Antibiotics
The first step in the management of perianal Crohn’s disease is usually the institution of antibiotics. Although there have been several control trials looking at the efficacy of metronidazole in the treatment of active Crohn’s disease, none of them have randomized for fistula response. Because there are no controlled trials showing efficacy we must rely on open-label studies and personal experience.
For example, in an uncontrolled study reported by Bernstein and colleagues (1987), 21 patients with fistula were evaluated, with clinical response being observed in 20 of the 21. Complete healing was seen in 10 of 18. High dose metronidazole was used in these studies, which can produce significant side effects in a large percentage of patients. Improvement was seen in about 6 to 8 weeks; however, my personal experience is that response is observed more quickly, usually within 2 to 3 weeks.
A follow-up report of 17 of these patients with 9 additional patients showed that relapse frequently occurred but patients often responded to restarting the metronidazole. Unfortunately, only about one-third of patients could successfully stop the metronidazole without experiencing a relapse. As noted, with the high dose administered, neuropathy may be encountered which may persist for years. Nausea, vomiting, and fatigue are also common side effects with metronidazole.
Although ciprofloxacin is now widely used for the treatment of Crohn’s disease fistulas, there are only anecdotal reports of response and closure. In these small studies, approximately 70% of patients responded to 1 to 1.5 g of cipro when used for up to 1 year. As with metronidazole, exacerbations were seen after the ciprofloxacin was withdrawn. The combination of ciprofloxacin and metronidazole was reported in an uncontrolled study to show approximately 85% response rate, with healing seen in slightly over 20% of the patients.
In summary, although lacking controlled data, current clinical experience, including my own, indicates that antibiotics are effective in the treatment of perianal Crohn’s disease and can be used for long periods of time. The dose of metronidazole should be maintained at < 1 g daily to prevent neuropathy. For patients who do not respond to either ciprofloxacin or metronidazole, the combination is indicated. Antibiotics are not likely to close a fistula but can be used to control this complication while awaiting more potent immunomodulatory therapy.
6-Mercaptopurine/Azathioprine
The first control trial designed to evaluate for fistula response in patients with Crohn’s disease was a randomized double blind placebo controlled trial studying 6-mercaptopurine (6-MP) versus placebo. This was a 2-year study in which patients were crossed over after 1 year. In addition to demonstrating clinical response and steroid sparing, 9 of 29 patients (31%) who received 6-MP closed their fistula. Only 1 of the 17 (6%) placebo patients showed closure. Additional healing was observed in 34% with 6-MP as compared with 18% with placebo. Although showing a strong trend to healing, the numbers of the study were not large enough to show statistical significance.
Subsequent data looking at this problem in an uncontrolled manner showed complete closure in 39% of patients with a response in another 26%. Of major importance was the slow response to this medication, which took 2 to 4 months to be effective. Of the 13 patients in this open-label study whose fistula closed completely, the fistula continued to be closed in 6 of the 13. This is contrasted with seven patients who stopped the drug and only two whose fistula remained closed and five whose fistula reopened. A subsequent meta-analysis showed that 54% (22 of 41) patients treated with 6-MP responded compared with only 6 of 29 (21%) treated with placebo.
The pooled odds ratio favoring fistula healing was 4.44. In the gastrointestinal (gastrointestinal) community the dosing and administration of 6-MP or azathioprine (AZA) has significant variation. Control trials used AZA in doses of 2 to 3 mg/kg daily, whereas 6-MP was used at a dose of 1.5 mg daily. My experience has suggested that < 1.5 mg/kg of 6-MP is required for a therapeutic response, and the mean dose that is needed is 75 mg daily. There is some synergistic effect when using 5-aminosalicylates with 6-MP, and this should be taken into account in looking at the dose of the immuno-suppressive.
There is no question but that 6-MP/AZA is an effective agent both in closing and maintaining closure of fistulas. These agents are quite safe long term with no definite evidence of the development of neoplasia or superinfections. Toxicity is mainly limited to allergic reactions, including rash, fever, pancreatitis, and to those patients who are thiopurine methyl transferase deficient. In this group lower doses must be used to prevent the development of leukopenia.
In summary, 6-MP/AZA should be instituted if fistulas do not close spontaneously or do not close with antibiotics. The presence of multiple fistulas is also an urgent indication for the institution of 6-MP/AZA. It is difficult to predict which patients are going to go on to a chronic course and my clinical experience suggests that 6-MP/AZA can alter the natural history of these fistulas. I have seen maintenance of closure for 20 plus years in many patients who continue to take these agents.
Methotrexate
Although there have been several controlled and uncontrolled studies evaluating the efficacy of methotrexate in the management of Crohn’s disease, none of them were designed to look at the response of fistula. In a small recently published study of 37 courses of methotrexate in Crohn’s disease patients, 4 of 16 (25%) had complete closure of fistula, whereas 5 of the 16 (31%) had partial response. The overall rate of fistula response was, therefore, 56%. Several of these patients had not responded to prior treatment with cyclosporin and/or 6-MP.
With the limited available data it would appear that 6-MP/AZA would be the first choice in the management of perianal fistulas. However, if patients are allergic or fail to respond then methotrexate in a dose of 25 mg administered intramuscularly on a weekly basis should be instituted. The toxicity of methotrexate is well known, including gastrointestinal upset, interstitial pneumonitis, and possible hepatic fibrosis. This medication is also contraindicated in those patients attempting to conceive.
Cyclosporin
There are currently no control trials demonstrating that cyclosporin is an effective therapy for Crohn’s disease perianal fistula. Multiple uncontrolled studies showed an overall clinical response rate of around 80%. When cyclosporin has been administered in low doses (s 5 mg/kg daily), it is ineffective. However, administration of the drug intravenously in a dose of 4 mg/kg has shown efficacy both in clinical response and in the management of fistula. This intravenous (IV) dose is comparable to an oral dose of 8 mg/kg. In a personal series of 16 patients, 14 (88%) responded to this regimen.
Complete closure of fistula was observed in 44% and moderate improvement in about 44%. The mean time to respond was slightly over 7 days. A recent study looking at IV cyclosporin in ulcerative colitis showed that a dose of 2 mg/kg was equally effective to 4 mg/kg, but this has not been studied in Crohn’s disease patients. We have usually tried to obtain monoclonal levels of cyclosporin between 300 to 500 ng/mL during the 7 to 10 day hospitalization with a trough level of approximately 300 ng/mL after discharge. We have seen relapse in 36% after discharge, but 64% maintain their response and allowed us to discontinue steroids, which may facilitate healing of fistulas. In order to avoid cyclosporin toxicity, experience is essential in dosing regimens.
Access to a laboratory that can give same day results is quite important, and we usually recommend referral to a tertiary center if experience is lacking in the gastroenterologist’s clinical practice. The main adverse effects of cyclosporin are abnormal renal function, hypertension, headache, and gingival hyperplasia. When used in combination with steroids Pneumocystis carinii pneumonia has been observed and prophylaxis should be instituted with Septra. After the steroid dose is lowered or discontinued, the Septra can also be stopped. We advise an examination under anesthesia prior to instituting cyclosporin therapy to ensure all abscesses have been drained.
Tacrolimus (FK506)
Several uncontrolled case series have been published suggesting that tacrolimus may be beneficial in the treatment of perianal fistulas. A small placebo controlled trial evaluated 43 patients with perianal fistula and compared tacrolimas with placebo. The dose of the oral tacrolimus was 0.20 mg/kg/d. Closure of at least 50% of fistulas at 2 visits 4 weeks apart occurred in 8% of placebo patients compared with 43% of patients treated with tacrolimus.
On the other hand, closure of all fistulas occurred in only 10% of tacrolimus-treated patients compared with 8% with placebo. Adverse events occurred, the most important being an increased serum creatinine level. This was the major toxicity but others were observed, including pares-thesias, tremors, and leg cramps. Nephrotoxicity was observed in 38% of patients compared with 0% with placebo. At the current time tacrolimus cannot be recommended for the management of perianal fistulas.
Infliximab
Infliximab, a chimeric monoclonal antibody, has shown efficacy not only in the treatment of active Crohn’s disease but also in the management of perianal fistula. An initial randomized double-blind placebo controlled trial was carried out on 94 patients who received infusions at weeks 0, 2, and 6. The primary endpoint was the reduction in the number of draining fistulas of a 50% on 2 visits 4 weeks apart. Fistulas were considered to be closed if drainage had stopped and could not be observed with gentle finger compression. At a 5 mg/kg dose the primary end-point was achieved in 62% of patients, with complete closure of fistula in 46% of patients compared with 13% receiving placebo.
The 5 mg/kg dose was the most effective and closed 55% of all fistula. The vast majority of fistulas closed after the second infusion. The main complication was the occasional development of a perianal abscess, which may have been related to the very rapid closure of the cutaneous end of the fistula tract before the rest of the fistula had healed. It is my habit to use concurrent antibiotics when trying to close perianal fistula with infliximab, as well as trying to discontinue steroids.
There has been an abstract report of a second control trial studying 306 patients who were initially treated with 3 doses of 5 mg/kg at 0,2, and 6 weeks. Patients who responded were then randomized into maintenance infliximab every 8 weeks compared with placebo. The primary endpoint was the “time to loss of response” through week 54. The median time to lose response in infliximab patients was > 40 weeks, whereas with placebo it was 14 weeks. At week 54,39% of patients who were receiving maintenance had complete closure of all draining fistula compared with only 19% of those receiving placebo. Despite this response rate, some patients treated with infliximab for fistulas may still ultimately require surgical intervention.
Our personal experience is to have a patient undergo an examination under anesthesia prior to instituting infliximab to ensure that all abscesses are being drained. The placement of setons may help in maintaining drainage while waiting for the infliximab to be effective. The development of human antichimeric antibodies in high percentages of patients receiving infliximab indicates that a concurrent immunomodulatory agent (such as 6-MP) be given so that the degree and duration of response may be enhanced. Infliximab has shown to increase the overall rate of infections, and serious complications, such as tuberculosis, histoplasmosis, coccidiomycosis, and others, have been observed following administration of this agent.
Jennifer is a clinical pharmacologist. She helps doctors of all medical specialties choose medications. Jennifer consults about drug dosage, interactions, and side effects.