Ed J Kuijper
Department of Medical Microbiology, Leiden University Medical Centre, the Netherlands
Martijn P. Bauer
Jaap T van Dissel
Department of Infectious Diseases, Leiden University Medical Centre, the Netherlands
On behalf of the European Society of Clinical Microbiology and Infectious Diseases (ESCMID)
Clostridium difficile infection (CDI) is a potentially fatal illness with an increasing incidence worldwide. Despite extensive ongoing research into CDI treatment, management of CDI still poses important problems, such as a high propensity to relapse and refractoriness to treatment, especially when there is an ileus and oral drugs cannot be administered.
This summary of the European guidelines discusses criteria for disease severity and provides updated recommendations for CDI treatment, indicating the level of evidence and strength of recommendation.
CDI may arise when a patient’s bowel is colonised by C. difficile after ingestion of spores, the spores subsequently germinate and the vegetative bacteria start producing toxins. The vegetative state of the bacterium is resistant to a varying but broad range of antibiotics and the spores are highly resistant to antibiotics and can withstand many forms of chemical attack, including most high-level disinfectants.
The most important problem in treating CDI is the high recurrence rate. The antibiotics needed to kill the vegetative bacteria do not kill the spores. Individuals who suffer a recurrence may enter a repetitive cycle of recurrences, leading to exhaustion and protein-losing enteropathy.
A second problem in treating CDI is the fact that, in severe forms of CDI, antibiotics may fail, resulting in progressive colitis with high morbidity and mortality. Several factors may play a role in this, such as a time-lag for antibiotics to reach adequate intracolonic levels1 and possibly the fact that a systemic inflammatory response due to severely damaged colonic mucosa may persist some time after removal of the etiological agent.
The objective of this guideline was to update the existing guidance document using recently published information concerning treatment of CDI.2
Criteria for the diagnosis of CDI
Pseudomembranous colitis, which is an endoscopic diagnosis, is caused by C. difficile in the vast majority of cases and therefore may suffice for the diagnosis of CDI in the absence of an obvious other cause. In the rest of the cases, a combination of symptoms and signs plus microbiological evidence of toxin-producing C. difficile in stool and absence of another cause is necessary.
Compatible clinical pictures are diarrhoea, ileus and toxic megacolon. Diarrhoea is defined as loose stools, i.e. taking the shape of the receptacle or corresponding to Bristol stool chart types 5 to 7,3 plus a stool frequency perceived as too high by the patient. Faecal incontinence may be a part of the disease.
Ileus in the context of CDI is defined as signs of severely disturbed bowel passage such as vomiting and absence of stool, combined with radiological signs of bowel distension.
Toxic megacolon is defined as radiological signs of distension of the colon combined with signs of a severe systemic inflammatory response.
We refer to the European Society of Clinical Microbiology and Infectious Diseases (ESCMID) guideline on diagnosis of CDI, which has recently been published.4 Since the published guidance documents, various molecular assays have become commercially available with sufficient test performances, allowing them to apply in routine diagnostics.5-7 The above-mentioned criteria are largely in line with the recommendations by the American Ad Hoc C. difficile surveillance working group8 and the European Study Group for C. difficile.9
Markers of severe colitis
Markers that could reasonably be assumed to correlate positively with severity of colitis are mentioned below, although we must stress that the prognostic value of these markers is uncertain. Obviously, markers should not be attributable to a concomitant disease, if they are to be regarded as a marker of severe CDI. Ideally, markers should be obtainable at the earliest time in the disease course to be a predictor of outcome.
- Fever (core body temperature > 38.5ºC)
- Rigours (uncontrollable shaking and a feeling of cold followed by a rise in body temperature)
- Haemodynamic instability, including signs of distributive (vasodilatory; septic) shock
- Signs of peritonitis, including decreased bowel sounds, abdominal tenderness, rebound tenderness and guarding
- Signs of ileus, including vomiting and absent passage of stool.
Admixture of blood with stools is rare in CDI and the correlation with severity of disease is uncertain.
- Marked leukocytosis (leukocyte count >15 ∙ 109/l)
- Marked left shift (band neutrophils >20% of leukocytes)
- Rise in serum creatinine (>50% above the baseline)
- Elevated serum lactate
- Hypo-albuminemiae (≤25g/l)
Colonoscopy or sigmoidoscopy:
- Pseudomembranous colitis
There is insufficient knowledge on the correlation of endoscopic findings compatible with CDI, such as oedema, erythema, friability and ulceration, and the severity of disease.
- Distension of large intestine
- Colonic wall thickening including low-attenuation mural thickening
- Pericolonic fat stranding
- Ascites not explained by other causes.
The correlation of haustral or mucosal thickening, including thumbprinting, pseudopolyps and plaques, with severity of disease is unclear.
Prognostic markers other than disease severity
- High age (≥65)
- Serious comorbidity and ICU admission
There has only been one placebo-controlled trial investigating the effectiveness of antibiotics for CDI and it had very few participants. Several antibiotics have been compared to each other. Oral administration of the glycopeptides vancomycin and teicoplanin appears most effective in inducing both clinical cure and microbiological cure, especially in severe CDI.
The difficulty is how to define severe CDI. In one prospective, randomised and blinded study,10 which evaluated the efficacy of vancomycin versus metronidazole according to disease severity, the diagnosis of severe CDI was based on age, body temperature, albumin level and leukocyte count. Vancomycin proved to be superior over metronidazole in cases of severe CDI.
Two trials investigating the efficacy of the toxin-binding polymer, tolevamer,11,12 also showed superiority of oral vancomycin over metronidazole in severe cases. A recent Cochrane systematic review13 has examined the available literature on antibiotic treatment options of CDI and concluded that teicoplanin is the most effective antibiotic treatment for moderate to severe CDI and vancomycin has no superiority over metronidazole. However, this review did not include the above-mentioned recent studies. It seems likely that the effectiveness of teicoplanin and vancomycin is in the same range.
Oral metronidazole is also very effective in inducing a response and has the advantage of low cost and the fact that it may contribute less to the emergence of vancomycin-resistant enterococci.
If metronidazole is indeed less effective than glycopeptides, this may be explained by the fact that only low levels metronidazole reach in the colon, since it is absorbed in the small intestine and then excreted again in the bile and in the inflamed colon, whereas glycopeptides are not absorbed.
Different doses of oral vancomycin have been used, but only one small, randomised trial14 has compared high- versus low-dose vancomycin and found no statistically significant difference. Since low doses of oral vancomycin result in high concentrations in stool, there is no need to treat with high doses, except in an attempt to reach sufficient concentrations in the colon when administering vancomycin by nasogastric tube in a patient with ileus.
Given the poor faecal concentrations of metronidazole achieved following a 500mg eight-hourly dose, lower doses (e.g. 250mg 6–8 hourly) should be less effective. Even a modest increase in the minimum inhibitory concentration (MIC) of metronidazole for C. difficile might result in insufficient faecal antibiotic concentrations to inhibit (vegetative) bacteria.
Metronidazole resistance is to be regarded as exceedingly rare. However, the emergence of reduced susceptibility to metronidazole has recently been reported in UK C. difficile strains.1,15 No reduced susceptibility to vancomycin was observed. The exact mechanism of reduced susceptibility to metronidazole remains to be determined. Notably, there is also evidence that inactivation of metronidazole occurs in the presence of gut contents, possibly due to metabolism by enterococci.16
Oral bacitracin and fusidic acid seem to be less effective than vancomycin and metronidazole, respectively, although this has not convincingly been demonstrated. Currently, there is insufficient evidence to advocate the use of the rifamycin derivative rifaximin, to which resistance has been noted, and the antiprotozoal/anthelminthic nitazoxanide, which has been shown to be statistically similar to metronidazole in a small prospective randomised trial,17 but whose non-inferiority to vancomycin could not be shown in another trial due to lack of power.18
Fidaxomicin (OPT-80), a macrocyclic antiobiotic and inhibitor of RNA polymerase of gram-positive bacteria, has recently been tested in a prospective randomised, controlled trial in comparison with vancomycine in the treatment of recurrent CDI. In total, 128 patients were included and a 45% reduction of recurrent infections was observed in the fidaxomicin group as compared to vancomycin-treated group. [Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) 2010, 12–15 September, Boston, USA, abstract L1-1305]. Clearly, these preliminary findings support further investigation of the clinical merits of fidaxomicin in CDI and a second phase III study is underway.
Duration of antibiotic therapy
The duration of antibiotics has been ten days in most studies. Occasionally, shorter duration (seven days) has been studied. We feel that there is insufficient evidence for a shorter duration of therapy with any antibiotic to consider shorter regimens a treatment option.
There is no definitive evidence that taper or pulse regimens with vancomycin are effective in reducing the incidence of relapses. This strategy is mainly based on favourable experience and the theoretical rationale that spores can still germinate long after the clinical symptoms have resolved.
McFarland et al19 retrospectively compared a standard course of antibiotics, vancomycin taper strategies (gradually decreasing the daily dose of vancomycin with 125mg to 750mg per day from varying starting doses) and vancomycin pulse strategies (125mg to 500mg of vancomycin every two to three days during a period of usually three weeks).
They found the recurrence rate to be lowest in pulse regimens (14%), followed by taper regimens (31%) and the standard regimen of vancomycin (54%; average for all dose groups). No other studies investigating taper or pulse regimens have been published. Further studies are needed.
Probiotics may be of value when added to antibiotics, but the studies that have investigated this suffer from major drawbacks such as small numbers, non-randomised allocation of antibiotics to which the probiotics were added, and lack of homogeneity between study groups. This is also the conclusion reached by a recent Cochrane systematic review.20 Therefore, there is insufficient evidence to recommend the addition of probiotics to antibiotics.
In addition, several reports of invasive disease have been reported resulting from the use of probiotics such as Saccharomyces boulardii in debilitated or immunocompromised patients.21,22 Moreover, probiotics were associated with increased mortality, partly due to nonocclusive mesenteric ischemia, in a randomised controlled trial in acute pancreatitis.23
Treatment when oral administration is not possible
The only parenteral antibiotic therapy for CDI, supported by case series, is metronidazole.24 Furthermore, several case reports regarding the use of intravenous immunoglobulin (IVIG) have been published, but the data do not provide sufficient evidence to support its use. Fifteen small, mostly retrospective and non-randomised, reports documented IVIG’s success in the treatment of protracted, recurrent or severe CDI.25
A recently performed phase 2 study of intraveneously administered human monoclonal antibodies against C. difficile toxins A and B did not reveal a significant effect on the course of the disease, with the exception of a decreased relapse rate.26 Thus, it is unknown how to best treat patients with ileus due to CDI.
There are some anecdotal reports on delivery of vancomycin to the gut by other means than orally, mainly through intracolonic delivery. Questions regarding the efficacy, optimal dosing and duration of treatment with intracolonic vancomycin are unanswered. The introduction of faecal collector drainage systems has facilitated the use of glycopeptide retention enemas in intensive care units (ICUs), but they are very expensive.
Tigecycline has been applied as ‘salvage’ therapy in two recent case series of patients with severe CDI complicated by ileus; these promising but preliminary findings require confirmation in prospective clinical trials.27,28 Faecal transplantation has been performed via instillation with a colonoscope, tube or enema, but there is insufficient, controlled evidence to recommend this approach as of yet.29-31
There are no prospective studies assessing which CDI patients benefit from surgical intervention. One study found that colectomy was most successful in a relatively early stage of the disease, i.e. before lactate exceeds 5.0mmol/l.32
Recommendations for medical treatment of
In the case of mild CDI (stool frequency <4 times daily; no signs of severe colitis), clearly induced by the use of antibiotics, it is acceptable to stop the inducing antibiotic and observe the clinical response, but patients must be followed very closely for any signs of clinical deterioration and placed on therapy immediately if this occurs.
Theoretic rationale, anecdotal evidence and one case-control study suggest that antiperistaltic and opiate agents should be avoided, especially in the acute setting.33 There is no evidence that switching to ‘low-risk’ antibiotics is effective, when the antibiotic treatment that triggered the episode of CDI cannot be stopped or its spectrum be narrowed. It seems rational, however, to always strive to use antibiotics covering a spectrum no broader than necessary. When the inciting antibiotic cannot be stopped, antibiotic treatment for CDI should be initiated. Furthermore, there is no proof that stopping gastric acid suppressants is effective, either.
In all other cases other than mild CDI, medical treatment for CDI should be started. Antibiotics may be started while awaiting diagnostics when there is sufficient clinical suspicion. We recommend treatment of an initial episode of CDI with the following antibiotics, according to disease severity (implementation category in brackets), when oral therapy is possible:
- Non-severe: metronidazole 500mg tid orally for ten days (A-I)
- Severe: vancomycin 125mg qid* orally for ten days (A-I).
* Oral vancomycin may be replaced by teicoplanin 100mg bid, if available.
CDI is judged to be severe when one or more of the markers of severe colitis mentioned under ‘definitions’ is present. It is unclear whether moderate disease in a patient with other unfavourable prognostic factors, such as high age and comorbidity, should be regarded as severe. This is left to the judgment of the treating physician. There is no evidence that various genotypes of C. difficile should be treated differently if disease severity does not differ. When oral therapy is impossible, we recommend the following antibiotics, according to disease severity (implementation category between brackets):
Non-severe: metronidazole 500mg tid intravenously for ten days (A-III)
- Metronidazole 500mg tid intravenously for 10 days+ (A-III)
- Intracolonic vancomycin 500mg in 100ml of normal saline every 4hrs–12hrs (C-III)
- And/or vancomycin 500mg qid by nasogastric tube (C-III).
Surgical treatment of CDI
Colectomy should be performed to treat CDI in any of the following situations:
- Perforation of the colon
- Systemic inflammation and deteriorating clinical condition not responding to antibiotic therapy; this includes the clinical diagnoses of toxic megacolon and severe ileus.
Since mortality from colectomy in patients with advanced disease is high, it is recommended to perform surgery in a less severe stage. A recent summary of 35 patients who underwent colectomy for fulminant C. difficile colitis revealed a mortality of 46%. None of the patients undergoing partial colectomy survived and the pre-operative presence of multisystem organ failure was an independent predictor of mortality.34
Unfortunately, no definite guidance can be given on the timing of colectomy. Serum lactate may, inter alia, serve as a marker for severity, and experts stress that one should operate before a threshold of 5.0mmol/l is reached.32 White cell count (WCC) in the first three days of disease is also a strong predictor of mortality and should be routinely monitored. At present, a WCC of 20 x 10(9)/L or greater may be the best cut-off value to objectively identify cases at higher risk of death.35
Recommendations for medical treatment of recurrent CDI
Observational data36 suggest that the incidence of a second recurrence after treatment of a first recurrence with oral metronidazole or vancomycin is similar. Therefore, we recommend treating a first recurrence of CDI as a first episode, unless disease has progressed from non-severe to severe. We recommend treatment of recurrent CDI with the following antibiotics (implementation category in brackets):
First recurrence: See recommendations for medical treatment of initial CDI.
Second recurrence and subsequent recurrences:
* If oral therapy is possible:
- Vancomycin 125mg qid* orally for at least 10 days (B-II)
- Consider a taper/ pulse strategy (B-II)
* Oral vancomycin may be replaced by teicoplanin 100mg bid, if available.
* If oral therapy is impossible:
- Metronidazole 500mg tid intravenously for 10-14 days+ (A-III)
- Retention enema of vancomycin 500mg in 100 ml of normal saline every 4hrs–12hrs (C-III)
- And/or vancomycin 500mg qid by nasogastric tube (C-III).
A recently reported randomised, double-blind, placebo-controlled phase 2 study of two neutralising, human monoclonal antibodies against C. difficile toxins A and B revealed a lower rate of recurrence of C. difficile infection among patients treated with monoclonal antibodies (7% versus 25%; P<0.001).26 However, this effect was only observed in outpatients with mild disease and not in inpatients, and the groups enrolled in the study are small. Phase 3 studies are currently designed and will include patients with more severe and life threatening disease.
The clinical experience with faeces transplantation to reduce relapse rate of CDI has recently been summarised by van Nood et al.37 A recovery of normal intestinal flora is postulated to be the mechanism for cure. In a recent study, the donor bacterial population was shown to persist at least for 24 weeks after infusion.31 Of 159 reported patients with recurrent CDI, 91% was cured after one or more infusions, but, clearly, such a result will be biased by positive reporting. After a promising pilot study,38 a randomised trial comparing donor faeces infusion to conventional antibiotic treatment with vancomycin has been initiated in the Netherlands (The FECAL trial).
Strength of recommendation and quality of evidence according to the Canadian Task Force
on the Periodic Health Examination
- Good evidence to support a recommendation
- Moderate evidence to support a recommendation
- Poor evidence to support a recommendation.
Quality of evidence:
- Evidence from ≥1 properly randomised, controlled trial
- Evidence from ≥1 well-designed clinical trial, without randomisation; from cohort or case-controlled analytic studies (preferably from ≥ centre); from multiple time-series; or from dramatic results from uncontrolled experiments
- Evidence from opinions of respected authorities, based on clinical experience, descriptive studies, or reports of expert committees.
- Kuijper EJ & Wilcox MW. Clin Infect Dis 2008;47:63-65.
- Bauer MP et al. Clin Microbiol Infect. 2009;15:1067-7.
- O’Donnell LJD et al. BMJ 1990;300:439-440.
- Crobach MJ et al. Clin Microbiol Infect. 2009;15:1053-66.
- Doing KM. Diagn Microbiol Infect Dis. 2010;66:129-3.
- Kvach EJ et al. J Clin Microbiol. 2010;48:109-14.
- Huang H et al. J Clin Microbiol. 2009;47:3729-31.
- McDonald LC et al. Infect Control Hosp Epidemiol 2007;28:140-145.
- Kuijper EJ et al. Clin Microbiol Infect 2006;12 Suppl 6:2-18.
- Zar FA et al. Clin Infect Dis 2007;45:302-307.
- Louie TJ et al. In: Program and abstracts of the 47th Interscience Conference on Antimicrobial Agents and Chemotherapy, 17-20 September 2007, Chicago, USA. Abstract K-425a.
- Bouza E et al. In: Program and abstracts of the 18th European Congress of Clinical Microbiology and Infectious Diseases, 19-22 April 2008, Barcelona, Spain. Abstract O464.
- Nelson R. Cochrane Database of Systematic Reviews 2007, Issue 3. Art. No: CD004610. DOI: 10.1002/14651858.CD004610.pub3.
- Fekety R et al. Am J Med 1989;86:15-19.
- Baines SD et al. J Antimicrob Chemother 2008;62:1046–1052.
- Nagy E & Földes J. J Antimicrob Chemother 1991;27:63-70.
- Musher DM et al. Clin Infect Dis 2006;43:421-427.
- Musher DM et al. Clin Infect Dis 2009;48: e41-6.
- McFarland LV et al. Am J Gastroenterol 2002;97:1769-1775.
- Pillai A & Nelson R. Cochrane Database of Systematic Reviews 2008, Issue 1. Art. No: CD004611. DOI: 10.1002/14651858.CD004611.pub2.
- Bassetti S et al. Am J Med 1998;105:71-72.
- Muñoz P et al. Clin Infect Dis 2005;40:1625-1634.
- Besselink MG et al. Lancet 2008;371:651-659.
- Friedenberg F et al. Dis Colon Rectum 2001;44:1176-1180.
- Abougergi MS & Kwon JH. Dig Dis Sci. 2010 Oct 6. (Epub ahead of print).
- Lowy I et al. N Engl J Med 2010;362:197-20.
- Herpers BL et al. Clin Infect Dis 2009;48:1732–1735.
- Lu CL et al. Int J Antimicrob Agents 2010;35:311-2.
- Russell G et al. Pediatrics 2010;126:e239-42.
- Silverman MS et al. Clin Gastroenterol Hepatol 2010;8:471-3.
- Grehan MJ et al. J Clin Gastroenterol. 2010;44:551-61.
- Lamontage F et al. Ann Surg 2007;245:267-272.
- Kato H et al. J Hosp Infect 2008; 70:194-195.
- Perera AD et al. Am Surg 2010;76:418-21.
- Bhangu A et al. J Investig Med 2010;58:621-4.
- Pépin J et al. Clin Infect Dis 2006;42:758-764.
- van Nood E et al. Euro Surveill. 2009;14(34):Pii=19316.
- Nieuwdorp M et al. Ned Tijdschr Geneeskd. 2008;152:1927-32.