Q J Med 2003; 96: 579-582
© 2003 Association of Physicians
Reducing delays in the diagnosis and treatment of Clostridium difficile diarrhoea
From the Department of Gastroenterology, Wycombe Hospital, High Wycombe, UK
Received 18 November 2002 and in revised form 29 April 2003
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Background: The diagnosis of Clostridium difficile diarrhoea is often delayed.
Aim: To assess and reduce delays in diagnosis and treatment
Design: Two-part study: retrospective audit then prospective observational.
Methods: The retrospective study audited cases positive for C. difficile culture or toxin A between June 2000 and January 2001. Cases were reviewed regarding demographic characteristics and the timing from onset of symptoms to testing and treatment (n = 27). In the prospective study, efforts were made to increase the awareness of medical staff about C. difficile diarrhoea, and testing for C. difficile toxin A assay was substituted for stool culture and external toxin analysis. Data were prospectively analysed for demographic characteristics and time from onset of symptoms to testing and treatment (n = 24).
Results: There were no significant differences in age (76 vs. 78 years), male to female ratio (1:2 vs. 1:4), reason for admission, specialty responsible, or mortality rate (40% vs. 37%) during the presenting admission. Preceding antibiotic use, the combination of antibiotics used and subsequent treatment also did not differ between groups. The time from onset of diarrhoea to sampling was reduced from 4.7 days (range 3–30) to 0.8 days (range 0–5; p 
0.05). The time from sampling to treatment was reduced from 7.7 days (5–34) to 2.8 days (2–9; p 0.05).
Discussion: The combination of increased awareness of C. difficile diarrhoea and the change to in-house toxin testing significantly reduced the time from onset of symptoms to sampling and treatment.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Clostridium difficile is a well-described nosocomial infection. It causes up to 20% of antibiotic-associated diarrhoea, and can display a clinical spectrum from an asymptomatic carrier state, through diarrhoea without colitis, and diarrhoea with non-pseudomembraneous colitis, to pseudomembraneous colitis and fulminant colitis.1 The infection is almost always related to recent or previous use of antibiotics. Elderly and patients with co-morbid disease are more susceptible to the infection.2 Hospital stay is prolonged in patients with C. difficile infection. They have an increased mortality prior to discharge, although a direct cause-effect relationship is often not apparent.3 Stringent measures of infection control and careful use of antibiotics can reduce the rate of infection,1 but depend on early detection of infected individuals. The diagnosis is usually made from the stool of infected patients using rapid enzyme immunoassays to toxin A or B. These tests have a high sensitivity and specificity and can be done within hours.4 However, some hospitals, as in our case, still use primary in-house stool culture with subsequent external toxin assay for confirmation of the diagnosis of C. difficile diarrhoea. Following Department of Health guidelines, this practice has now become much less common.5 Culture usually needs 2–3 days, which may lead to delay and thus affect outcome. We studied the implications of this approach for the timing of diagnosis and treatment of C. difficile diarrhoea. We prospectively compared this with the effects of primary in-house toxin assay for C. difficile in combination with a program to increase awareness about this infection.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
We analysed the delays at each stage of diagnosis in patients with C. difficile diarrhoea and sought to improve aspects of care, where this could be done reasonably simply. We undertook a retrospective audit of all cases positive for C. difficile culture or toxin A between June 2000 and January 2001. The clinical notes and microbiological records were analysed for demographic data obtained on all patients using the hospital database. In 27/48 cases, we were able to obtain the clinical notes and microbiology record. These cases were analysed for demographic data, use of antibiotics in the preceding month and the timing from the onset of diarrhoea to sampling, availability of culture/toxin results and treatment.
Following the retrospective study, in-house C. difficile toxin A testing (Oxoid) was introduced as the primary diagnostic test. This has a reported sensitivity of 91% and a specificity of 98%.6 The awareness about C. difficile was raised by presentations to junior and senior medical staff (Grand round and medical audit meetings). A prospective study of cases identified as being C. difficile toxin-positive was undertaken between March and August 2001 and analysed in the same way as the retrospective data.
Data were compared using Students t-test. Patients remained under the clinical teams caring for them.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
In the retrospective study, 48 patients were identified. Demographic data were obtained for 27 of 48 cases from the notes, and showed a mean age of 76 years (range 24–95), with a male to female ratio of 1:2. Underlying diagnoses were most commonly stroke, pneumonia, obstructive airways disease and cardiac failure. Mortality as the final outcome of hospital stay in the group as a whole was 40% as (16/42), and antibiotics had been used in 97% of subjects in the preceding month. Commonly-used antibiotics were augmentin and cefuroxime. A combination of antibiotics was used in 38% (10/27) of patients. Mean time from onset of diarrhoea to sampling was 4.7 days (range 3–30). Mean time from sampling to positive culture was 3.1 days (range 1–7). Mean time from sampling to a definitive diagnosis was 9.3 days (range 6–18). Mean time from onset of diarrhoea to treatment was 7.7 days (range 5–34).
In the prospective study, 24 patients were analysed. Demographic data were obtained from the notes, and showed a mean age of 79 years (range 38–93) with a male to female ratio of 1:1.4. Underlying diagnoses were most commonly stroke, pneumonia, obstructive airways disease and cardiac failure. Mortality as the final outcome of their hospital stay in the group as a whole was 37% (9/24), and antibiotics had been used in 100% of subjects in the preceding month. Commonly-used antibiotics were again augmentin and cefuroxime. A combination of antibiotics was used in 45% (11/24) of patients. Mean time from onset of diarrhoea to sampling was 0.8 days (range 0–5). Mean time from sampling to a definitive diagnosis was 1.9 days (range 1–4). Mean time from onset of diarrhoea to treatment was 2.8 days (range 2–9).
Overall, the time from onset of diarrhoea to treatment was reduced from a mean of 7.7 days (range 5–34) to a mean of 2.8 days (2–9; p 0.05). There was no difference in the number of patients treated with metronidazole (63% vs. 62%). Patients who did not receive metronidazole for C. difficile diarrhoea had improved spontaneously, been discharged or died prior to diagnosis (Tables 1 and 2).
|
|
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
This study showed that there is a substantial delay in diagnosis of C. difficile in many patients. Our retrospective study was suggestive that this amounted to 9.3 days from onset of diarrhoea to definitive diagnosis. We have demonstrated that a substantial part of this delay was because patients with possible C. difficile diarrhoea were not identified early as being at risk, and there was a delay in sampling stool cultures. As a result, we changed policy to suggest that all patients who had diarrhoea following a course of antibiotics in the preceding month should have stool cultures sent to the lab as soon as possible, and it should not be hoped that this was a transient event which would resolve spontaneously. The second area of delay was that there is necessarily a delay in using stool culture as a primary screening test since it takes 2–3 days to detect C. difficile. In addition, not every patient who has C. difficile will be toxin positive or have C. difficile as the cause for their diarrhoea. This may lead to over treatment, but our study did not analyse this aspect of care. By introducing an in-house toxin A assay, we showed a significant reduction in time to determine an accurate diagnosis. We almost certainly avoided treating some patients inappropriately. Although not quantified, it is likely that the additional cost of the test is compensated for by a reduction in lab costs and a reduction in unnecessary treatment costs. Having adopted this policy, it is important however to remember that some patients may develop C. difficile diarrhoea from low levels of toxin production, which can potentially be missed on toxin assays. Where there is a high index of clinical suspicion of C. difficile diarrhoea, Clostridium should be cultured even if the toxin assay is negative. We identified no such cases in our small study. Other modes of diagnosis such as endoscopic and histological confirmation of pseudomembranous colitis remain an essential part of the workup of patients with unexplained diarrhoea. Recent studies have suggested an increase in the percentage of toxin-A-negative/B-positive strains, which involve a minimum of 3% of cases in the UK. It would therefore be sensible to use toxin assays that test for toxin A and B, as suggested by the PHLS Anaerobe Reference Unit.8 A local survey in 2001 showed that these recommendations are only being slowly implemented: 4/9 hospitals in our region only tested for toxin A (unpublished data).
Although our study is small it seems to be representative of patients with C. difficile diarrhoea. The demographic characteristics of our patients do not differ from those described in the literature.1–3,8
This study shows that a simple change in the diagnostic workup of C. difficile diarrhoea from the primary in-house stool culture and subsequent external toxin assay to primary in-house C. difficile toxin assay in combination with a program to raise the awareness about C. difficile diarrhoea can significantly reduce both the time from onset of diarrhoea to sampling and the time from onset of diarrhoea to treatment. Our study was not designed to look at outcome measures such as hospital stay and mortality, which did not differ in the two groups. The total number of cases with C. difficile diarrhoea in the prospective study was half of those in the retrospective study (24 vs. 48). This may be related to seasonal variation, with increased use of antibiotics for pneumonias in the winter months, but might also reflect an impact of the earlier detection of cases with earlier use of barrier nursing. Routine screening of hospital patients for C. difficile has not been advocated, as treatment of the asymptomatic carrier state is not recommended.1,9 However, the early detection of cases with C. difficile diarrhoea is vital in preventing hospital outbreaks. Bender et al.10 found that 30% of patients on a chronic care ward were C. difficile-positive during an outbreak, while the corresponding figure was 5% on a ‘skilled nursing facility’. Standard infection control procedures, including hand washing, use of disposable gloves and barrier nursing, are important in the prevention of spread and can be re-enforced by early diagnosis.11 In our study, presentations aimed at junior and senior doctors led to a significant reduction in the time from onset of diarrhoea to sampling. Where doctors change their posts frequently, it might be necessary to introduce a nurse-led protocol to facilitate and sustain early sampling of stool in patients who develop diarrhoea on broad-spectrum antibiotics. Specific antibiotic policies could be used in conjunction with measures to facilitate early detection and have halved the number of C. difficile-associated diarrhoeas in a large teaching hospital.12
|
Acknowledgments |
|---|
Our thanks to Dr Lyons and Dr Waghorn and the Department of Microbiology at Wycombe Hospital for their help.
|
Footnotes |
|---|
Address correspondence to Dr M.B. Frenz, Department of Gastroenterology, Wycombe Hospital, High Wycombe, Bucks HP11 2TT. e-mail: mfrenz{at}doctors.org.uk
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Kelly CP, Pothoulakis C, LaMont JT. Clostridium difficile colitis. N Engl J Med 1994; 330:257–62.
2. Rudensky B, Rosner S, Sonnenblick M, Van Dijk Y, Shapira E, et al. The prevalence and nosocomial acquisition of Clostridium difficile in elderly hospitalised patients. Postgrad Med J 1993; 69:45–7.
3. Wilcox MH, Cunniffe JG, Trundele C, Redpath C. Financial burden of hospital-acquired Clostridium difficile infection. J Hosp Infect 1996; 34:23–30.[CrossRef][Web of Science][Medline]
4. Patel JB, Donahue AM, Nachamkin I. Performance of two rapid, single-use immunoassays for the detection of Clostridium difficile toxin A. Diagnost Microbiol Infect Dis 2001; 39:21–4.[CrossRef][Web of Science][Medline]
5. Department of Health/Public Health Laboratory Service joint working group. Clostridium difficile infection. Prevention and Management. Heywood, BAPS Health Publications Unit, 1994.
6. Bentley AH, Patel NB, Sidorczuk M, Loy P, Fulcher J, et al. Multicentre evaluation of a commercial test for the rapid diagnosis of Clostridium difficile-mediated antibiotic-associated diarrhoea. Eur J Clin Microbiol Infect Dis 1998; 17:788–90.[CrossRef][Web of Science][Medline]
7. Brazier JS, Stubbs SL, Duerden BI. Prevalence of toxin A negative/B positive C. difficile strains. J Hosp Infect 1999; 42:248–9.[Web of Science][Medline]
8. Cheng SH, Lu JJ, Young TG, Perng CL, Chi WM. Clostridium difficile associated diseases: comparison of symptomatic infection versus carriage on the basis of risk factors, toxin production and genotyping results. Clin Infect Dis 1997; 25:157–8.[Web of Science][Medline]
9. Bartlett JG. Management of Clostridium difficile infection and other antibiotic-associated diarrhoeas. Eur J Gastroenterol Hepatol 1996; 8:1054–61.[Web of Science][Medline]
10. Bender BS, Bennett R, Laughton BE, Greenouhg WB 3rd, Gaydos C, et al. Is Clostridium difficile endemic in chronic care facilities? Lancet 1986; ii:11–13.
11. Settle CD, Wilcox MH. Review article: antibiotic-induced Clostridium difficile infection. Aliment Pharmacol Ther 1996; 10:835–41.[CrossRef][Web of Science][Medline]
12. Ludlam H, Brown N, Sule O, Redpath C, Coni N, et al. An antibiotic policy associated with reduced risk of Clostridium difficile-associated diarrhoea. Age Ageing, 1999; 28:578–80.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||