TWO-STAGE TESTING NEEDED FOR C DIFF
Patients suspected of carrying the C diff bacteria should have their stool samples tested twice to avoid misdiagnosis, according to researchers at St George?s, University of London and St George?s Healthcare NHS Trust.? Currently, patients across the NHS are diagnosed with a single test, based on the level of a toxin produced by the bacteria in stool samples, but researchers recommended that two different tests are used to avoid getting false positives and missing real cases.?
Today?s findings are based on a review of 18 studies on the effectiveness of six of the most commonly used tests.? However, in related research, St George?s Hospital has been piloting a two-stage testing process over the past six months.
Clostridium difficile, or C diff, is a hospital-acquired infection, which can lead to diarrhoea and severe inflammation of the bowel and can be fatal. A quick and accurate diagnosis is important to halt the spread of the bacteria, alongside good hand hygiene and regular cleaning of wards, bedpans and toilets.? Because C diff, a naturally-occurring micro-organism, can thrive when the normal balance of the gut is upset, patients on antibiotics are at risk.? The elderly are also at risk – over 80% of C diff infected people are over 65.
The St George?s research showed that no one test for C diff is sufficiently accurate. A review of 18 studies on the effectiveness of current C diff tests revealed that as many as 20% of positive results may be false. With 55,393 reported cases in England in the financial year 2007-08, according to the Office for National Statistics, researchers estimate that around 11,000 patients may have been falsely diagnosed with the infection. Tests were also found to be missing up to 20% of real C diff cases.
The studies looked at the six most commonly used testing kits available in the UK and how likely they were to find a positive or negative result, compared with a reference test. The results of the research are to be published on The Lancet Infectious Diseases website on Saturday 1 November and in the December issue of the journal.
Researchers warn that the misdiagnosis of C diff can have very serious consequences. Lead author of the study, Dr Timothy Planche, Consultant in Medical Microbiology, St George?s Healthcare NHS Trust, says: ?A false negative result could mean that infected patients don?t get the right treatment and could pass the infection on to others. Conversely, patients receiving a false positive result may receive inappropriate treatment and be placed in wards along with infected patients, putting them at risk of contracting the infection.?
Whilst the tests were similar in their accuracy, some were more likely to return a false positive result, whereas others were more likely to miss a case. This variation may make it difficult for patients to make meaningful comparisons in rates of incidence between hospital trusts.
The researchers recommend a two-stage testing process, involving two different tests. The first test would be quick test to weed out the negatives. Stool samples returning a positive result on the first test would then be given a second test in order to weed out the false positives. St George?s Hospital and St George?s, University of London are at the forefront of implementing this two-stage approach and are currently analysing its impact.
Corresponding author of the study, Professor Sanjeev Krishna, Professor of Molecular Parasitology and Medicine at St George?s, University of London, says: ?This research is an excellent example of an NHS Trust and a university working together to improve patient care.?
The paper will be discussed at The Lancet Infectious Diseases Conference on Healthcare-associated infections from December 11-12 at the QEII Conference Centre in London.
Notes to editors
For more information or interviews, contact Tamsin Starr at St George?s, University of London Communications on 0208 725 1139; 07786 982028 or at media@sgul.ac.uk.
St George?s, University of Londonprovides training to a full range of more than 2,600 healthcare and sciences students on one site. The college offers courses in medicine and biomedical sciences, as well as midwifery, nursing, physiotherapy, radiography and social work in conjunction with Kingston University. The college is dedicated to promoting by excellence in teaching, clinical practice and research, the prevention, treatment and understanding of disease. It is recognised globally for the quality of its research, with a high reputation in areas such as infection, diseases of the heart and circulation, cell signalling and epidemiology. Other areas of expertise include genetics, health and social care sciences and mental health.?
St George?s Healthcare NHS Trust treated more than 500,000 patients each year from its three sites in south west London. The Trust provides everything you would expect from your local NHS hospital, including AandE, maternity services and care for the young and elderly.? It also has specialised and expert care for some of the most serious and complicated injuries and illnesses including major trauma, neurology, cardiac care, cancer, renal transplantation and stroke.?
Effectiveness of six of the most commonly used tests
This is a digest of the information available in Table 2 and Figure 4 of the study. Please note these are approximate figures after the results of several studies have been combined and there was considerable variation between studies.
| Test | Percentage of true positives missed* | Percentage of positive results that are false%? |
| TechLab Tox A/B II | 16.7% | 12% |
| Meridian Premier | 5.2% | 22% |
| TechLab Tox A/B Quik Chek | 16.1% | 3% |
| Remel Xpect | 18% | 29% |
| Meridian Immunocard | 10.4% | 7% |
| BioM?rieux VIDAS | 24.3% | 45% |
* please see Table 2 ? column “Sensitivity” ? this number is? equivalent to (1-median sensitivity) expressed as a percentage
? this is information from figure 4. We assume a prevalence of C. diff ?of 10% – which is approximately the current situation. The number here is (1-positive predictive value) expressed as a percentage. Read off the graph at a prevalence of 10%
