Abstract
Prevention and management of Clostridium difficile infection (CDI) can be improved by rapid and reliable diagnostics. The Vidas C. difficile glutamate dehydrogenase assay had performance comparable to that of the Quik Chek-60 assay (overall agreement, 95%) and a sensitivity of >93%; thus, it is suitable as the first test in two-stage algorithms for a CDI diagnosis.
TEXT
Accurate and timely diagnosis of Clostridium difficile infection (CDI) is a key step in optimizing patient management and reducing cross-infection risk, and diagnosis has historically relied on the detection of C. difficile toxins in fecal samples (1). An alternative target, C. difficile-specific glutamate dehydrogenase (GDH), first identified in 1991 by Lyerly et al., has been shown to be highly conserved between PCR ribotypes of C. difficile (2, 3). Current United Kingdom and European guidance recommends GDH as a possible first assay in a two-stage diagnostic algorithm for CDI, most commonly alongside toxin detection (1, 4, 5). The poor prognostic value of current C. difficile toxin enzyme immunoassays (EIA) means that they should not be used as standalone assays for the diagnosis of CDI (6–8).
We prospectively tested fecal samples routinely submitted for C. difficile testing between July 2012 and January 2013 to laboratories in one United Kingdom hospital (Leeds Teaching Hospital NHS Trust, Leeds, United Kingdom) and two U.S. hospitals (Wishard Health Services, Indianapolis, IN, USA, and Tricore Reference Laboratory, Albuquerque, NM, USA). Samples included were <3 days old and had been refrigerated to follow the international good practices for C. difficile diagnosis. All samples, once made anonymous, were tested at the receiving laboratory using a new enzyme-linked fluorescence assay (ELFA), the Vidas C. difficile GDH assay (bioMérieux, France), a comparator GDH EIA, Quik Chek-60 (Techlab, USA), and two culture methods. Samples were frozen at −70°C before shipping to Leeds for testing with an in-house PCR assay for the GDH gene gluD. Samples used in this service evaluation were residual diagnostic material and did not require ethical approval or consent in the United Kingdom. In the United States, approval of the ethics committee (institutional review board) was granted, while the requirement for informed consent was waived.
Fecal samples were directly inoculated onto C. difficile chromID culture media (bioMeriéux, France). In addition, fecal samples were alcohol shocked in 50% alcohol before inoculation on to cycloserine-cefoxitine fructose agar (CCFA) (Remel, USA). All plates were incubated at 37°C in an anaerobic environment for 24 h (chromID) or 48 h (CCFA) before inspection for suspect colonies (according to the manufacturers' instructions). The identity was confirmed with a Microgen latex agglutination kit (Microgen, United Kingdom) for the United Kingdom site or with Gram stain and Prodisc kits (Remel, USA) for U.S. sites. Both commercial immunoassays were performed as per the manufacturers' instructions.
In-house PCR specific for C. difficile gluD was performed at Leeds on samples that had been frozen at −70°C, either at Leeds or before transportation from the other sites. Briefly, samples were defrosted and then diluted 1/10 in 1 ml STAR buffer (Roche, Germany) with the addition of 1/10 chloroform before being spun at 16,000 × g for 10 min in a centrifuge. An internal control (Yersinia ruckeri) was added to each sample before DNA was extracted on the QiaXtractor using the DX kit (Qiagen Ltd., United Kingdom). Template DNA was added to the Brilliant QPCR multiplex master mix (Agilent, United Kingdom) along with primers and probes for gluD and Yersi (Table 1). Amplification was performed on a Stratagene MX3000P (Agilent, United Kingdom) using the following thermocycling conditions: 95°C for 10 min followed by 45 cycles at 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s. A previous evaluation determined that a gluD cycle time value of <35 cycles indicated a positive PCR result (data not shown).
TABLE 1.
Sequence of primers and probes used in GDH (gluD) PCR assaya
| Oligonucleotide name | Sequence | 5′ modificationb | 3′ modificationc |
|---|---|---|---|
| Yersi F1 | GGAGGAAGGGTTAAGTGTTA | ||
| Yersi R1 | GAGTTAGCCGGTGCTTCTT | ||
| Yersi P1 | GCGAGTAACGTCAATGTTCAGTGC | Cy5 | BHQ2 |
| gluD F3 | GTCTTGGATGGTTGATGAGTAC | ||
| gluD R2 | TTCCTAATTTAGCAGCAGCTTC | ||
| gluD P1 | AAGCCAGTTGAATTTGGTGG | FAM | BHQ1 |
Shown are the sequences of primer and probes, and their modifications, that were used in the GDH PCR assay at Leeds. Yersi primers and probes and gluD primers were described previously (9, 10), and the gluD probe was designed in house (Leeds).
Cy5, cyanine 5; FAM, 6-carboxyfluorescein.
BHQ, black hole quencher.
In total, 1,914 samples were tested during the study; 1,906 had complete data for all the assays and were used for comparisons. Over half of the samples (53.1%) came from patients aged >60 years (Table 2), with slightly more women than men (ratio, 1.42 females:1 male). Liquid feces made up 62% of the total samples tested across all three sites (Table 2); only one site (Tricore) tested formed fecal samples (2.5% of the total).
TABLE 2.
Baseline characteristics of patients included in the study
| Characteristic | No. (%) of patients at: |
|||
|---|---|---|---|---|
| All sites | Leeds | Wishard | Tricore | |
| Total | 1,914 (100.0) | 524 (27.4) | 466 (24.3) | 924 (48.3) |
| Age | ||||
| Child <2 yr | 3 (0.2) | 0 (0.0) | 3 (0.6) | 0 (0.0) |
| Child 2–12 yr | 79 (4.1) | 12 (2.3) | 19 (4.1) | 48 (5.2) |
| Adolescent 13–21 yr | 58 (3.0) | 11 (2.1) | 17 (3.6) | 30 (3.2) |
| Adult 22–59 yr | 757 (39.6) | 149 (28.4) | 217 (46.6) | 391 (42.3) |
| Adult ≥60 yr | 1,017 (53.1) | 352 (67.2) | 210 (45.1) | 455 (49.2) |
| Age class | ||||
| Pediatric (<22 yr) | 140 (7.3) | 23 (4.4) | 39 (8.4) | 78 (8.4) |
| Adult (≥22 yr) | 1,774 (92.7) | 501 (95.6) | 427 (91.6) | 846 (91.6) |
| Gender | ||||
| Female | 1,123 (58.7) | 296 (56.5) | 267 (57.3) | 560 (60.7) |
| Male | 790 (41.3) | 228 (43.5) | 199 (42.7) | 363 (39.3) |
| Nature of specimen | ||||
| Formed | 47 (2.5) | 0 (0.0) | 0 (0.0) | 47 (5.1) |
| Liquid | 1,186 (62.0) | 380 (72.5) | 363 (77.9) | 443 (47.9) |
| Semiformed | 681 (35.6) | 144 (27.5) | 103 (22.1) | 434 (47.0) |
Of the two reference culture media used, the chromID C. difficile medium was more sensitive than the Remel CCFA medium (67.0% versus 62.5%) compared with GDH PCR. The 95% confidence interval (CI) of the difference between the sensitivities of the assays did not include zero (difference, −4.5%; 95% CI, −7.5% to −1.4%), indicating a significant difference between the sensitivity of the two culture media (11).
The Vidas C. difficile GDH assay was 93.0% sensitive and 91.8% specific compared with chromID C. difficile (Table 3). The assay was slightly more sensitive compared with Remel CCFA (95.8%) but, conversely, was less specific (90.0%) (Table 3). The overall levels of agreement of the Vidas GDH assay with chromID C. difficile and Remel CCFA were 92% and 91%, respectively. There was geographical variance in the performance of the assay; sensitivity was lower at Leeds Teaching Hospitals NHS Trust but not statistically different from the two U.S. sites, while specificity was statistically lower at Tricore (data not shown). It should be noted, however, that neither medium performed as well as the Vidas C. difficile GDH assay compared with the GDH PCR assay (Table 4). In a three-way comparison, the 95% CI of the differences between sensitivities of the assays did not include zero, indicating a significant difference of the sensitivity of the Vidas C. difficile GDH assay versus both culture media (Table 4) (11). The sensitivity of the Vidas C. difficile GDH assay did not alter significantly if the equivocal GDH PCR results were treated as positive, negative, or void (Table 5). While the Vidas C. difficile GDH assay had higher sensitivity, it was significantly less specific than both culture media (Table 4), indicating that this assay may be useful only as a screening assay (for example, as part of a two-step diagnostic algorithm).
TABLE 3.
Sensitivities and specificities of GDH EIAs in comparison with those of two commercial selective C. difficile culture media
| Comparator assay (total no. of samples tested) | Result of comparator assay (no. of samples) |
EIAa |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Vidas C. difficile GDH |
Quik Chek-60 |
|||||||||||
| Positive | Negative | Equivocal | Invalid | Result (no. of tested samples) | Sensitivity (% [95% CI]) | Specificity (% [95% CI]) | Overall agreement (% [95% CI]) | Result (no. of tested samples) | Sensitivity (% [95% CI]) | Specificity (% [95% CI]) | Overall agreement (% [95% CI]) | |
| ChromID medium (1,914) | 357 | 1,557 | 0 | 0 | Positive (332); negative (1,429) | 93.0 (89.9–95.2) | 91.8 (90.3–93.0) | 92.0 (90.7–93.1) | Positive (332; negative (1,505) | 93.0 (89.9–95.2) | 96.7 (95.6–97.4) | 96.0 (95.0–96.8) |
| Remel CCFA (1,914) | 313 | 1,601 | 0 | 0 | Positive (300); negative (1,441) | 95.8 (93.0–97.6) | 90.0 (88.4–91.4) | 91.0 (89.6–92.2) | Positive (303); negative (1,520) | 96.8 (94.2–98.3) | 94.9 (93.8–95.9) | 95.2 (94.2–96.1) |
Overall agreement of the Vidas C. difficile GDH assay (bioMérieux, France) or the Quik Chek-60 assay (Techlab, USA) when compared with two different reference culture media for C. difficile (comparator assays).
TABLE 4.
Three-way comparison between Vidas GDH assay and commercial media using GDH PCR as the reference methoda
| Assay performance | Vidas GDH |
Remel CCFA |
ChromID C. difficile |
|||||
|---|---|---|---|---|---|---|---|---|
| % | 95% CI | % | 95% CI | GDH minus Remel (95% CI) | % | 95% CI | GDH minus ChromID (95% CI) | |
| Sensitivity | 71.2 | 66.7 to 75.3 | 62.5 | 57.8–67.0 | 8.7 (5.6 to 11.9) | 67.0 | 62.4–71.3 | 4.2 (1.7 to 6.8) |
| Specificity | 89.4 | 87.7 to 90.9 | 96.8 | 95.8–97.6 | −7.4 (−8.9 to −6.1) | 95.1 | 93.9–96.1 | −5.7 (−7.2 to −4.3) |
The two C. difficile reference culture media are compared to the GDH PCR assay. The analysis presented used GDH PCR equivocal results as positive results.
TABLE 5.
Effect of analyzing equivocal GDH PCR results as positive, negative, or void on the performance of the Vidas C. difficile GDH assay using AUROC analysisa
| Equivocal GDH PCR result | AUROC of Vidas C. difficile GDH assay compared with GDH PCR (%) | 95% CI |
|---|---|---|
| Positive | 0.83 | 0.80–0.86 |
| Negative | 0.86 | 0.83–0.88 |
| Voidb | 0.86 | 0.84–0.89 |
AUROC, area under receiver operator curve.
GDH PCR equivocal results analyzed as void were removed from the analysis.
In this multicenter comparison study, we found that the Vidas C. difficile GDH assay was comparable in performance to the commercially available GDH EIA (Quik Chek-60, Techlab, USA), with an overall agreement of 95% (Table 6). The Vidas assay has the advantage of being automated, with good traceability and more comprehensive quality control than the Quik Chek. It is, however, slower (40 min run time) and requires a larger sample volume (200 μl versus 25 μl).
TABLE 6.
Agreement between the Vidas GDH assay and the commercial comparator assay Quik Chek-60 or the GDH gluD PCR assay
| Comparator assay (total no. of samples tested)a | Result of comparator assay (no. of samples) |
Vidas C. difficile GDH |
||||||
|---|---|---|---|---|---|---|---|---|
| Positive | Negative | Equivocal | Invalid | Result (no. of tested samples) | Positive agreement (% [95% CI]) | Negative agreement (% [95% CI]) | Overall agreement (% [95% CI]) | |
| Quik Chek 60 (1,914) | 384 | 1,530 | 0 | 0 | Positive (374); negative (1,444) | 97.4 (95.3–98.6) | 94.4 (93.1–95.4) | 95.2 (94.2–96.1) |
| GDH PCR (1,906)b | ||||||||
| Equivocal as positive | 424 | 1,482 | 45 | 0 | Positive (302); negative (1,325) | 71.2 (66.7–75.3) | 89.4 (87.7–90.9) | 85.4 (83.7–86.9) |
| Equivocal as negative | 379 | 1,527 | 45 | 0 | Positive (289); negative (1,454) | 76.3 (71.7–80.3) | 88.9 (87.2–90.3) | 86.4 (84.7–87.8) |
Shown are the levels of agreement between the Vidas C. difficile GDH assay (bioMérieux, France) and the commercial comparator assay Quik Chek 60 (Techlab, USA) or the GDH PCR assay.
Results for the comparison of the Vidas C. difficile GDH assay (bioMérieux, France) and the GDH PCR assay have been analyzed with the equivocal GDH PCR results included as either a positive or a negative result.
There are some limitations to our study. We did not use a gold standard reference method for a CDI diagnosis; however, GDH alone cannot reliably diagnose CDI and is diagnostic only in conjunction with a toxin detection assay (6). Culture and PCR for the gluD gene, therefore, are more representative comparators when assessing GDH detection assays. This does pose some difficulties when comparing our results with other publications, as most incorporated a reference method for diagnosing CDI. Notably, however, a meta-analysis showed that, for the three studies that compared a GDH assay with culture, the sensitivities of the GDH assays examined were 95.0%, 93.4%, and 93.5%, that is, comparable with the Vidas C. difficile GDH assay studied here (12).
GDH assays have proved to be useful as the first assay of a two-stage algorithm for CDI diagnosis (4, 5). It is important to emphasize that the second assay should detect toxins A and B of C. difficile, as detection of the toxin has been shown to correlate with both mortality and severity of infection (6, 13). The Vidas C. difficile GDH assay is a sensitive method that makes it suitable as a first assay in these recommended diagnostic algorithms.
ACKNOWLEDGMENTS
This work was supported by funding from bioMérieux (France).
We thank Matthew Phelan for his assistance with the statistical analysis.
M.H.W. reports grants and personal fees from Actelion; grants and personal fees from Cubist, Astellas, Merck, Optimer, Sanofi Pasteur, and Summit during the conduct of the study; personal fees from AstraZeneca, Durata, Nabriva, Novacta, Pfizer, Roche, Basilea, The Medicines Company, and VH Squared; grants and personal fees from Cerexa, Abbott, bioMérieux, Da Volterra, and European Tissue Symposium; and other disclosures from Alere apart from the submitted work. K.A.D. reports grants from Astellas during the conduct of the study. C.E.B., K.A.M., R.S., S.Y., T.E.D., D.D.F., and R.J.B. have no conflicts to report.
REFERENCES
- 1.Crobach MJ, Dekkers OM, Wilcox MH, Kuijper EJ. 2009. European Society of Clinical Microbiology and Infectious Diseases (ESCMID): data review and recommendations for diagnosing Clostridium difficile-infection (CDI). Clin Microbiol Infect 15:1053–1066. doi: 10.1111/j.1469-0691.2009.03098.x. [DOI] [PubMed] [Google Scholar]
- 2.Lyerly DM, Barroso LA, Wilkins TD. 1991. Identification of the latex test-reactive protein of Clostridium difficile as glutamate dehydrogenase. J Clin Microbiol 29:2639–2642. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Carman RJ, Wickham KN, Chen AM, Lawrence AM, Boone JH, Wilkins TD, Kerkering TM, Lyerly DM. 2012. Glutamate dehydrogenase (GDH) is highly conserved among Clostridium difficile ribotypes. J Clin Microbiol 50:1425–1426. doi: 10.1128/JCM.05600-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.United Kingdom Department of Health. 2014. Clostridium difficile: updated guidance on diagnosis and reporting. http://www.dh.gov.uk/en/Publicationsandstatistics/Publications/PublicationsPolicyAndGuidance/DH_132927.
- 5.Wren MWD, Sivapalan M, Kinson R, Shetty NP. 2009. Laboratory diagnosis of Clostridium difficile infection. An evaluation of tests for fecal toxin, glutamate dehydrogenase, lactoferrin and toxigenic culture in the diagnostic laboratory. Br J Biomed Sci 66:1–5. [DOI] [PubMed] [Google Scholar]
- 6.Planche T, Davies K, Coen P, Finney J, Monahan I, Morris K, O'Connor L, Oakley S, Pope C, Wren M, Shetty N, Crook D, Wilcox M. 2013. Differences in outcome according to Clostridium difficile testing method: a prospective multicentre diagnostic validation study of C. difficile infection. Lancet Infect Dis 13:936–945. doi: 10.1016/S1473-3099(13)70200-7. [DOI] [PMC free article] [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 and cytotoxigenic culture methods. J Clin Microbiol 47:3211–3217. doi: 10.1128/JCM.01082-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Planche T, Aghaizu A, Holliman R, Riley P, Poloniecki J, Breathnach A, Krishna S. 2008. Diagnosis of Clostridium difficile infection by toxin detection kits: a systematic review. Lancet Infect Dis 8:777–784. doi: 10.1016/S1473-3099(08)70233-0. [DOI] [PubMed] [Google Scholar]
- 9.Lund M, Nordentoft S, Pedersen K, Madsen M. 2004. Detection of Campylobacter spp. in chicken fecal samples by real-time PCR. J Clin Microbiol 42:5125–5132. doi: 10.1128/JCM.42.11.5125-5132.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Paltansing S, van den Berg RJ, Guseinova RA, Visser CE, van der Vorm ER, Kuijper EJ. 2007. Characteristics and incidence of Clostridium difficile-associated disease in The Netherlands, 2005. Clin Microbiol Infect 13:1058–1064. doi: 10.1111/j.1469-0691.2007.01793.x. [DOI] [PubMed] [Google Scholar]
- 11.Clinical and Laboratory Standard Institute. 2008. User protocols for evaluation of qualitative test performance; approved guideline—2nd ed CLSI document EP12-A2. Clinical and Laboratory Standard Institute, Wayne PA. [Google Scholar]
- 12.Shetty N, Wren MWD, Coen PG. 2011. The role of glutamate dehydrogenase for the detection of Clostridium difficile in fecal samples: a meta-analysis. J Hosp Infect 77:1–6. doi: 10.1016/j.jhin.2010.07.024. [DOI] [PubMed] [Google Scholar]
- 13.Longtin Y, Trottier S, Brochu G, Paquet-Bolduc B, Garenc C, Loungnarath V, Beaulieu C, Goulet D, Longtin J. 2013. Impact of the type of diagnostic assay on Clostridium difficile infection and complication rates in a mandatory reporting program. Clin Infect Dis 56:67–73. doi: 10.1093/cid/cis840. [DOI] [PubMed] [Google Scholar]