase of the instrument. Use of the C.Diff Quik Chek Complete assay as the first step in this algorithm to eliminate the need for secondary testing for up to 88% of specimens would vastly decrease the cost associated with confirmatory testing by PCR.
The turnaround time from specimen collection to the reporting of results for our current routine testing (batch testing by the GDH-EIA once per day and, if the result is positive, reflexing to the LF-TOX test and then, finally, reflexing to toxigenic broth culture if the LF-TOX test is negative) would range from a minimum of ~30 h to up to ~3 days. Our institution implements contact isolation upon the order of a diagnostic test for C. difficile, and deterring the costs associated with C. difficile disease has become a significant issue for a majority of medical institutions (9, 11, 13, 22). Implementation of a new testing algorithm by use of the C.Diff Quik Chek Complete assay with reflex to the Xpert PCR for discrepant results would allow patient results to be available within 3 h after the specimen reaches the laboratory and would save approximately 85% of annual isolation costs in our medical center (as only 15% of patients for whom the test is ordered truly have C. difficile disease). Adoption of this new algorithm not only will result in an up to 3-day reduction in the turnaround time for the reporting of results but also will increase our ability to detect C. difficile-positive patients, as it is more sensitive than our current method (20).
A limitation of this study is that only discrepant samples were further analyzed by the toxigenic culture assay, with assumptions being made that if samples are positive for the GDH antigen, for the toxin A and B antigens, and by the Xpert C. difficile PCR, they would also be positive by toxigenic culture; and those that were negative by all five assays would have been negative by toxigenic culture. This is a reasonable assumption, based on the known performance characteristics of the each of these assays (17, 24), and similar assumptions have been made in a previously published study (5); however, this is still a limitation of our study that must be taken into account. Our study was strengthened by use of a respected referral laboratory for testing of the Xpert C. difficile PCR assay-positive samples that were negative by in-house toxigenic culture, which resulted in three additional samples being identified as true positives. The referral laboratory had the added advantage of using culture directly onto CCFA plates, in addition to anaerobic broth culture. Another limitation to our study could relate to the relatively small sample size of 284 specimens. However, although the studies of both Quinn et al. (17) and Swindells et al. (24) (discussed below) did test all specimens by comparison with a known “gold standard,” our sample included more than 100 additional specimens compared to the numbers used in those two studies.
Our results agree with those of Quinn and colleagues (17) and Swindells and colleagues (24), who both compared the C.Diff Quik Chek Complete assay to PCR. Both groups found that the C.Diff Quik Chek Complete assay had sensitivities of 78.3% and from 61.1 to 73.3%, respectively, whereas our study found a sensitivity of 60%. This difference could be due to the number of specimens in each evaluation, as we tested additional specimens, or again, it could be due to geographical differences associated with the different ribotypes of the organisms. In addition, during our study we found that one specimen tested toxin A/B positive and GDH negative by the C.Diff Quik Chek Complete assay, but such results were not seen by Quinn et al. (17) or Swindells et al. (24), which, again, is likely associated with the additional specimens that we evaluated. Our findings also agree with the results of the studies of Quinn et al. (17) and Swindells et al. (24) on the specificity of the C.Diff Quik Chek Complete assay, for which our data also showed a specificity of >99%. Quinn and colleagues concluded that for a reasonable expense, an FDA-cleared, user-friendly product is not currently available for confirmatory testing of C. difficile disease (17). In contrast, we found that the Xpert C. difficile PCR assay, when it was used with the C.Diff Quik Chek Complete assay, is a product that will allow the real-time, confirmatory testing of discrepant results, finally revolutionizing laboratory testing for C. difficile infection, as also discussed by Swindells and colleagues (24). This algorithm will also allow timely decisions for the appropriate placement of patients into isolation to be made.
In summary, use of the combination of the C.Diff Quik Chek Complete assay with reflex to Xpert C. difficile PCR testing for discrepant results provides a rapid, easy, and cost-effective means of accurately diagnosing C. difficile disease. Use of a test to screen for both GDH and toxins A and B will allow the laboratory to detect more samples with true-positive and true-negative results (>88%) without having to test these specimens by the more expensive PCR assay. Use of only a test for GDH and testing of all GDH-positive samples by PCR, as well as use of only a test for toxins A and B and testing of all toxin A/B-negative samples, would unnecessarily increase the number of PCR tests. The material costs per test for each of these assays are $11.50 (reimbursement cost, $34.36) for the C.Diff Quik Chek Complete assay (hands-on time, ~9 min per specimen) and $33.38 (reimbursement cost, $50.27) for the Xpert C. difficile PCR assay (hands-on time ~5 min per specimen). Thus, use of the C.Diff Quik Chek Complete assay as a screen, followed by use of the Xpert C. difficile PCR assay if the C.Diff Quik Chek Complete assay results are discrepant, would keep the expense for diagnosis much lower than the use of PCR alone and would be more accurate than the use of testing for GDH and/or toxins A and B alone (Table (Table2).2). Use of a GDH-EIA format and, if the result of that assay is positive, reflexing to the LF-TOX assay and then reflexing to the Xpert C. difficile PCR assay, if necessary, would give similar findings. Choosing the appropriate algorithm should be based on the specimen volume, work flow needs, and the clinical impact. One must keep in mind that if any assay is used in the batch mode, the turnaround time will be lengthened. Rapid reporting of results for C. difficile toxin will assist infection control with the isolation of patients with active disease, save the institution the costs associated with isolation, as well as save patients from unnecessarily being placed into isolation.
We thank Diane Citron of the R. M. Alden Research Laboratories for additional testing of discrepant samples, Inverness Medical and Cepheid for providing materials for this study, and Kaiser Permanente infection control practitioners Anne Eades and Dana Trocino for contact isolation information.
Published ahead of print on 7 April 2010.