Find synonyms Find exact match
Disclosures: R Balfour Sartor, MD Consultant/Advisory Boards: Salix (IBS, hepatic encephalopathy, possible IBD [Rifaximin]); Dannon/Yakult (Microbiotic [Probiotics]); GlaxoSmithKline (IBD [Probiotic foods]).] Paul Rutgeerts, MD, PhD, FRCP Grant/Research Support: Centocor; Merck; UCB; Abbott (IBD treatment). Consultant: Centocor; Merck; UCB; Abbott; Millennium/Takeda; Genentech/Hoffman LaRoche; Neovacs: Merck/Serono; Bristol Myers Squibb; Robarts; Tillotts; Pfizer; Falk Pharma; Galapagos (IBD treatment). Speaker's Bureau: Centocor; Merck; UCB; Abbott (IBD treatment). Shilpa Grover, MD, MPH Employee of UpToDate, Inc.
Contributor disclosures are reviewed for conflicts of interest by the editorial group. When found, these are addressed by vetting through a multi-level review process, and through requirements for references to be provided to support the content. Appropriately referenced content is required of all authors and must conform to UpToDate standards of evidence.
INTRODUCTION — Antibiotics have a well-established role in the treatment of septic complications of the inflammatory bowel diseases (IBD), such as abscesses and wound infections. In contrast, their benefit in the treatment of the primary disease processes of Crohn's disease, ulcerative colitis, and pouchitis has not been well-established in carefully-designed clinical trials, although they are used commonly [1-4]. The rationale for antibiotic therapy in IBD is based upon a large body of evidence demonstrating that luminal bacteria and perhaps fungi have an important role in the pathogenesis of IBD [5,6]. (See "Immune and microbial mechanisms in the pathogenesis of inflammatory bowel disease".)
Treatment with antibiotics has the potential to influence the course of IBD by several mechanisms:
●Decreasing the concentrations of bacteria and fungi in the gut lumen
●Altering the composition of the intestinal microbiota to favor beneficial bacteria
●Decreasing bacterial tissue invasion and treating microabscesses
●Decreasing bacterial translocation and systemic dissemination
Some of the antibiotics that have been used to treat IBD may also act as immunomodulators and thereby exert their benefit by mechanisms other than their antimicrobial effects [7,8]. In addition, some antibiotic regimens have been designed to treat mycobacterial infection, which has been hypothesized to have a role in the development of Crohn's disease .
Therapeutic manipulation of the intestinal microbiota can also be accomplished with other strategies including prebiotics (ie, dietary components that promote the growth of beneficial bacteria) and probiotics (ie, beneficial bacteria). How these strategies might be used together in the treatment of IBD remains a topic of active investigation. (See "Probiotics for gastrointestinal diseases".)
This topic review summarizes clinical trials of antibiotic treatment for IBD, focusing primarily on Crohn's disease, which has been the subject of the greatest number of clinical studies.
Luminal disease — Several controlled trials and observational studies in the treatment of active Crohn's disease with antibiotics have been published [9-17]. Most were small (16 to 213 subjects), short-term (range 2 to 24 weeks), and have important methodologic limitations. Furthermore, these studies used different entry criteria, permitted use of various concomitant therapies, and focused on different endpoints, thus limiting direct comparisons between studies. However, in a 2011 meta-analysis of randomized controlled trials, antibiotics were superior to placebo for the induction of remission of active Crohn's disease (RR of active disease not in remission 0.85, 95% CI 0.73-0.99) . In addition, this meta-analysis found that antibiotics were more effective than placebo for maintenance of remission of Crohn's disease (RR of relapse 0.62, 95% CI 0.43-0.96). In a 2012 meta-analysis of 10 randomized trials, antibiotic therapy was associated with an improvement in clinical symptoms as compared with placebo (odds ratio 1.35 [95% CI 1.16-1.58]) .
Clinical experience with antibiotics in the treatment of IBD has far outpaced published scientific evidence. The published literature and clinical experience considered together suggest a modest benefit of metronidazole (10 or 20 mg/kg/day) or the combination of metronidazole and ciprofloxacin for primary or adjunctive therapy of colonic Crohn's disease, but not for isolated small intestinal disease. Several other single and combination antibiotic regimens also appear to be effective, although there is somewhat less experience. However, many uncertainties remain, including the magnitude and duration of the benefit, the optimal dose and duration of therapy, and strategies to reduce adverse effects. Metronidazole in particular can be associated with permanent peripheral neuropathy when used for prolonged periods, especially at higher doses.
Most antibiotics only transiently alter luminal bacterial concentrations; bacteria repopulate the mucosa after therapy ceases . Antibiotics also promote proliferation of resistant strains.
One study demonstrated that chronic use of metronidazole eliminated Bacteroides species in patients with Crohn's disease for at least six months . Furthermore, Bacteroides clearance was associated with disease remission suggesting that treatment of anaerobes is important. However, ciprofloxacin has poor anaerobic coverage, but may be effective in clearing E. coli, which have been implicated in ileal Crohn's disease [22-24].
In one report, the combination of metronidazole and ciprofloxacin significantly decreased mucosally-associated bacteria after 1 and 7 to 14 days of treatment . However, mucosally associated bacteria rebounded to higher than pretreatment levels 1 to 18 weeks after antibiotic exposure ceased, but returned to pretreatment levels by six months after cessation of antibiotics. Neither Bacteroides nor Enterobacteriaceae (E. coli) groups were permanently suppressed.
Thus, the relationship between the bacterial targets of antimicrobial agents and therapeutic efficacy remains obscure. Whether results observed with specific antimicrobial agents can be generalized to others is uncertain.
The following summarizes the largest studies in which the patient populations, clinical endpoints, and interventions were best described.
●One study included 105 patients with active disease who were randomly assigned to metronidazole (20 mg/kg or 10 mg/kg) or placebo for 16 weeks . Only 56 patients completed the study. Withdrawals were due to deterioration of symptoms, adverse events, or protocol violations. Among those who completed the study, there was a significant decrease in the Crohn's disease activity index, and a decrease in CRP with metronidazole therapy. However, there were no differences in remission rates between those receiving metronidazole or placebo. Patients with colonic involvement fared better than those with isolated ileal involvement, which has been a consistent finding in several studies. The high dropout rate in this study makes the results difficult to interpret, although the results suggest a benefit of this therapy for Crohn’s colitis and ileocolitis.
●Another trial included 78 patients with active disease who were randomly assigned to either metronidazole or sulfasalazine in a double-blind, cross-over study for two four-month periods . Metronidazole was slightly more effective than sulfasalazine.
●One controlled trial included 47 patients with moderately active, resistant disease who were randomly assigned to ciprofloxacin (500 mg twice daily) monotherapy or placebo for six months . At the end of the study, Crohn's disease activity scores were significantly lower in the antibiotic group.
●Another study included 40 patients with mild to moderate flares of Crohn's disease who were randomly assigned to ciprofloxacin (1 g daily) or Pentasa (4 g daily) for six weeks . At the end of the study complete remission was observed in a similar proportion of patients (55 versus 56 percent, respectively).
These findings do not necessarily indicate that the two drugs are equivalent since the study was not designed to determine equivalence. Furthermore, because there was no placebo arm, it is difficult to know for certain whether the improvement in either arm could have been due to the natural history of the disease.
Side effects of ciprofloxacin include photosensitivity, tendonitis, and rare tendon rupture, inhibition of cartilage growth in fetuses and children, thrush, candidiasis, and rare prolongation of the QT interval .
Combination metronidazole and ciprofloxacin
●One study of combination therapy included 134 patients with ileal Crohn's disease (with or without right-sided colonic involvement) who were randomly assigned to ciprofloxacin (500 mg twice daily) plus metronidazole (500 mg twice daily) or placebo for eight weeks . All patients received oral budesonide (9 mg once daily). At the end of the study no significant differences were observed in the rates of remission in the antibiotic or placebo groups (33 versus 38 percent, respectively). More patients with disease in the colon were in remission after antibiotics (53 versus 25 percent), although the results did not achieve statistical significance.
Discontinuation of treatment because of adverse events occurred significantly more often in the antibiotic group (20 versus 0 percent). The authors concluded that the addition of these antibiotics to budesonide was ineffective in the treatment of active ileal Crohn's disease but that treatment may be beneficial when the disease affects the colon.
Response to treatment correlated with serologic response to bacterial or fungal antigens . Although not statistically significant, patients with predominant serologic responses to OmpC and I(2) had the highest remission rates, suggesting that serologic responses to commensal bacteria may predict therapeutic responses to antibiotics.
●Another report included 41 patients who were randomly assigned to combination therapy with ciprofloxacin (500 mg twice daily) plus metronidazole (250 mg four times daily) or methylprednisolone (0.7 to 1 mg/kg per day followed by a taper) for 12 weeks . At the end of the study, clinical remission was observed in a larger proportion of patients in the steroid group (63 versus 46 percent) although the results did not achieve statistical significance.
●An open-label study included 72 patients who were treated with ciprofloxacin (500 mg twice daily) plus metronidazole (250 mg three times daily) for 10 weeks . A clinical response was observed in 67 percent of patients who were not taking concurrent prednisone and 90 percent of patients who were receiving prednisone. The response was also higher in those with colonic disease with or without ileal involvement compared with those with ileal disease alone (86 versus 64 percent). Five patients discontinued antibiotics because of adverse events.
●Rifaximin (an oral antibiotic with a broad antimicrobial spectrum and minimal systemic absorption) was associated with improvement in several observational, uncontrolled studies [28-30]. A controlled trial of 83 patients comparing rifaximin 800 mg twice daily, 800 mg once daily to placebo for 12 weeks found that treatment failure was significantly less likely with rifaximin twice daily than placebo . The doses used in the above reports ranged from 200 mg three times daily to 800 mg twice daily and response was generally seen within four weeks.
●In a phase II randomized trial of a formulation of rifaximin with increased intestinal availability (extended intestinal release [EIR]), 402 patients with moderately active Crohn’s disease were assigned to receive rifaximin-EIR 400 mg twice daily, rifaximin-EIR 800 mg twice daily, rifaximin-EIR 1200 mg twice daily, or placebo for 12 weeks . Remission rates at the end of treatment were higher in the patients treated with rifaximin-EIR 800 mg twice daily compared with those treated with placebo (62 versus 43 percent). Remission rates in those treated with 400 mg twice daily (54 percent) or 1200 mg twice daily (47 percent) did not differ significantly from placebo.
●An open-label study included 25 patients with active disease who were treated with clarithromycin (250 mg twice daily) for four weeks with continued treatment for up to 12 weeks in those who showed a response . A response was observed in 64 percent of patients while 48 percent achieved clinical remission. Treatment was continued in 11 patients for a median of 28 weeks during which 73 percent remained in remission. However, no benefit after three months of 1 gm once daily was seen by the same group of investigators in a later placebo-controlled trial involving 41 patients with active disease, although a possible transient benefit at one month was noted on a post-hoc analysis . An uncontrolled trial of clarithromycin 200 mg twice daily for four weeks demonstrated response in 57 percent and remission in 36 percent of 14 Japanese patients .
Anti-tuberculosis antibiotic trials — Several observations have suggested a link between Crohn's disease and mycobacteria, which provided the rationale for several small trials of antituberculous therapy in patients with Crohn's disease.
At least two meta-analyses suggested an overall benefit, although it is unclear if response was due to the antituberculosis treatment or nonspecific effects of the antibiotics. However, a subsequent, large (213 patients) placebo-controlled trial of clarithromycin, rifabutin, and clofazimine for two years found no lasting benefit, although a significant improvement with antibiotics was seen at four months (66 percent remission on antibiotics versus 50 percent with placebo) . It is likely that benefit was due to the effects of this treatment on commensal bacteria rather than Mycobacterium avium paratuberculosis (MAP). Unfortunately, mucosal levels of MAP were not measured before and after therapy in this study. Methods to analyze in vitro antibiotic sensitivity to human Crohn's disease MAP isolates show sensitivity to clarithromycin and azithromycin, but relative resistance to isoniazid . Surprisingly, 7 out of 10 human MAP isolates were sensitive to ciprofloxacin in this study. (See "Investigational therapies in the medical management of Crohn disease".)
Fistulizing disease — Few studies have been published that focus primarily on fistulizing Crohn's disease. Most have been case series in patients with perianal disease. A 2011 meta-analysis that included three trials found that in patients with perianal fistulas, treatment with either ciprofloxacin or metronidazole resulted in reduced fistula drainage more often than placebo (RR 0.8, 95% CI 0.66-0.98) .
Despite the limited available evidence, these studies have led to use of antibiotics in the treatment of patients with new onset, simple perianal fistulae. Remission rates exceeding 50 percent have been described with high dose metronidazole (20 mg/kg per day in divided doses) . However, 50 percent of patients taking metronidazole for a mean 6.5 months developed neurologic side effects requiring dose reduction or discontinuation . In a second small study, improvement (closure of at least 50 percent of fistulae) occurred in 40 percent of patients treated with ciprofloxacin, compared with 14 percent with metronidazole and 13 percent with placebo. Early termination of treatment occurred in ≤13 percent of ciprofloxacin and placebo-treated patients, but in 71 percent treated with metronidazole .
Lower antibiotic doses also appear to be effective and reduce the risk of side effects. A double-blind, placebo-controlled study of either ciprofloxacin 500 mg twice daily or placebo in combination with infliximab 5 mg/kg for 18 weeks showed a significantly increased response with ciprofloxacin (73 percent) compared with placebo (39 percent) . Of potential importance, patients with Crohn's-related NOD2 variants had worse outcomes to antibiotics and infliximab (0 percent) than did patients with NOD2 wild type genotypes (33 percent healing) . In an uncontrolled study of concomitant therapy of perianal fistulas, patients who received azathioprine plus antibiotics achieved a significantly better response (48 percent) than those without immunosuppression (15 percent) . (See "Perianal complications of Crohn disease".)
We commonly use antibiotics, either metronidazole (10 to 20 mg/kg per day), ciprofloxacin (500 mg twice daily), or a combination of these agents, as first-line treatment of uncomplicated perianal fistula following drainage of associated abscesses. Antibiotics are continued for three months in most situations, with close observation for peripheral neuropathy in metronidazole treated patients. We use antibiotics in combination with 6-mercaptopurine, azathioprine or infliximab for more complex or refractory cases. Further studies need to be performed with agents that target Gram positive skin bacteria, which were found to be the predominant bacterial species colonizing perianal fistulas . (See "Perianal complications of Crohn disease".)
Postoperative recurrence — Optimal strategies to prevent postoperative recurrence of Crohn's disease are evolving. One trial showed that the combination of metronidazole and azathioprine was superior to treatment with metronidazole alone . A potential role for antibiotics is supported by the observation that recurrent disease develops only when the mucosa is reexposed to luminal contents, suggesting that bacteria may have a role in promoting disease recurrence . (See "Medical prophylaxis of postoperative Crohn disease", section on 'Antibiotics'.)
ULCERATIVE COLITIS — Controlled trials of narrow spectrum antibiotics in ulcerative colitis have not demonstrated a consistent benefit. Thus, they have little, if any, role in the treatment of active disease except in patients with septic complications related to fulminant colitis in whom they may help to avert a life-threatening infection. However, limited studies raise the possibility of an effect of broad spectrum antibiotic therapy. (See "Management of severe ulcerative colitis".)
The following illustrate the range of findings in the largest reports:
●70 patients with mild or moderately active ulcerative colitis were randomly assigned to ciprofloxacin (250 mg twice daily) or placebo for 14 days in addition to oral and rectal corticosteroids as well as olsalazine . No significant differences in remission rates were observed (71 versus 72 percent).
●39 patients with severe ulcerative colitis were randomly assigned to intravenous metronidazole (500 mg every eight hours) or placebo for five days . All patients also received intravenous and rectal corticosteroids. No significant differences in improvement were noted in the two groups (74 versus 70 percent).
●39 patients with severe ulcerative colitis were randomly assigned to either intravenous metronidazole (500 mg three times daily) and tobramycin (4 mg/kg three times daily) or placebo in addition to parenteral nutrition, intravenous and rectal corticosteroids . No significant differences in clinical improvement were observed (65 versus 64 percent). Similar results from the same group were reported in a second placebo-controlled trial comparing ciprofloxacin with placebo in severe ulcerative colitis .
●In contrast to the above reports, a benefit of a six-month course of ciprofloxacin (500 to 750 mg twice daily) was suggested in a placebo-controlled trial of 83 patients who had responded poorly to conventional therapy with steroids and mesalamine . Treatment failure (defined endoscopically and clinically) was observed in 44 percent of those receiving placebo compared with only 21 percent of those receiving ciprofloxacin. This study has not yet been confirmed.
●A controlled trial included 28 patients with moderate to severe, steroid refractory disease who were randomly assigned to either rifaximin 400 mg twice daily or placebo for 10 days, while intravenous corticosteroids were continued . Although overall clinical outcomes were not different, stool frequency, rectal bleeding, and sigmoidoscopy scores were significantly improved in the rifaximin-treated group. Similar findings were described in an uncontrolled study . Similarly, the broad spectrum combination of amoxicillin, tetracycline, and metronidazole was beneficial in a small (20 patient) two-week study .
●In one study, levels of enteric bacteria in the colonic biofilm transiently decreased following treatment with metronidazole and ciprofloxacin, but mucosal bacteria rebounded to higher than pretreatment levels within one week of stopping antibiotics . This provides a basis for a lack of sustained clinical responses.
These data are reflected in several meta-analyses. A 2011 meta-analysis suggested a modest benefit of antibiotics as compared with placebo in the induction of remission in patients with active ulcerative colitis (RR of active ulcerative colitis 0.64, 95% CI 0.43-0.96) . In a 2012 meta-analysis of nine randomized trials, antibiotic therapy was associated with an improvement in clinical symptoms as compared with placebo (OR 2.17, 95% CI 1.54-3.05) .
POUCHITIS — Overgrowth with commensal bacteria has been hypothesized to have an important role in the development of pouchitis in patients with an ileal pouch-anal anastomosis, based on a response to short-term treatment with metronidazole and other antibiotics in most patients [44,46,49,50]. A clinical benefit has also been observed after treatment with a combination of probiotics to prevent relapse of chronic, relapsing pouchitis . (See "Pouchitis: Epidemiology, clinical manifestations, and diagnosis" and "Probiotics for gastrointestinal diseases".)
BACTERIAL OVERGROWTH — Small intestinal bacterial overgrowth is relatively common in Crohn's disease due to intestinal strictures, resection of the ileocecal valve or internal fistulae. Several antibiotic regimens, including ciprofloxacin, metronidazole and rifaximin, are highly effective in normalizing hydrogen breath tests, but have variable effects on improving symptoms of bloating, diarrhea, and abdominal pain [52,53]. Rifaximin only transiently cleared bacterial overgrowth, with hydrogen breath tests becoming abnormal 30 days after cessation of therapy .
CLOSTRIDIUM DIFFICILE INFECTION — C. difficile toxin-induced inflammation is an important risk factor for inflammatory bowel disease (IBD) flares and can be a complication of antibiotic therapy. Several studies suggest an increased frequency of C. difficile infection in patients with IBD, including those with ileal pouches after total colectomy [54-57]. In one series, 61 percent of C. difficile-infected IBD patients had been exposed to antibiotics . Treatment with metronidazole or oral vancomycin clears the C. difficile in most patients, but response of IBD patients to treatment is variable with increased frequency of metronidazole resistance . Factors associated with worse outcomes (death or colectomy) among patients with IBD and C. difficile in one study included a serum albumin less than 3 g/dL, hemoglobin less than 9 g/dL, and a serum creatinine above 1.5 mg/dL .
Given the high frequency of superinfection and high rate of colectomy, all hospitalized patients with a flare of disease should be evaluated for C. difficile toxin A and B. A study suggests that IBD patients with C. difficile superinfection who receive immunosuppressant medications have a worse outcome .
SUMMARY AND RECOMMENDATIONS
Active luminal Crohn's disease — The published literature and clinical experience considered together suggest a modest benefit of metronidazole (10 or 20 mg/kg/day) or the combination of metronidazole and ciprofloxacin 500 mg twice daily for primary or adjunctive therapy of colonic Crohn's disease, but not for isolated small intestinal disease. While evidence for rifaximin exists, there is somewhat less experience with this agent. However, there remain many uncertainties including the magnitude and duration of the benefit, the optimal dose and duration of antibiotic therapy, whether combination therapy is superior to single agents, and strategies to reduce adverse effects. Metronidazole in particular can be associated with permanent peripheral neuropathy when higher doses are used for prolonged periods. (See "Overview of the medical management of mild to moderate Crohn disease in adults".)
Fistulizing Crohn's disease — We commonly use antibiotics (either metronidazole or ciprofloxacin alone or in combination) as first-line treatment of uncomplicated perianal fistula following drainage of associated abscesses. Antibiotics are continued for three months in most situations. We use antibiotics in combination with 6-mercaptopurine, azathioprine, or infliximab for more complex or refractory cases.
Postoperative recurrence of Crohn's disease — The available data suggest a modest benefit of short-term metronidazole and similar agents in preventing postoperative recurrence in patients with ileocolonic anastomosis. However, their role is uncertain particularly considering the need for long-term therapy and the potential for adverse effects. Some data suggest that the combination of a three-month course of metronidazole and long-term azathioprine is superior to metronidazole alone, but this observation needs to be replicated. (See "Medical prophylaxis of postoperative Crohn disease".)
Ulcerative colitis — Controlled trials of narrow spectrum antibiotics in ulcerative colitis have not demonstrated a consistent benefit. Thus they have little, if any, role in the treatment of active disease except possibly in patients with disease refractory to traditional medications or in those with fulminant colitis in whom they may help to avert a life-threatening infection. There is some developing evidence that broad spectrum rifaximin or combinations of antibiotics may have a role in treating ulcerative colitis, but more extensive studies need to confirm these preliminary results. (See "Management of severe ulcerative colitis".)
Pouchitis — Antibiotics have an important role in the treatment of pouchitis. Metronidazole, ciprofloxacin, and rifaximin are effective, although ciprofloxacin is better tolerated than metronidazole and may be more effective. The combination of ciprofloxacin with either metronidazole, rifaximin, or tinidazole can be effective in refractory disease. Determining coliform sensitivities may help plan individualized therapies in patients with pouchitis that becomes refractory to treatment. (See "Pouchitis: Epidemiology, clinical manifestations, and diagnosis".)
All topics are updated as new information becomes available. Our peer review process typically takes one to six weeks depending on the issue.