Reprinted from
www.antimicrobe.org Reprinted from
www.antimicrobe.org
Clostridium difficile - Antimicrobial Therapy
General Considerations and Treatment for 1st and 2nd Episodes
As is shown in Table 2, metronidazole, vancomycin, teicoplanin, and fusidic acid have all demonstrated >90% efficacy in randomized comparative trials of C. difficile diarrhea treatment. The most clinical experience has been with vancomycin and metronidazole. Metronidazole, unlike vancomycin, is well-absorbed and fecal concentrations are low or absent in treated patients without diarrhea. However, bactericidal fecal concentrations are present in patients with C. difficile diarrhea and decrease as the diarrhea improves, suggesting that metronidazole may diffuse from the serum compartment directly through inflamed colonic mucosa or that increased intestinal transit time results in reduced absorption.
Although vancomycin was originally considered the drug of choice by several experts, metronidazole is the preferred first line agent for most cases of CDAD for several reasons. First, the efficacy of metronidazole has been confirmed in a subsequent prospective randomized trial demonstrating high cure rates and similar clinical recurrence rates when compared with vancomycin and teicoplanin. The earlier prospective comparison of metronidazole and vancomycin has been criticized because patients with positive stool cultures and negative stool toxin assays were included in the study and several reviews have cautioned that metronidazole be reserved for less severe cases of C. difficile diarrhea. Cure and recurrence rates in the more recent study and in the earlier prospective study were the same when patients with pseudomembranous colitis documented by endoscopy were analyzed separately suggesting that severity of disease does not influence the response to metronidazole. Another report also documented similar cure and recurrence rates, however, the duration of symptoms in this retrospective study was shorter in patients treated with vancomycin than for those who received metronidazole [3 vs. 4.6 days, respectively]. Neither of the two prospective studies showed any difference in the time to response between patients treated with these two agents. Because the response rates to either metronidazole or vancomycin are so high, a prospective study of enormous patient size would be necessary to demonstrate any subtle differences in clinical efficacy. The clinical experience with metronidazole for C. difficile diarrhea was reviewed at one institution over a 10-year period in which metronidazole was emphasized as first-line therapy (70%, 13%, and 17% of patients received metronidazole, vancomycin, and no treatment, respectively). The drug intolerance rate, treatment failure rate, and recurrence rate for 632 metronidazole-treated patients was 1%, 2%, and 6%, respectively.
Second, metronidazole is the least expensive treatment for C. difficile diarrhea at a cost of $4.00 for a 10-day treatment course (250 mg four times daily) compared to $175.00 or $873.00 for a 10-day course of vancomycin (125 mg or 500 mg four times daily) (local Chicago, IL pharmacy costs). Third, concern over the increase in vancomycin resistance among enterococci and, potentially, other important hospital-acquired pathogens has led to the recommendation that vancomycin use be limited. In particular, the Hospital Infection Control Practices Advisory Committee (HICPAC) has suggested that oral vancomycin be reserved for the treatment of patients who fail to respond to metronidazole or in severe, potentially life threatening illness. This recommendation has, in general, received wide-spread acceptance and most experts now recommend metronidazole as the initial specific agent of choice except for particular situations. Long term follow-up of >700 women who received metronidazole for vaginal trichomoniasis did not demonstrated increased cancer-related morbidity or mortality despite evidence of carcinogenic potential in some rodent studies. The safety of metronidazole in children has not been proven and metronidazole is a pregnancy category B drug. Despite having no FDA approval for this indication, we believe that metronidazole should be considered the drug of choice for C. difficile diarrhea.
Vancomycin was the first effective therapy for C. difficile diarrhea and has been the drug to which all subsequent therapies have been compared. Cure rates close to 100% for vancomycin at a dosage of 500 mg given orally three to four times daily for 10 days have been documented in all the prospective studies reported (Table 2). This response rate drops to 75% when the regimen is decreased to 125 mg orally four times daily for 5 days (Table 2). However, there were no treatment failures when the 'lower dose' vancomycin regimen (125 mg four times daily) was given for 10 days. Despite the predictably high fecal levels achieved and its remarkable clinical efficacy, ~15% of treated patients will still have a recurrence (Table 2). Teicoplanin has demonstrated efficacy similar to that achieved with vancomycin and metronidazole in two prospective clinical studies when studied at two different doses given twice a day for 10 days. Fusidic acid has been studied prospectively and although similar, high cure rates were demonstrated, the clinical recurrence rate was higher than for those treated with teicoplanin. Fusidic acid was also associated with at higher rate of side affects (gastrointestinal discomfort) compared with vancomycin and teicoplanin therapy. Clinical cure rates and C. difficile eradication rates with bacitracin are somewhat lower than with vancomycin and bacitracin should be considered as a second line agent in the treatment of C. difficile diarrhea. Treatment with the ion-exchange resin, colestipol, is clearly inferior to the agents described above and is not recommended for initial treatment.
The optimal dosage of most effective treatments for C. difficile diarrhea have not been determined but duration of therapy for 10 days appears necessary to achieve cure rates of 90%. Based on the available data we recommend the following dosages; metronidazole 250 mg orally four times daily or 500 mg three times daily and vancomycin 125 mg orally four times daily. Further studies will be necessary before teicoplanin and fusidic acid dosage recommendations can be made. All therapies should be administered orally and antiperistaltic agents should not be administered. An early controlled trial indicated that phenoxylate-atropine (Lomotil) used alone was deleterious to patients with C. difficile diarrhea and anecdotal reports suggest that its use may predispose to toxic megacolon. Loperamide has also been implicated in toxic megacolon. A recent retrospective study did not demonstrate any different response in patients who received antimotility agents (usually loperamide or codeine phosphate) in conjunction with specific therapy, but carefully designed prospective studies need to be conducted before any general recommendations can be made regarding the use of antimotility agents for C. difficile diarrhea.
Historically, patients with primary infection due to C. difficile (1st episode) who respond to therapy but develop a 2nd episode have responded to retreatment with the same therapy, an approach also endorsed by previous guidelines. A recent study conducted in the era of BI/NAP1 reiterated that 1st and 2nd episodes of CDAD can be effectively treated with metronidazole. However in this study, regardless of choice of therapy (metronidazole or vancomycin), complication rates were higher than previously reported for either drug. Another recent study evaluated the approach of combining rifampin with metronidazole versus monotherapy with metronidazole for 1st episode cases of CDAD (Table 2). This study showed no incremental benefit gained by the addition of rifampin to metronidazole. Also, recent studies utilizing metronidazole as the comparator agent in trials of investigational therapies such as nitazoxanide have demonstrated non-inferiority of the newer therapies to metronidazole (Table 2). In addition, vancomycin’s efficacy has been re-evaluated as a result of being used as a comparator for investigational therapies such as tolevamer. The efficacy of vancomycin was similar to high-dose tolevamer and superior to low-dose tolevamer (Table 2). Although these recent studies involved relatively small numbers of patients and severely-ill patients were excluded from participation, they do provide some insight into the efficacy of these established treatment options relative to newer agents.
Caveats to the general principle of utilizing the same therapy to manage 1st and 2nd episodes of CDAD which may influence the clinician’s decision to use a different agent for the 2nd episode include the presence of markers for severe disease and whether or not the gastrointestinal tract is functioning (e.g., presence of ileus, toxic megacolon). Treatment options for fulminant CDAD and for patients with compromised gastrointestinal function necessitating alternative therapy are discussed later in this chapter.
Table 2. Randomized, Comparative Trials of Oral Therapy for Clostridium difficile Diarrhea*
Agent
Regimen
Patients Studied
Cure Rate (%)
Time to Resolution
Relapse Rate (%)
References
Metronidazole
500 mg tid x 10d
31
94
3.2 d
16
Wenisch 1996 (139)
250 mg qid x 10d
76
87, 95
2.4 d, NS
42
Teasley 1983 (128) Musher 2006 (97)
Metronidazole + rifampin
500 mg tid x 10d
19
63
7.0d
Lagrotteria 2006 (75)
Vancomycin
500 mg tid x10d
31
94
3.1d
15
Wenisch 1996 (139)
500 mg qid x 10d
87
100
2.6 – 3.6 d
29
Teasley 1983, deLalla 1992,
Dudley 1986 (35,40,128)
125 mg qid x 7d
21
86
4.2 d
NS
Young 1985 (144)
125 mg qid x 5d
12
75
< 5.0 d
19
Mogg 1980 (93)
125 mg qid x 10d
80
91
2.0 d
7
Louie 2006 (83)
Teicoplanin
400 mg x bid x 10d
28
96
2.8 d
8
Wenisch 1996 (139)
100 mg bid x 10 d
26
96
3.4 d
28
deLalla 1992 (35)
Fusidic acid
500 mg tid x 10 d
29
93
3.8 d
28
Wenisch 1996 (139)
Bacitracin
20k-25k U qid x 7-10 d
36
78
2.5-4.1 d
23
Dudley 1986, Young 1985 (40, 144)
Nitazoxanide
500 mg bid x 7 d
40
90
NS
23
Musher 2006 (97)
500 mg bid x 10 d
36
89
NS
10
Musher 2006 (97)
Tolevamer
1 gm tid x 14 + d
72
67
4.0 d
4.9 d
Louie 2006 (83)
2 gm tid x 14 + d
70
83
2.5 d
NS
Louie 2006 (83)
Rifaximin
200 mg tid x 10 d
10
90
NS
NS
Boero 1990 (7)
Colestipol
10 gm qid x 5 d
14
36
< 5.0 d
Mogg 1980 (93)
Placebo
-- x 5 d
14
21
< 5.0 d
Mogg 1980 (93)
Recurrent CDAD (≥3rd episode)
Recurrent disease remains a challenge for clinicians, researchers, and particularly for the patients themselves. It is apparent from Table 2 that most people respond to treatment for C. difficile diarrhea. Recurrence of diarrhea after initially effective therapy is extremely common, however, occurring at rates of 5 to 30% regardless of the initial treatment regimen, similar to the experience with the hamster model of C. difficile diarrhea. As mentioned above, the mechanism of diarrhea recurrence may be different following metronidazole or vancomycin therapy as a result of decreasing metronidazole fecal levels on resolution of diarrhea or the bacteriostatic effect of vancomycin on C. difficile at the high concentrations achieved during therapy. In addition, diarrhea recurrence may be due to relapse of the original infecting strain or due to exogenous reinfection with a new strain. There is no evidence that recurrence or failure of initial treatment response in the case of metronidazole is due to acquisition of resistance to the initial therapeutic agent. A small group of patients suffer multiple recurrences of C. difficile diarrhea, frustrating to the physician as well as to the patient. Typically, these patients respond promptly to treatment after each recurrent episode but redevelop symptoms and positive stool assays within one to two weeks after discontinuation of treatment. As previously mentioned, nearly all cases of C. difficile diarrhea are precipitated by antimicrobial disruption of normal intestinal flora. It is possible that the specific agent used for treatment of the initial episode of diarrhea is the precipitating agent for the recurrent episode.
Several strategies have been explored for managing patients with multiple recurrences (Table 3). When interpreting the efficacy of these strategies, one must consider a number of pitfalls that cloud the interpretation of the efficacy of the intervention and limits the widespread extrapolation of the results to a broader population. Chief among these are the introduction of type 2 error and investigator bias due in part to small sample sizes and the open-labeled nature of these case reports, case series, and in some cases clinical trials. This being stated, recurrent CDAD remains a significant challenge resulting in nearly continuous treatment with vancomycin in a small number of patients. For these patients, it is helpful to have options. Discussed below is a distillation of studies and case series that may be considered for recurrent CDAD.
Table 3. Empirical Treatment Strategies for Patients with Multiple Recurrences of C. difficile Diarrhea
Strategy
References
Saccharomyces boulardii
McFarland 1994, Surawicz 1989, Surawicz 2000 (62, 81,82)
Lactobacillus GG
Biller 1995, Gorbach 1987 (10,38)
Rectal infusion of feces or a synthetic fecal bacterial flora
8 Reports summarized by Borody 2000 (19)
Oral infusion of donor stool via NG tube
Aas (3)
Administration of a non-toxigenic C. difficile strain
Seal 1987 (119)
Vancomycin and rifampin combination
Buggy 1987 (23)
Vancomycin in tapering doses/pulsed doses
Tedesco 1985 (89, 130)
Treatment with active agent followed by 2 week course of rifaximin
Johnson, CID 2007 (in press)
Intravenous gamma globulin
Hassett 1995, Warny 1995, Leung 1991, Wilcox 2004, McPherson 2006 (59, 77, 92, 137, 140)
Whole bowel irrigation
Liacouras 1996 (80); Persky 2000 (110)
No treatment with careful observation
Gerding 1995, Bartlett 1992 (6, 36)
Of the biotherapeutic approaches, treatment with the yeast Saccharomyces boulardii is probably the best studied. In the initial report of 13 patients with multiple diarrhea recurrences who received vancomycin for 10 days and S. boulardii for 28 days, 11 patients had no further recurrences. A subsequent randomized, placebo-controlled study showed that S. boulardii in combination with standard therapy was more effective than standard therapy alone in preventing recurrences in patients who had a history of more than one C. difficile diarrhea episode. A more recent study in which the standard therapy and therapeutic dose was stratified showed a significant decrease in recurrent episodes among patients treated with high-dose vancomycin (2 gm/day) and S. boulardii compared with to high-dose vancomycin and placebo. Neither low-dose vancomycin or metronidazole with or without S. boulardii was effective. Still today, the data to support the use of lactobacillus-containing and S. boulardii-containing regimens are poor and conflicting. The best study supporting the use of S. boulardii did not convince the Food and Drug Administration to approve it for this use. A recent meta-analysis suggested that probiotics are effective; however, because of the heterogeneity of study methodologies and patient populations, it is not scientifically possible to conduct a meta-analysis of the probiotic literature. Critical examinations of an earlier meta-analysis of probiotic use cast similar doubt upon the authors’ conclusions for related reasons. A systematic review of probiotic efficacy indicated that the current literature does not support the use of probiotics for CDAD. Moreover, the literature is evolving to demonstrate that the so-called non-pathogenic strains of the various fungi and bacteria used in the currently marketed probiotics have caused numerous cases of bacteremias due to Lactobacillus species and fungemias due to S. boulardii in both immunocompetant and immunocompromised hosts.
Other, non-antimicrobial, biotherapeutic approaches tested in open trials include treatment with rectal infusions of feces from normal hosts and the infusion of a mixture of bacteria simulating a normal flora. A recent review of the eight reports of this approach involving instillation of feces or fecal bacteria was optimistic, indicating a good cure rate without recurrence for most patients. Despite the aesthetic concerns of this approach and the potential concern for transmission of other infectious agents, investigators have tried to refine this method using different carrier suspensions and modes of infusion. An additional novel approach that was partially successful in two patients was introduction of a non-toxigenic strain of C. difficile.
Antimicrobial approaches that have been studied in small open trials include combination therapy with vancomycin (125 mg PO qid) and rifampin (600 mg PO bid) given for 7 days and a study of 163 patients evaluating a mixture of strategies including treatment with different vancomycin doses for 10 days, treatment which was followed by tapering doses for a mean of 21 days, or treatment with vancomycin (or no treatment) immediately followed by pulsed dosed vancomycin (a single dose of 125, 250, or 500 mg given every 3 days) for a mean of 27 days. High total daily dose (2g/d) vancomycin for 10 days performed slightly better than traditional total daily dose (500 mg/d) vancomycin in terms of reduced recurrence rates. Treatment followed by tapered vancomycin and treatment followed by pulsed vancomycin resulted in a significantly reduced recurrences when compared with treatment alone, with the latter regimen being associated with the least amount of recurrences. Tedesco and colleagues also studied vancomycin tapered regimens in 22 patients. The regimen consisted of: week 1 (500 mg/d), week 2 (250 mg/d), week 3 (125 mg/d), weeks 4-6 (pulsed dosed vancomycin 125 mg every 3 days). Collectively, these descriptive studies demonstrate that tapered and pulsed dosed vancomycin regimens seem to be effective in reducing recurrences and large scale prospective studies are warranted. Metronidazole has also been studied in a pulsed and tapered regimen but too few patients were evaluable to draw any conclusions. The concept of the pulsed regimen takes advantage of the fact that our current treatments (e.g., vancomycin, metronidazole) are only effective against vegetative forms (but not spore forms) of C. difficile. By allowing for a relatively antibiotic-free period in the gastrointestinal tract, it is theorized that spore forms will recrudesce into vegetative forms rendering them susceptible to subsequent therapy.
Cholestyramine, an anion-exchange binding resin has also been anecdotally reported as useful in this setting. However, the efficacy of anion-exchange binding resins in the treatment of primary C. difficile diarrhea has been poor when evaluated under more rigorous conditions (Table 1). The only placebo controlled trial that has evaluated the toxin binding capability of anion exchange resins was conducted using colestipol, which demonstrated that these agents were comparable to placebo in terms of reducing fecal excretion of C. difficile toxins. Animal models have confirmed cholestyramine’s poor toxin binding affinity relative to other toxin binding compounds. Cholestyramine does, however, bind to a variety of drugs, including the very treatments themselves (e.g., vancomycin), resulting in a reduction of biological activity in stool. Based on the lack of efficacy data to support the use of cholestyramine and the potential for deleterious effects when combined with oral treatments for C. difficile-associated disease, their use cannot be recommended. Intravenous administration of immune globulin (IVIG) may also benefit a subgroup of patients with multiple recurrences of C. difficile diarrhea. In two of the studies, patients had low levels of serum anti-toxin A IgG and the patient in the third study had selective IgG1 deficiency. The largest study published to date of IVIG was a retrospective, observational evaluation of 14 patients. Six of 14 patients responded clinically, and no relapse occurred within the timeframe reported. The doses used ranged from 150-400 mg/kg administered as a single-dose (one patient received a second dose). The median response time was 10 days. In another study, 3 of 5 patients treated with doses of IVIG between 300 to 500 mg/kg (usually 400 mg/kg), were deemed successes, with resolution occurring within 11 days. An upside to IVIG is that it may provide therapeutic option for patients with severe/relapsing disease where no other therapeutic options are available. Unfortunately, marginal efficacy, lack of data regarding the optimal dose, cost (~$U.S. 1,500/dose for a 70kg patient) and frequent shortages are significant disadvantages.
Table 1. In Vitro Susceptibility of Clostridium difficile to Potential Therapeutic
Antimicrobial Agents (Data Selected from Representative Studies).
MIC (mg/ml)
Agent
n
MIC50
MIC90
Range
Reference
Rifampin
55
< 0.001
0.002
<0.001-0.002
Bacon 1991 (8)
Metronidazole
50
0.29
0.6
0.25-1.0
Bannatyne 1987 (9)
Ramoplanin*
70
0.25
0.25
0.12-0.5
Biavsco 1991 (16)
Tiacumicin B* 15
15
0.25
0.25
0.12-0.5
Swanson 1991 (126)
Tiacumicin C*
15
0.5
1.0
0.25-1.0
Swanson 1991 (126)
Teicoplanin
70
0.5
0.5
0.25-1.0
Biavsco 1991 (16)
Vancomycin
70
1.0
2.0
1.0-2.0
Biavsco 1991 (16)
Tinidazole
50
0.55
3.74
0.5-4.0
Bannatyne 1987 (9)
Bacitracin
110
(MIC mode = 16 mg)
8.0-32.0
Young 1985 (144)
Fusidic Acid
20
(17/20 susceptible to 10 mg, disc testing)
Burdon 1979 (24)
* Investigational agents
One report suggested that whole bowel irrigation with a polyethylene glycol solution (Golytely) followed by a course of vancomycin was successful in terminating multiple recurrences of C. difficile colitis in two young children. A more recent report combined Golytely lavage with administration of donated feces directly through a colonoscope to all segments of the colon. The theoretical advantage of this refinement over previous fecal replacement strategies was a more thorough reduction of the resident flora and recolonization of the entire colon. Finally, no treatment with careful observation has also been advocated in selected patients.
Recently, 8 patients ranging in age from 43-88 years with 4-8 prior episodes of CDAD and receiving between 79-372 total days of treatment were placed on an unconventional regimen as described momentarily (Clin Infect Dis 2007;44:846-848.). In these patients, a variety of treatment strategies had been previously attempted, including combinations of standard therapy (metronidazole or vancomycin) and rifampin, probiotics, tapered or pulsed vancomycin regimens. Although diarrhea would cease for a period of time, symptoms would invariably return (mean 10.5 days, range 1-59 days). These patients were treated until symptoms resolved and then immediately started on rifaximin 400 to 800 mg daily (divided into 2 to 3 doses) for 2 weeks. In this small sample, 7 of 8 patients remained symptom free for a range of 51 to 431 days. One patient was reported to have a recurrent episode, was retreated with a 2-week course of rifaximin and has remained symptom free for 9 months. One patient developed high-level resistance during therapy (MIC values for the pre- and post-therapy isolates were 0.0078 µg/mL and >256 µg/mL, respectively), although the patient remained asymptomatic. The results of this case series are promising for a very frustrating clinical problem facing many clinicians today; however, it should be kept in mind that this use of rifaximin is considered “off-label” (rifaximin is not approved by the FDA for the treatment of CDAD) and that resistance was clearly documented to occur during exposure to rifaximin.
Fulminant CDAD:
The hypervirulent C. difficile strain (BI/NAP1) has been associated with more severe disease by some investigators. A retrospective observational study that included infection with BI/NAP1 strains demonstrated that whether metronidazole or vancomycin was used for 2nd episode cases, the complication rates were higher for both options compared with historical controls. However, if severe CDAD is suspected, several determinations should be made. First, is ileus or toxic megacolon suspected? If so, see section below and early surgical consultation must be emphasized. For severe CDAD (defined as elevated white blood cell count, ascites, and/or hypotension) but in the absence of ileus or toxic megacolon, oral vancomycin may be considered for initial therapy and is recommended by some experts. The patient should be monitored daily for response to therapy, including the number and consistency of bowel movements as well as for markers necessitating surgical evaluation.
Severe Ileus/Toxic Megacolon:
The most serious manifestation of C. difficile disease is toxic megacolon which, paradoxically, may present without diarrhea. Treatment of patients with toxic megacolon or severe illness is difficult and controversial but several attempts have been made to achieve effective antimicrobial concentrations at the site of infection when the oral route is compromised. Some authors advocate treatment with intravenous metronidazole or with intravenous vancomycin at dosages 2 grams per day, placement of a long catheter in the small intestine and instillation of vancomycin. Another approach is to administer vancomycin by rectal enema. A strategy used successfully at one institution in six patients with severe ileus included vancomycin administered by nasogastric tube and by retention enema plus intravenous metronidazole (Table 4). If these approaches are unsuccessful surgical intervention or colonic decompression should be considered (see Adjunctive Therapy).
Table 4. Empirical Treatment Protocol for Clostridium difficile-Infected Patients with Severe Ileus* _______________________________________________________________________________
Vancomycin 500 mg per rectum every 6 hA, plus
Vancomycin 500 mg via NG tube every 6 hB, plus
Metronidazole 500 mg IV every 6 h
______________________________________________________________________________
* Reference: 67
ALiquid intravenous formulation diluted in 100 cc of normal saline: Insert #18 Foley catheter into rectum 4 to 8 inches,
fill the balloon to 30 cc, gently pull catheter down, instill vancomycin and clamp catheter for 60 minutes, deflate balloon
and remove catheter.
BLiquid intravenous formulation diluted with at least 10 cc of fluid: clamp nasogastric tube for 60 minutes after each instillation.
When Oral Therapy Is Not Possible:
Although several options exist for the treatment of C. difficile diarrhea (Table 2), all well-studied regimens employ oral therapy. When the oral route is compromised, intravenous therapy may be considered. Colonic concentrations of vancomycin are negligible after intravenous administration and there is little support for this option. Fecal concentrations of metronidazole, however, are similar when metronidazole is given orally or intravenously in the setting of acute C. difficile diarrhea. Anecdotal experience supports intravenous metronidazole therapy for C. difficile diarrhea. A recent retrospective review of 10 patients who received at least two days of intravenous metronidazole as the initial therapy for acute C. difficile colitis in whom oral therapy was not possible showed symptomatic improvement in the majority of patients during therapy without subsequent complications requiring surgical intervention. A randomized, prospective study is needed, but this therapy alone may be inadequate in patients with severe adynamic ileus. Therefore, in patients with severe manifestations of C. difficile disease, other methods to ensure effective antimicrobial concentrations at the site of infection should also be considered (Table 4).