Abstract
We compared the Qiagen artus C. difficile QS-RGQ kit, a new nucleic acid amplification test for the detection of Clostridium difficile toxins in stool specimens, with the Cepheid Xpert C. difficile test. The sensitivity, specificity, positive predictive value, and negative predictive value for the Qiagen artus C. difficile QS-RGQ test were 100%, 89.5%, 60.9%, and 100%, and those for the Cepheid Xpert C. difficile test were 100%, 90%, 62.2%, and 100%, respectively.
TEXT
Diagnosis of Clostridium difficile infection (CDI) is based on clinical presentation and laboratory tests. In recent years, nucleic acid amplification tests (NAATs) for the direct detection of C. difficile toxin genes in stool samples have become commercially available and are a highly sensitive alternative to the less sensitive enzyme immunoassays (EIAs), the time-consuming toxigenic culture (TC), and the cell culture cytotoxicity neutralization assay (CCNA). Qiagen has released a CE-IVD-marked (European certification for in vitro diagnostic devices) and FDA-cleared assay for the extraction and simultaneous detection of C. difficile toxin A and toxin B genes in stool specimens using the QIAsymphony RGQ platform. In this study, we compared the Qiagen artus C. difficile QS-RGQ kit (Qiagen artus test) with the Cepheid Gene Xpert C. difficile test (Cepheid Xpert test) and toxigenic culture. To our knowledge, this is the first evaluation of the Qiagen artus C. difficile QS-RGQ kit.
(These data were presented in part as a poster at the 54th Interscience Conference on Antimicrobial Agents and Chemotherapy, 5 to 9 September 2014, Washington, DC.)
From January 2014 to May 2014, 201 loose stool specimens submitted for C. difficile testing from inpatients at the University Hospital Cologne, a 1,400-bed tertiary-care facility, were included in the study. All stool samples were analyzed upon delivery or after overnight storage at −20°C. Specimens were analyzed by the Qiagen artus C. difficile QS-RGQ kit and the Cepheid Gene Xpert C. difficile test strictly according to the manufacturers' instructions. As a reference method, every stool sample was processed for enriched toxigenic culture (TC). The stool sample was plated on cycloserine-cefoxitin-fructose agar (CCFA) after alcohol shock. Suspicious colonies were confirmed by matrix-assisted laser desorption ionization–time of flight (MALDI-TOF) mass spectrometry. Toxin detection was performed using the toxin A/B enzyme immunoassay (EIA) (Ridascreen Clostridium difficile toxin A/B assay; r-biopharm). Isolates negative for toxin A/B by EIA were further tested by PCR (in-house PCR adapted from that described by Houser et al. [1], the Cepheid Xpert test, and the Qiagen artus test). Samples negative for TC and positive for only one NAAT were defined as discrepant samples. Stored DNA from discrepant samples was retested with the Qiagen artus test. Retesting with the Cepheid Xpert test was not possible, as the original stool samples were not stored. In addition, the charts of patients with discrepant results were reviewed to clarify the diagnosis of CDI. We determined diarrhea (defined as three or more loose stools in 24 h) and the clinical C. difficile score by using a clinical prediction scale for hospital-onset CDI (2).
From the 201 stool samples, two samples were reported as invalid by the Cepheid Xpert test (1%). Both were negative by TC and the Qiagen artus test. We excluded these samples from performance calculation. No sample was reported as invalid by the Qiagen artus test. Of the remaining 199 samples, 28 (14%) were positive by TC and both of the NAATs. In addition, 21 samples were positive by at least one NAAT but remained negative by TC. In 11 of these samples, no growth was detected on CCFA. For 10 samples, suspicious colonies on CCFA were identified by MALDI-TOF as species other than C. difficile and toxin A/B EIA results were negative. For one sample, suspicious colonies were identified as C. difficile but the toxin A/B EIA and toxin PCR results for this isolate remained negative. Of the 21 TC-negative samples, 14 were positive by both PCR assays and 7 were positive by one PCR assay only (Qiagen artus test, n = 4; Cepheid Xpert test, n = 3). No culture-positive, PCR-negative sample was observed. The sensitivities, specificities, positive predictive values, and negative predictive values were, respectively, 100%, 89.5%, 60.9%, and 100% for the Qiagen artus test and 100%, 90%, 62.2%, and 100% for the Cepheid Xpert test, compared to TC (Table 1). The overall agreement between the two NAATs was excellent (kappa, 0.9). Of the 46 stool samples that were positive by the Qiagen artus test, 40 were reported tcdA positive and tcdB positive, five were reported tcdA positive and tcdB negative, and one sample was reported tcdA negative and tcdB positive.
TABLE 1.
Comparison of the Qiagen artus C. difficile QS-RGQ kit and the Cepheid Xpert C. difficile test to toxigenic culture
| Assay and result | No. of samples with indicated result by toxigenic culture |
Performancea compared with toxigenic culture |
||||
|---|---|---|---|---|---|---|
| Positive | Negative | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | |
| Qiagen artus C. difficile QS-RGQ kit | 100 | 89.5 | 60.9 | 100 | ||
| Positive | 28 | 18 | ||||
| Negative | 0 | 153 | ||||
| Total (n = 199)b | 28 | 171 | ||||
| Cepheid Xpert C. difficile test | 100 | 90 | 62.2 | 100 | ||
| Positive | 28 | 17 | ||||
| Negative | 0 | 154 | ||||
| Total (n = 199)b | 28 | 171 | ||||
PPV, positive predictive value; NPV, negative predictive value.
Two samples were excluded from analysis due to invalid results by the Cepheid Gene Xpert C. difficile test.
Stored DNA eluates from the seven discrepant samples were retested by the Qiagen artus test. Cycle threshold (CT) values of the internal controls were not different between the first and second runs (30.52 ± 1.7 versus 29.48 ± 0.33). Results of the retest and of the retrospective analysis of patient charts are shown in Table 2.
TABLE 2.
Characterization of specimens with discrepant results by the Qiagen artus C. difficile QS-RGQ kit and the Cepheid Gene Xpert C. difficile test
| Specimen | Resulta of: |
CDI scoreb | |||
|---|---|---|---|---|---|
| Initial test |
Qiagen artus retest | ||||
| Qiagen artus | Cepheid Xpert | TC | |||
| p88 | tcdB+, tcdA+ | Negativec | Negative | Negativec | Positive |
| p101 | tcdB+, tcdA+ | Negativec | Negative | Negativec | Positive |
| p104 | Negativec | tcdB+, RT027+, cdt+ | Negative | tcdB+, tcdA− | Positive |
| p125 | tcdB+, tcdA− | Negativec | Negative | tcdB−, tcdA+ | Negative |
| p132 | tcdB−, tcdA+ | Negativec | Negative | Negativec | Negative |
| p176 | Negativec | tcdB+, RT027−, cdt− | Negative | Negativec | Positive |
| p188 | Negativec | tcdB+, RT027−, cdt− | Negative | Negativec | Positive |
Qiagen artus, Qiagen artus C. difficile QS-RGQ assay; Cepheid Xpert, Cepheid Gene Xpert C. difficile test; TC, toxigenic culture. tcdB+ or tcdB−, gene target for toxin B detected or not detected, respectively; tcdA+ or tcdA−, gene target for toxin A detected or not detected; RT027+ or RT027−, tcdC gene deletion at nucleotide 117 indicating ribotype 027 detected or not detected; cdt+ or cdt−, binary toxin gene detected or not detected.
The clinical plausibility of C. difficile infection (CDI) was determined using a clinical prediction scale for hospital-onset CDI (2).
Negative for all genes tested in the indicated assay.
This is the first study evaluating the performance characteristics of the Qiagen artus C. difficile QS-RGQ kit in a routine clinical setting. As described previously (1, 3–5), we found that PCR was more sensitive than TC. This has also been observed when CCNA was used as a reference method (6). Nevertheless, the use of a less sensitive EIA for toxin detection in TC can lead to falsely low sensitivity of TC (7, 8). In our study, we identified only one sample that was positive for C. difficile by culture but negative by toxin EIA. This isolate was also negative for toxin A/B by PCR. Therefore, we consider the 14 samples with positive results by the two PCR assays but negative results by TC in our study to be truly positive for toxigenic C. difficile. As all NAATs may detect colonization instead of infection, an accurate clinical diagnostic workup is mandatory to interpret positive results. In our study, seven samples were positive by one NAAT only. Five of the seven patients were clinically diagnosed as having CDI. Retesting of the stored DNA eluates of these samples by the Qiagen artus test resulted in new discrepancies compared to the results of TC and clinical outcomes (Table 2). The reason for the discrepancies remains unresolved. Since we could show that the CT values of the internal controls were similar between the two test runs, degradation of DNA in the sample is not likely. DNA in these stool samples is seemingly close to the limit of detection, and discrepant results have to be expected as stochastic events. Individual sample-specific factors (for example, inhibitors) may also be influencing the PCR process. Thus, we propose that retesting of the stored DNA with the Qiagen artus test is not useful to clarify discrepant results. The Qiagen artus test detects two different targets, tcdA (toxin A) and tcdB (toxin B). The contribution of toxin A and toxin B to disease is not well understood. To date only strains that are either tcdA and tcdB positive or tcdA negative and tcdB positive have been shown to be pathogenic in humans (9), although in vitro studies suggest a pathogenic role for tcdA-positive, tcdB-negative isolates (10). In our study, we identified five stool samples with tcdA-positive, tcdB-negative results by the Qiagen artus test. Four of these isolates were positive for tcdB by the Cepheid Xpert test, indicating that a tcdB gene is present but was not detected by the Qiagen artus test. In accordance with previously published data (1), our results suggest that the analytical sensitivity of the tcdA target might be higher than the analytical sensitivity of the tcdB target, leading to an overestimation of the number of tcdA-positive, tcdB-negative strains. In line with the manufacturer's instructions, we therefore recommend that strains be interpreted and reported as positive for toxigenic C. difficile if either target is positive by the Qiagen artus test.
Our data show that under routine conditions, the Qiagen artus C. difficile QS-RGQ kit performs equally as well as the Cepheid Xpert C. difficile test. In our study, both of these NAATs showed identical sensitivities and specificities. Optimization of workflow depends on laboratory environment, clinical demands, and platform availability.
ACKNOWLEDGMENTS
A travel grant was received from IntelligentMDx/Qiagen GmbH for presentation of the data as a poster at ICAAC 2014.
G.P. has received honoraria for lectures on PCR studies from Cepheid, Becton Dickinson, Thermo Fisher, and Qiagen.
REFERENCES
- 1.Houser BA, Hattel AL, Jayarao BM. 2010. Real-time multiplex polymerase chain reaction assay for rapid detection of Clostridium difficile toxin-encoding strains. Foodborne Pathog Dis 7:719–726. doi: 10.1089/fpd.2009.0483. [DOI] [PubMed] [Google Scholar]
- 2.Chandra S, Thapa R, Marur S, Jani N. 2014. Validation of a clinical prediction scale for hospital-onset Clostridium difficile infection. J Clin Gastroenterol 48:419–422. doi: 10.1097/MCG.0000000000000012. [DOI] [PubMed] [Google Scholar]
- 3.Zidaric V, Kevorkijan BK, Oresic N, Janezic S, Rupnik M. 2011. Comparison of two commercial molecular tests for the detection of Clostridium difficile in the routine diagnostic laboratory. J Med Microbiol 60:1131–1136. doi: 10.1099/jmm.0.030163-0. [DOI] [PubMed] [Google Scholar]
- 4.Novak-Weekley SM, Marlowe EM, Miller JM, Cumpio J, Nomura JH, Vance PH, Weissfeld A. 2010. Clostridium difficile testing in the clinical laboratory by use of multiple testing algorithms. J Clin Microbiol 48:889–893. doi: 10.1128/JCM.01801-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Buchan BW, Mackey TL, Daly JA, Alger G, Denys GA, Peterson LR, Kehl SC, Ledeboer NA. 2012. Multicenter clinical evaluation of the Portrait Toxigenic C. difficile assay for detection of toxigenic Clostridium difficile strains in clinical stool specimens. J Clin Microbiol 50:3932–3936. doi: 10.1128/JCM.02083-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Berry N, Sewell B, Jafri S, Puli C, Vagia S, Lewis AM, Davies D, Rees E, Ch'ng CL. 2014. Real-time polymerase chain reaction correlates well with clinical diagnosis of Clostridium difficile infection. J Hosp Infect 87:109–114. doi: 10.1016/j.jhin.2014.03.005. [DOI] [PubMed] [Google Scholar]
- 7.Eastwood K, Else P, Charlett A, Wilcox M. 2009. Comparison of nine commercially available Clostridium difficile toxin detection assays, a real-time PCR assay for C. difficile tcdB, and a glutamate dehydrogenase detection assay to cytotoxin testing and cytotoxigenic culture methods. J Clin Microbiol 47:3211–3217. doi: 10.1128/JCM.01082-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Ylisiurua P, Koskela M, Vainio O, Tuokko H. 2013. Comparison of antigen and two molecular methods for the detection of Clostridium difficile toxins. Scand J Infect Dis 45:19–25. doi: 10.3109/00365548.2012.708780. [DOI] [PubMed] [Google Scholar]
- 9.Carter GP, Rood JI, Lyras D. 2010. The role of toxin A and toxin B in Clostridium difficile-associated disease: past and present perspectives. Gut Microbes 1:58–64. doi: 10.4161/gmic.1.1.10768. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kuehne SA, Cartman ST, Heap JT, Kelly ML, Cockayne A, Minton NP. 2010. The role of toxin A and toxin B in Clostridium difficile infection. Nature 467:711–713. doi: 10.1038/nature09397. [DOI] [PubMed] [Google Scholar]