Clostridium difficile associated diarrhoea: diagnosis and treatmentBMJ 2005; 331 doi: https://doi.org/10.1136/bmj.331.7515.498 (Published 01 September 2005) Cite this as: BMJ 2005;331:498
Clostridium difficile associated diarrhoea is a serious condition with a mortality of up to 25% in frail elderly people.1 It affects older, frailer, hospitalised patients and also younger patients who are immunosuppressed.
Cross infection by C difficile is common in neonatal units, but neonates do not seem to develop C difficile associated diarrhoea.
The diagnosis of C difficile associated diarrhoea depends on:
Presence of diarrhoea, defined as an increase in stool liquidity usually accompanied by an increased frequency of bowel motions. A formal cut-off is the passing of more than 300 ml of liquid stool in 24 hours
Detection of toxins produced by C difficile in the stools.
The patient may also experience abdominal pain and have systemic features of malaise, fever, dehydration, and delirium. A pseudomembranous colitis is present in severe cases. In this state there is sloughing of the colonic epithelium, which is severely inflamed due to the cytotoxic action of C difficile.
C difficile associated diarrhoea is classically associated with clindamycin, but it may occur after exposure to a wide range of antibiotics.
Symptoms usually start during antibiotic treatment or shortly afterwards. Symptoms can be delayed by a few weeks, so it is worth asking patients whether they are currently taking, or have recently taken, antibiotics.
How does it happen?
C difficile is spread by the faecal-oral route, albeit indirectly through spores left on surfaces. C difficile is an anaerobic, Gram positive, spore forming bacterium that is the major identifiable cause of antibiotic associated diarrhoea.
In a proportion of those colonised, perhaps around one third,7 C difficile produces toxins that cause diarrhoea. The two principal toxins, A and B, share 63% of amino acid sequence homology and act on small GTP-binding proteins (these are part of the large family of guanine nucleotide binding proteins that are coupled to many membrane receptors and are implicated in many diseases).
Toxin B is around 1000 times more cytotoxic than toxin A.8 Toxin A is also an enterotoxin in that it loosens the tight junctions between the epithelial cells that line the colon, which helps toxin B enter into epithelial cells.
Some strains also produce a binary toxin, but its role in human disease is uncertain.9
C difficile associated diarrhoea is therefore characterised by a progression from an uncolonised state, through to C difficile colonisation, followed by toxin production. In part this depends on the specific strain of C difficile, with one strain (toxigenic S-type 5236) responsible for around 70% of cases in the UK.6 However, host factors that predispose to colonisation and toxin production are equally important.
One systematic review identified several risk factors for C difficile associated diarrhoea10:
Increasing age (excluding infancy)
Severe underlying disease
Non-surgical gastrointestinal procedures
Presence of a nasogastric tube
Receiving anti-ulcer medications
Stay on intensive care unit
Long duration of hospital stay
Long duration of antibiotic course
Receiving multiple antibiotics.
Another systematic review identified clindamycin, cephalosporins, and penicillins as the classes of antibiotics most associated with C difficile associated diarrhoea.11
One study found that the risk factors for colonisation in a medicine of the elderly hospital (age, admission from another hospital, use of non-cephalosporin and cephalosporin antibiotics) were different from the risk factors for the transition between culture-positive and toxin-positive status (antibiotic use only).12
A failure to mount an immune response is associated with colonisation and toxin production.13 14 In addition, the case mix on wards, and bays within wards, is important because of a potential herd immunity effect that can prevent epidemic outbreaks of C difficile associated diarrhoea.15 For example, if only a few patients on a ward have been exposed to antibiotics, most of those who have not been treated with antibiotics will retain their normal colonic bacterial flora and will not develop C difficile associated diarrhoea.
For an outbreak to occur, each case of C difficile associated diarrhoea must give rise to more than one new case. This depends not only on the infectivity of the agent, in this case C difficile, but also on the proportion of surrounding patients who are susceptible to infection. So as the proportion of patients on the ward who have received antibiotics and who are old rises, a threshold is reached where the product of the infectivity of C difficile and the proportion of susceptible patients is greater than one, and an outbreak of C difficile associated diarrhoea occurs.
Infectivity depends not just on the characteristics of the specific strain, but on factors that impact on cross infection (hygiene, bed spacing, and shared toilet facilities).
How should I treat it?
Treatment can be divided into:
Therapy for patients with C difficile associated diarrhoea
Measures to prevent other patients developing it.
Patients with C difficile associated diarrhoea
Therapy for patients with C difficile associated diarrhoea comprises:
Supportive measures (adequate fluid and electrolyte replacement)
Withdrawal of current antibiotic therapy if possible
Antibiotic treatment to eradicate C difficile.
Standard first line antibiotic therapy is metronidazole 400 mg administered orally three times a day. An alternative is oral vancomycin 125 mg four times a day, or at a higher dose for severe episodes. One study suggested this should be used as first line treatment for patients with albumin < 25 g/l or for patients who are in intensive care.16 Resistance to metronidazole and vancomycin is not reported, but changing antibiotic is often tried empirically after one week if symptoms have not resolved.
Placebo arms in clinical trials suggest that C difficile associated diarrhoea will resolve without antibiotic therapy in around 20% of patients. An earlier systematic review found that metronidazole and vancomycin produced comparable results.17
If a patient can't take oral medications, a nasogastric tube will be needed to administer antibiotics because an enteral route is required to treat the infection.
Evidence is inadequate to support the use of prebiotics (nutrients that encourage “normal” colonic bacterial flora) or probiotics (live microbial supplements, for example containing bifidobacteria, lactobacilli, and so on) for treating patients with established C difficile associated diarrhoea.18
Preventing C difficile associated diarrhoea
Prevention of C difficile associated diarrhoea consists of general infection control measures and those more specific for C difficile.
Diagnosis now tends to be made by detecting toxin, and so laboratories no longer culture C difficile. This hampers wider public health control of outbreaks and monitoring of antibiotic sensitivities because there may be a delay in detecting emerging epidemic strains.19 It also impairs diagnostic evaluation because cytotoxin is absent from the stools of a small proportion of endoscopically proved cases.20
Each unit should evaluate its case mix and, therefore, its risk of an outbreak of C difficile associated diarrhoea. For those at high risk, strict control of antibiotic prescribing may be useful.21
One study showed that several measures significantly reduced the incidence of C difficile associated diarrhoea22:
Revision and enforcement of the isolation policy to include wearing of gloves and gowns, and hand washing by healthcare staff; patients isolated in single or double bedded rooms or cohort bays; equipment such as thermometers and stethoscopes dedicated to each patient
Monthly educational programme for all healthcare workers
Phenolic disinfectant used for environmental cleaning
Triclosan (0.03%) soap used for hand washing
Sterilisation department centralised
Cart washer installed for cleaning wheelchairs and stretchers that have not come into direct contact with patients
Aggressive surveillance activity.
Recurrence can occur in up to one third of cases. It is sometimes difficult to determine whether this is due to failure of eradication of C difficile and subsequent recolonisation with normal colonic bacterial flora, or to a new infection of a susceptible patient. Strain typing may help, but its use is limited because the “epidemic” strain accounts for around 70% of infections, and colonisation with more than one strain is not uncommon.
You should treat patients whose stools contain a positive culture but do not contain toxins only if you strongly suspect that they have C difficile associated diarrhoea and if they are systemically unwell. This is because C difficile may be found incidentally in the stools of patients whose diarrhoea is due to other causes.
Use high dose vancomycin as first line treatment for patients on the intensive care unit or for patients with low albumin levels.
It is not unusual for an initial episode of C difficile associated diarrhoea to affect the patient's nutritional status and for the serum albumin to remain low at the time of recurrence. This is an indication to use high dose vancomycin.
A report to the Department of Health in 2004 on appropriate public health surveillance of C difficile reviewed current research.28 Key themes were infection control, antibiotic restriction policies (especially of third generation cephalosporins), and use of probiotics. It also noted that vaccines against C difficile toxins have been successful in animal models and that early safety trials in humans have been satisfactory. However, active immunisation may not be effective in people most at risk of C difficile, who characteristically fail to mount an immune response to C difficile infection.
Moreover, local colonic immunoglobulin A (IgA) production may be more important in protecting against C difficile associated diarrhoea than humoral IgG, and colonic IgA production is impaired in patients with C difficile associated diarrhoea.29 In view of this, passive immunisation, for example with pooled human immunoglobulin, may be a more promising strategy.30
Here is a small sample of the questions that you can find at the end of this module. To see all the questions and to get the answers, go to www.bmjlearning.com/ and search for “clostridium difficile”
A 79 year old man in a nursing home has mild diarrhoea and his stool samples are positive for C difficile toxin. He is otherwise well. Which of the following treatments would you give him?
He should be given probiotics
He should be treated with metronidazole 400 mg by mouth three times a day
He should be treated with vancomycin 250 mg four times a day
A 70 year old woman on ITU develops severe diarrhoea, abdominal pain, pyrexia, and delirium. Stool samples are culture-positive for C difficile, but toxin is not detected. What should be the first step in her management?
She should be treated with vancomycin 250 mg four times a day
She should be treated with metronidazole 400 mg three times a day
She should be treated expectantly, and antibiotics started only if her stools become toxin positive
A 75 year old man is found to have C difficile in his stools. But he is asymptomatic and there are no toxins in his stools. What should you advise?
Observe him off treatment
In what proportion of people who are colonised with C difficile does the bacterium produce toxins that cause diarrhoea?
You diagnose a 65 year old man with C difficile associated diarrhoea. You want him to start metronidazole, but he is unwilling and asks whether his symptoms could resolve on their own. What percentage of patients with C difficile associated diarrhoea will get better without antibiotic therapy?
Clostridium difficile associated diarrhoea is the most common cause of antibiotic associated diarrhoea in hospital
Oral metronidazole 400 mg three times a day and vancomycin 125 mg four times a day are the treatments of choice. You should usually continue treatment until symptoms stop
Probiotics may be useful for preventing recurrence
Evidence for the effectiveness of infection control measures is limited
In the meantime, clinicians will need to weigh the possible advantages of using vancomycin first line, especially at high dose, against the potential problems of inducing greater vancomycin resistant enterococci.1
Models and analysis
Looking to the future, one of the major difficulties is applying randomised controlled trial methodology to studying a condition that depends on herd immunity and for which there are many known confounding variables. These “herds” are spatially and temporally discrete but, in general, the number of cases of C difficile associated diarrhoea within a herd is too small for a randomised controlled trial. Mathematical modelling is therefore an alternative.
One promising approach is to apply stochastic models (models that deal with small numbers of chance events) to observational data.31 Such random stochastic events can be simulated by Monte Carlo methods. However, events are linked spatially and temporally and are therefore not individually independent.
Fortunately, mathematical methods such as Markov chain models can be used. These were developed to understand paths of cosmic particles as they collide with atoms on entering the earth's atmosphere. These models can help us conduct research on patients with C difficile associated diarrhoea.
This article is based on a module that is available on BMJ Learning (http://www.bmjlearning.com/). Access to the site is free but registration is required
Competing interests None declared.