The recommendations in this guideline are based on the evidence identified and the experience of the committee.
The committee agreed that although diagnostics and reporting were out-of-scope for this guideline, a recommendation should be included on where to find such information. They concluded, from experience, that people should see Public Health England's updated guidance on the diagnosis and reporting of Clostridioides difficile.
The committee discussed that, in practice, there has been a change in the definition of the severity of C. difficile from the 4 categories (mild, moderate, severe and life threatening) used by Public Health England to 3 categories (non-severe, severe and life threatening). However, the Public Health England categories still apply because this is current national guidance.
The committee discussed the findings of the economic model, which took into account severity by adjusting for older age, increased risk of recurrence, increased hospitalisation and a higher risk of fulminant colitis (see the economic analysis in the evidence review for full details; there was a lack of useful direct evidence for severity that could be used in the economic model). The economic model found that severity did not cause a substantial change in which antibiotic was the most cost effective. Therefore, the committee agreed that the main reason to assess severity was to identify the appropriate place of care, overall management, and any subsequent improvement or worsening. They also agreed that an assessment of whether the current infection was a first or further episode (relapse or recurrence) of C. difficile infection should be included. This was because recurrence was a driver in the economic model and determines antibiotic choice (see also the rationale on choice of antibiotic).
The committee recognised that C. difficile infection most commonly affects people who are taking or have recently taken antibiotics. They discussed that, even though antibiotics being taken may be associated with the C. difficile infection, the person may still need antibiotics for the original infection. They agreed that, in line with good antimicrobial stewardship, prescribers should review existing antibiotic treatment and:
stop it unless essential, or
if an antibiotic is still essential, consider changing to one with a lower risk of causing C. difficile infection.
The committee agreed that it is good prescribing practice to review the continuing need for existing proton pump inhibitor (PPI) treatment in people with suspected or confirmed C. difficile infection, in line with NICE's guideline on medicines optimisation. They were aware that, although some associations have been made between PPI use and the risk of C. difficile infection or recurrence, there is no definitive evidence of a causal or exacerbator effect. Also, no evidence from systematic reviews or randomised controlled trials was found to support stopping current PPI treatment. The committee agreed that suddenly stopping a PPI during an acute episode of infection may cause additional gastric symptoms. Additionally, some people will need ongoing gastroprotection for a clinical indication. However, they were aware that many people may be taking a PPI without a clear indication, so concluded that the use and need for a PPI should be reviewed.
The committee agreed that, when people present with suspected or confirmed C. difficile infection, it is good prescribing practice to review other medicines with gastrointestinal activity or adverse effects (such as laxatives) being taken. They also agreed that it is good practice to review other medicines being taken that may have detrimental effects if people are acutely ill and dehydrated. These include non-steroidal anti-inflammatory drugs, angiotensin-converting enzyme inhibitors, angiotensin‑2 receptor antagonists and diuretics.
The committee agreed that an oral antibiotic should be offered for suspected or confirmed C. difficile infection in adults, young people and children.
For adults presenting in the community, the committee agreed that GPs may be unfamiliar with diagnosing and treating C. difficile infection, so may want to seek prompt specialist advice from a microbiologist or infectious diseases specialist before starting treatment. The committee noted that, for people in hospital, once diagnosed, they would be under the care of a multidisciplinary team, which would ensure appropriate review and care.
The committee discussed the lack of evidence on treating C. difficile infection in children and young people. They were aware that, in practice, very few children have C. difficile infection. The committee agreed that a positive test for C. difficile in children 2 years and under is often because of high carriage rates of the bacteria rather than because of actual infection. They considered that this may lead to overprescribing of antibiotics. The committee concluded that treatment for C. difficile infection in children and young people should only be started by a microbiologist, paediatric infectious diseases specialist or paediatric gastroenterologist, or after advice from such a specialist.
The committee discussed the most appropriate route of administration of antibiotics for C. difficile infection. They agreed that the enteral route is best because sufficient concentrations within the intestinal lumen need to be reached. The committee concluded that it is preferable to take antibiotics orally or, if this is not possible, enterally in some other way (such as a nasogastric or enteral feeding tube, or rectally). They advised seeking specialist advice on administration from a specialist gastroenterologist or pharmacist if the oral route is not available.
The committee agreed that, in line with the general management of gastroenteritis (see the NICE clinical knowledge summary on adult gastroenteritis and NICE's guideline on diarrhoea and vomiting caused by gastroenteritis in under 5s), prescribers and other care staff should monitor and manage fluid loss and gastroenteritis symptoms. Antimotility drugs such as loperamide should be avoided because they slow down the action of the gut. This can lead to C. difficile toxins being retained for longer, which may make a person more unwell.
Bezlotoxumab was not recommended as adjunctive therapy to antibiotics to prevent recurrent C. difficile infection. The committee discussed the clinical evidence, which showed that bezlotoxumab was more effective than placebo at preventing recurrence. However, they also reviewed the economic analysis and agreed that adding bezlotoxumab to either vancomycin or fidaxomicin was not a cost-effective option (there is a 0% probability of it being cost effective at £30,000 per quality-adjusted life year [QALY] gained). The committee agreed that this finding was robust, even in people with a higher risk of recurrence, and were confident in making a recommendation for bezlotoxumab not to be used.
The committee noted that faecal microbiota transplantation (FMT; a procedure done in a small number of specialist centres) was not effective as a first-line treatment for C. difficile infection compared with vancomycin. They were aware that long-term safety data on, and regulations about the use of, FMT are minimal compared with medicines. They were aware of variation in mortality rates associated with FMT use, and that there is almost no evidence for its use in children. NICE's interventional procedures guidance on FMT for recurrent C. difficile infection states that 'current evidence on the efficacy and safety of FMT for recurrent Clostridium difficile infection is adequate to support the use of this procedure provided that normal arrangements are in place for clinical governance, consent and audit'. In the economic model, FMT was placed as a third-line treatment (for people with continuing symptoms after first- and second-line antibiotics) that may help prevent serious complications. The committee agreed that FMT may be useful in adults who have had 2 or more previous episodes of C. difficile infection in addition to the current episode to prevent recurrence of C. difficile infection. They were aware of ongoing developments around the screening of faecal microbiota donors to identify multidrug-resistant organisms.
For more details, see the summary of the evidence on treating initial or first recurrent C. difficile infection.
maintaining fluid intake to avoid dehydration (and on the symptoms or signs of dehydration that people should be aware of)
the need to help reduce the spread of C. difficile infection, which is contagious (that is, people should follow the advice in the NICE clinical knowledge summary on adult gastroenteritis and in NICE's guideline on diarrhoea and vomiting caused by gastroenteritis in under 5s)
when to seek medical help.
The committee were aware that C. difficile infection should be managed as a diagnosis in its own right. They agreed that the management and progress of suspected or confirmed C. difficile infection should be monitored during treatment. This could include assessing the severity of the infection and symptoms, and the need for hydration. The committee concluded that, from their experience, it would be good practice to reassess people if symptoms or signs of infection do not improve as expected or worsen rapidly or significantly at any time. They agreed that daily review is usual in hospital and that, in the community, people should be given appropriate safety netting advice to ensure that they return for reassessment if needed.
The committee also agreed that clinicians should consider stopping antibiotics for C. difficile infection if they have been started for clinically suspected C. difficile infection before stool sample test results are available and subsequent results do not confirm infection.
The committee agreed that people with suspected or confirmed C. difficile infection in the community should be referred to hospital if they are severely unwell, or their symptoms or signs worsen rapidly or significantly at any time. For life-threatening infection, an urgent referral is needed. The committee recognised that there are some individual factors (such as age, frailty and comorbidities) for which it may also be appropriate to consider referral to hospital. This is because they are associated with a higher risk of complications or recurrence.
The committee agreed that people who develop C. difficile infection while in hospital are unlikely to be having care from a microbiologist or infectious diseases specialist at diagnosis. However, once diagnosed they should be under the care of a multidisciplinary team to ensure appropriate review and care. The team could include, as needed, a microbiologist, infectious diseases specialist, gastroenterologist, surgeon, pharmacist and dietitian. This would depend, for example, on the severity of illness and need for surgery.
The committee discussed the evidence for the effectiveness and cost effectiveness of the different antibiotic options for treating C. difficile infection. They were aware that antibiotic resistance is not a major concern when treating C. difficile infection.
Oral vancomycin was recommended by the committee as the first-line antibiotic for a first episode of C. difficile infection of any severity. Fidaxomicin was recommended as the second-line antibiotic for a first episode of C. difficile infection of any severity when vancomycin is ineffective (treatment failure). The committee noted that, although fidaxomicin was more effective than vancomycin for sustained symptomatic cure in the network meta-analysis, the cost of fidaxomicin is substantially higher. In the base-case analysis, there was only a 2% probability of first-line fidaxomicin being cost effective compared with first-line vancomycin (at £30,000 per QALY gained). They also discussed that vancomycin treatment failure should not be judged too early. Diarrhoea can take 1 to 2 weeks to resolve, and it is not usually possible to determine whether antibiotic treatment for C. difficile is ineffective until day 7.
The committee agreed that, when teicoplanin and second-line metronidazole were excluded from the economic model, the remaining results clearly showed that vancomycin was the most cost-effective first-line antibiotic across a range of scenarios. This was the case when results from people at both higher and lower risks of recurrence were included (in particular, it was more cost effective as a first-line option than either metronidazole or fidaxomicin). They also agreed that fidaxomicin was the appropriate second-line option.
The committee noted that, from experience, some hospital trusts use fidaxomicin for first-line treatment of C. difficile infection in people who are older or frailer as a strategy to reduce recurrence and readmission. The aim is to offset the cost of using fidaxomicin by reducing future costs. The committee were made aware of a real-world evaluation of fidaxomicin in which its use first line had a greater effect on reducing mortality than its use second line after vancomycin. However, they heard that the economic model considered a range of benefits and harms (including deaths), as well as the costs of each strategy. Vancomycin (not fidaxomicin) was still the most cost-effective first-line option, even in people at higher risk of recurrence. The committee concluded that a recommendation to use fidaxomicin first line would incur unreasonably large opportunity costs that are not appropriate in the wider context of overall healthcare resource allocation. There are possible rare exceptions when vancomycin may not be acceptable, such as for an infection that is vancomycin resistant.
The committee noted that, when taken orally, vancomycin is not well absorbed from the gut into the circulation (although absorption may increase if the gut is damaged). So, the likelihood of side effects (such as ototoxicity) is lower with oral than with intravenous administration, although there is still a need to monitor for this in some people (see medicines safety). They also discussed the development of drug-resistant bacteria, in particular, vancomycin-resistant enterococci. However, they agreed with expert testimony that this is not a major concern in clinical practice when vancomycin is used orally for C. difficile infection.
The committee discussed that vancomycin capsules would usually be the preferred formulation for taking vancomycin orally. They were aware that vancomycin powder for solution is also licensed to be taken orally for C. difficile infection, and that it is used in some settings (particularly if people cannot take solid oral medicines). However, they discussed that locally agreed protocols should be in place to reduce the risk of medication errors around reconstitution and administration, and to take account of the practicalities of administration, particularly in community settings. This is discussed further in Specialist Pharmacy Service guidance on choosing between oral vancomycin options.
Fidaxomicin was recommended by the committee for a further episode of C. difficile infection of any severity occurring within 12 weeks of symptom resolution. They defined this as a relapse. For a further episode of C. difficile infection occurring more than 12 weeks after symptom resolution (defined as recurrence), either vancomycin or fidaxomicin was recommended, with choice being an individualised patient decision.
The committee noted there was no clinical evidence comparing vancomycin with fidaxomicin in a population having a further episode of C. difficile infection after initial cure. Their decisions were therefore heavily influenced by the threshold analyses around risks of future recurrence. This was because they agreed that 1 key difference with a further episode of infection is the higher risk of subsequent additional recurrences. The committee noted that the risk of future recurrence needed to be around 30% to 40% for fidaxomicin to be cost effective as a first-line option compared with vancomycin (at £30,000 per QALY gained). While they did not believe that this would be the case for all people with a recurrent infection, they did agree that there would be people with a risk of recurrence that high. They therefore agreed that it was appropriate for both vancomycin and fidaxomicin to be first-line options for further episodes. They concluded that the choice would come down to an individualised patient decision based around severity, the risk of additional recurrences (which increases after each recurrent episode) and the time between recurrences. The committee favoured fidaxomicin for more severe, more recent or multiple recurrent episodes. They thought that vancomycin would be suitable for less severe or first recurrent episodes, or if there had been a long time between episodes.
The committee were aware that there is poor agreement on the definition of relapse or recurrence in C. difficile infection, both nationally and internationally. They discussed different time periods and agreed, based on expert opinion, that 30 days from resolution of symptoms was too short a time period to define recurrence. They thought that further symptoms within this time period after initial symptom resolution were more likely to represent relapse with the same strain of C. difficile infection. The committee heard that, in practice, further symptoms within 12 to 24 months may be considered a recurrence, likely with a different strain of C. difficile infection. However, they were also aware of evidence that suggested recurrence generally relates to a further episode within 20 weeks. Defining relapse or recurrence is outside of the remit of the committee, and evidence on this issue was not searched for. So, the committee agreed that it could not be certain about the time period but thought that 12 weeks was a reasonable cut-off point between relapse and recurrence.
The committee agreed that specialist advice should be sought about the choice of antibiotics for C. difficile infection that has not responded to either first- or second-line antibiotics, or for a life-threatening infection. However, they recognised that, in practice, specialists will often initially recommend high-dose oral vancomycin with or without intravenous metronidazole for this. If ileus is present, specialists may use vancomycin rectally.
Teicoplanin was not recommended by the committee for treating C. difficile infection. It was ranked first in the network meta-analysis results. However, the committee were concerned about the extensive limitations of the 2 small studies of teicoplanin included in the network meta-analysis, both of which were at considerable risk of bias. The committee noted that the point estimate of effect was important. However, the 95% confidence intervals were wide, revealing much uncertainty in the estimate. This meant that there was little difference from, and overlap with, the estimate of effect for vancomycin. The committee were also aware of the limited clinical experience with using teicoplanin in the UK for C. difficile infection. They concluded that further research was needed on teicoplanin for treating C. difficile infection and made a recommendation for research.
The committee had an initial discussion about the findings from the economic model. They noted that, if the results from the studies of teicoplanin were considered robust, it would come out clearly as the most cost-effective first-line treatment. However, they were not convinced by either the sample size or quality of the studies on teicoplanin. They agreed there was not enough clinical evidence to recommend it, so focused on the economic model results excluding teicoplanin.
Metronidazole was not recommended by the committee for treating C. difficile infection. The committee agreed that not using metronidazole first line for mild and moderate C. difficile infection represented a change in practice for some clinicians. However, they were confident in the evidence that metronidazole was neither clinically nor cost effective compared with vancomycin. In the network meta-analysis results, metronidazole was ranked lowest out of all the antibiotics available in the UK (below teicoplanin, fidaxomicin, vancomycin, rifaximin and fusidic acid). In the economic modelling, when the costs of rehospitalisation were included in the analysis, metronidazole was a less cost-effective first-line treatment than vancomycin, which was dominant in most scenarios (meaning using vancomycin was both less costly and more effective than using metronidazole). From the evidence, metronidazole had lower initial cure rates and higher recurrence rates than vancomycin. The committee heard that metronidazole is comparatively inexpensive compared with other antibiotic treatments. However, they discussed that, from experience, many hospital trusts have already moved away from using metronidazole, prompted by lower efficacy compared with other antibiotics and potential side effects. The committee also heard expert testimony that cure or improvement may take longer with metronidazole compared with other antibiotic treatments. A longer period before treatment becomes effective is concerning. This is because it may lead to increased transmission of the infection, particularly in hospital or residential care settings. Neither of these issues were addressed in the economic model.
When considering the economic model, the committee agreed that it was appropriate to exclude strategies in which metronidazole was used as a second-line intervention. They noted that a limitation of the analysis was that interventions were assumed to be equally effective as second-line options compared with first-line options. This was because there were no data to test this assumption. They agreed that, when C. difficile is not clinically cured using first-line vancomycin or fidaxomicin it is likely to represent infection that is harder to treat. So, it would also be less likely to respond to metronidazole, meaning it would not be effective as a second-line agent.
The committee recognised that intravenous metronidazole may be a treatment option in the rare event that C. difficile infection fails to respond to either vancomycin or fidaxomicin, or in people with a life-threatening infection. The committee noted that, from experience, intravenous metronidazole (as an adjunct to vancomycin by the enteral route) is used in practice for some people in these circumstances.
The committee noted the evidence showing no statistically significant difference in clinical effectiveness with low-dose (125 mg four times a day) compared with high-dose (500 mg four times a day) vancomycin. The committee concluded that the standard licensed dose of oral vancomycin 125 mg four times a day for 10 days was sufficient to treat a first episode of mild, moderate or severe C. difficile infection, or a further episode of infection more than 12 weeks after symptom resolution (recurrence). The committee were also aware that specialists may use higher licensed doses of oral vancomycin (up to 500 mg four times a day) for C. difficile infection not responding to first- or second-line antibiotics or for life-threatening infection.
Oral vancomycin can be taken as capsules or the powder for solution can be reconstituted and taken orally as a drink or by nasogastric tube (a licensed use). The committee discussed that capsules would be the preferred formulation, particularly in community settings, for ease of use and to avoid any safety concerns around reconstitution and administration.
A tapered or pulsed regimen of vancomycin was not recommended because, in the evidence review, its use was limited to studies in which there was co-administration of FMT. The committee were aware that there are ongoing trials which might provide evidence for wider use of pulsed or tapered vancomycin.
The committee noted the evidence suggesting that fidaxomicin 400 mg daily was more clinically effective than 100 mg or 200 mg daily. They concluded that the standard licensed dose of oral fidaxomicin 200 mg twice a day for 10 days was sufficient to treat C. difficile infection.
The committee considered the comparison of the standard and extended-pulsed regimens of fidaxomicin in the economic model. The unlicensed extended-pulsed regimen of fidaxomicin is 200 mg twice a day on days 1 to 5, then 200 mg once a day on alternate days from days 7 to 25. The committee noted that the point estimates were in favour of extended-pulsed fidaxomicin. However, there was considerable uncertainty in this conclusion (with a 36% chance of standard fidaxomicin being more cost effective than extended-pulsed fidaxomicin at £30,000 per QALY gained). Also, the absolute magnitude of the differences was small. The committee agreed that there was insufficient evidence of benefits from the extended-pulsed regimen to justify recommending an unlicensed treatment regimen over a licensed one.
The committee agreed that treatment for C. difficile infection in children and young people should only be started by, or after advice from, a specialist. And that antibiotic choice can be based on recommendations for adults, taking into account the varying licensed indications for children and the availability of suitable products.
Vancomycin capsules are licensed to treat C. difficile infection only in people aged 12 years and over (see the vancomycin capsules summary of product characteristics). Vancomycin powder for solution taken orally is licensed to treat C. difficile infection in all age groups (see the vancomycin powder for solution summary of product characteristics).
Fidaxomicin tablets are licensed to treat C. difficile infection in children with a body weight of at least 12.5 kg (see the fidaxomicin summary of product characteristics). Fidaxomicin granules for oral suspension are licensed to treat C. difficile infection from birth (and are likely to become available in the UK). However, there is a caution for use in babies less than 6 months and in babies with body weight less than 4 kg (see the Medicines and Healthcare products Regulatory Agency information on fidaxomicin granules).
For more detail see the summary of the evidence on antibiotic dose.
The committee agreed that, although preventing C. difficile infection through good antimicrobial stewardship, infection control and environmental hygiene were out-of-scope for this guideline, a recommendation should be included on where to find such information. They concluded, from experience, that people should see Public Health England's guidance on C. difficile infection: how to deal with the problem, and NICE's guidance on healthcare-associated infections and antimicrobial stewardship.
The committee also discussed the importance of ensuring that a diagnosis of C. difficile infection is recorded in a person's medical records. This is particularly important when transferring from one care setting to another, so that it can be taken into account before prescribing any future antibiotics, to help minimise the risk of recurrent episodes.
The committee noted the lack of evidence of clinical or cost effectiveness to prevent C. difficile infection with antibiotics. They recognised that there was some evidence for rifaximin preventing further recurrences from a single study in people who already had recurrent infection. However, the intensive way in which antibiotics were used in the study has raised concerns about the possible emergence of rifamycin resistance, which has been reported in C. difficile infection cases, and prolonged flora disturbance.
The committee also recognised the limited evidence of benefit for:
fidaxomicin in preventing C. difficile infection in people having a haematopoietic stem cell transplant who had fluoroquinolone prophylaxis
vancomycin in preventing C. difficile infection in people who are in hospital.
The economic model only included treatment options, including adjunctive treatment with bezlotoxumab (which is used to prevent recurrent infection) and FMT to determine sequencing of treatments. It did not include comparisons for preventing a first episode of C. difficile infection with antibiotics, prebiotics or probiotics. The committee concluded that, because of the lack of evidence and concerns about antimicrobial resistance, antibiotics should not be offered for preventing C. difficile infection.
The committee noted the lack of convincing evidence of effect for prebiotics (oligofructose), which showed little difference in preventing C. difficile-associated outcomes in the included studies. They concluded that prebiotics conferred no benefit and that people taking antibiotics should not be advised to take prebiotics to prevent C. difficile infection.
The committee agreed that there is some evidence of a small effect with probiotics in preventing C. difficile infection. However, there were many limitations in the evidence, including:
a high number needed to treat
aggregation of the results of different types of probiotics in meta-analyses
the lack of effectiveness when using confirmed cases only (in everyone, but particularly in children).
The committee also noted concerns from expert testimony about the high prevalence of C. difficile infection in the placebo arms of some studies, which does not reflect clinical practice in the UK. The single study done in a UK setting found no evidence of effect for probiotics in people aged over 65 years. They further noted that NHS England's guidance on conditions for which over the counter items should not routinely be prescribed in primary care states that probiotics should not routinely be prescribed.
The committee concluded that, because of concerns about the evidence base (including cost effectiveness), people taking antibiotics should not be advised to take probiotics to prevent C. difficile infection.