Real-time polymerase chain reaction correlates well with clinical diagnosis of Clostridium difficile infection

https://doi.org/10.1016/j.jhin.2014.03.005Get rights and content

Summary

Aim

To determine the clinical utility of a rapid molecular assay for Clostridium difficile infection (CDI) in an acute hospital setting.

Methods

From March to September 2011, stool specimens from inpatients in two acute hospitals with suspected CDI were tested prospectively by routine cell culture cytotoxin neutralization assay (CCNA), real-time polymerase chain reaction (PCR) using the GeneXpert (Cepheid Inc., Sunnyvale, CA, USA), and a dual testing algorithm [glutamate dehydrogenase (GDH)/toxin enzyme immuno-assay, Premier, Launch Diagnostics, Longfield, UK]. All patients with positive PCR, CCNA or discrepant results were reviewed by a multi-disciplinary team (treating clinician, gastroenterologist, microbiologist and infection control nurse).

Results

C. difficile detection rates were 11.7% (PCR), 6% (CCNA) and 13.8% (GDH). Out of 1034 stool specimens included in the study, 974 (94.1%) had concordant CCNA and PCR results. Eighty-nine percent (886/985) had concordant CCNA, PCR and GDH results, and 94.4% (930/985) had concordant GDH and PCR results. Using clinical diagnosis as the reference, PCR had sensitivity of 99.1%, specificity of 98.9%, positive predictive value (PPV) of 91.9% and negative predictive value (NPV) of 99.9%. CCNA on a single sample had sensitivity of 51%, specificity of 99.4%, PPV of 91.9% and NPV of 94.3%. GDH had sensitivity of 83.8%, specificity of 94.5%, PPV of 64.7% and NPV of 97.9%. Almost twice as many patients were positive by PCR compared with CCNA (121 vs 62); 54/59 of those with discrepant results were clinically confirmed as CDI.

Conclusion

Rapid diagnosis of CDI using PCR was timely, accurate and correlated well with clinical diagnosis.

Introduction

Clostridium difficile causes nosocomial antibiotic-associated diarrhoea, with a range of mild to severe disease, pseudo-membranous colitis, toxic megacolon and a potentially fatal outcome.1, 2 The epidemiology of CDI is changing. More recently, patients without traditional risk factors, younger patients and patients in the community have been diagnosed more frequently with CDI. In a recent study, a large proportion (47.6%) of undiagnosed cases of CDI occurred due to lack of clinical suspicion, especially in younger patients and in the community.3 Thirty-day case fatality with CDI has been observed to be higher when there is infection with a hypervirulent strain.4 Recent literature suggests that ribotype analysis alone may not be a good predictor of severe CDI.5

The diagnosis of CDI in clinical laboratories has traditionally been performed by toxin detection. This was done using either enzyme immuno-assay (EIA) for Toxins A/B or cell culture cytotoxin neutralization assay (CCNA) which, historically, has been the gold standard test for C. difficile detection. The detection limits for Toxin A/B EIA are much higher than those for CCNA. Recently, toxin EIAs have been reported to be insensitive and non-specific.6, 7, 8, 9 Toxigenic culture may be more sensitive but it introduces long delays in the reporting of results.

Recent advances in the detection of C. difficile in human stool samples include development of in-house PCR methods, new commercial molecular assays [BD GeneOhm Cdiff PCR (BD Diagnostics/GeneOhm Sciences, San Diego, CA, USA), ProGastro Cd PCR (Gen-Probe Prodesse, San Diego, CA, USA), Cepheid Xpert C. difficile PCR (Cepheid Inc., Sunnyvale, CA, USA), Illumigene C. difficile (Meridian Bioscience, Inc., Cincinnati, OH, USA)] and glutamate dehydrogenase (GDH) detection assays.9, 10 These new methods are more sensitive than Toxin A/B EIA and have been evaluated in several laboratories for the development of local testing algorithms. Tests for GDH, an enzyme produced by C. difficile and other organisms, require additional analysis due to the lack of specificity in order to determine whether a toxigenic strain of C. difficile is truly present.9

In one study, Xpert C. difficile testing yielded the highest sensitivity and NPV, in the shortest time, of all the individual and multiple test algorithms evaluated.11

In order to control CDI efficiently, especially during outbreaks or if strains with increased virulence are circulating, there is a need for rapid identification and separation of patients with and without CDI. This can only be achieved with a reliable diagnostic assay that provides rapid and accurate results, ideally without the need for batch testing.

This study aimed to assess whether a new rapid random access commercial PCR test, Xpert C. difficile, correlates well with clinical CDI diagnosis and can be relied upon as a single diagnostic assay in the laboratory. PCR was compared with CCNA and a dual GDH/Toxin A/B EIA algorithm.

Section snippets

Methods

This study was approved by the Public Health Wales Research and Development Committee. Ethical approval was not deemed necessary as the specimens were requested routinely in accordance with the Abertawe Bro Morgannwg University Health Board (ABM UHB) C. difficile care pathway for clinical diagnosis and management, and no additional specimens were collected for study purposes.

This was a prospective study, conducted at two acute hospital sites within ABM UHB. Eligibility criteria included

Results

In total, 1051 PCR tests were performed, of which 33 (3.14%) initially gave an invalid result. The test was repeated for 22 of these 33 samples, all of which gave a valid result on the second run. Eleven specimens were excluded as there was insufficient specimen to repeat the PCR test.

Another six specimens were excluded from the analysis, as one was a duplicate sample from the same episode, two had invalid CCNA results, one had an invalid PCR result but insufficient quantity of specimen to

Discussion

As described previously,11, 12 this study found that PCR was more sensitive than CCNA (99.1% vs 51%) and GDH (99.1% vs 83.8%) for the detection of CDI when using clinical diagnosis as the reference. The sensitivity of CCNA was lower than expected on single specimens. CCNA had a higher detection rate when multiple specimens were tested within a 14-day period, as approximately one-quarter of these discrepant PCR/CCNA patients tested positive by CCNA on further sampling, although these subsequent

Acknowledgements

The authors wish to thank the nurses and doctors from Singleton and Morriston Hospitals, Dr. Mike Isaac, Katherine Davies, Ian Thomas, Hilary Williams, Chris Jones, Angela Farr, Lynne Ray and all the staff of PHW Microbiology ABM (Swansea), the ABM Health Board Infection Control Team and the pharmacists for their support. This study was presented, in part, at the Federation of Infection Societies Scientific Meeting, 16th–18th November 2011, Manchester, UK; Welsh Association for Gastroenterology

References (20)

There are more references available in the full text version of this article.

Cited by (0)

View full text