Infliximab in the treatment of medically refractory indeterminate colitis

Authors


Dr E. A. Vasiliauskas or Dr K. A. Papadakis, Inflammatory Bowel Disease Center, Cedars-Sinai Medical Center, 8631 W. Third Street, Suite 430E, Los Angeles, CA 90048, USA.
E-mail: VasiliauskasE@cshs.org; PapadakisK@cshs.org

Summary

Aim : To examine the outcome of infliximab intervention in refractory indeterminate colitis.

Methods : Twenty patients with severe, medically refractory indeterminate colitis were treated with infliximab. All patients initially received infliximab, 5 mg/kg, intravenously and, in some patients, the dose was subsequently increased to 10 mg/kg. The number of infusions ranged from one to 16 per patient. Indeterminate colitis was defined as colitis that could not be classified with certainty as Crohn's disease or ulcerative colitis based on traditional clinical, endoscopic and histopathological criteria. The clinical response to infliximab was classified as complete response, partial response or non-response.

Results : Fourteen of the 20 patients (70%) showed a complete response to infliximab treatment, two showed a partial response and four showed no response. The four non-responders underwent colectomy with ileal pouch-anal anastomosis. The resected colon specimen was consistent with ulcerative colitis in all four cases, although two were subsequently re-classified as Crohn's disease. Eight additional patients were subsequently re-classified as having Crohn's disease on longer follow-up evaluation, whilst eight continued to have features of indeterminate colitis. The response rate to infliximab treatment was similar in both groups.

Conclusions : Infliximab is effective in approximately two-thirds of patients with indeterminate colitis, and thus may be considered for patients with refractory disease prior to colectomy. The follow-up time afforded by infliximab treatment may allow for more accurate classification of the disease in a significant proportion of patients whose colitis has indeterminate features at initial presentation.

Introduction

Convincing clinical, radiographic and pathological criteria for distinguishing between ulcerative colitis and Crohn's disease of the colon were not reported until the 1960s.1–3 Well-described macroscopic and microscopic histopathological criteria allow for most patients with inflammatory bowel disease to be clearly classified as having Crohn's disease or ulcerative colitis following the examination of colectomy specimens. Nevertheless, as the repertoire of colonic response to injury is limited, colitis due to Crohn's disease and ulcerative colitis may share common histopathological features and the discriminating characteristics may not be well defined, especially in the setting of acute, severe colitis. Based on an examination of the surgically excised colon specimen alone, a clear distinction between the two disorders cannot be made with confidence in 4–29% of cases.1, 3–10 The term ‘indeterminate colitis’ was originally used to describe this type of colitis in surgical specimens that could not be classified with certainty as ulcerative colitis or Crohn's disease, and was therefore proposed as a provisional classification prior to establishing a definitive diagnosis.6

Subsequently, the definition of indeterminate colitis has often been extended to the pre-colectomy evaluation of patients with chronic idiopathic colitis, often with somewhat arbitrary definitions. It is still debated whether indeterminate colitis represents simply a problem of classification, a distinct clinical entity from ulcerative colitis and Crohn's disease, incomplete evaluation or early/incomplete disease expression at the time of evaluation. Unfortunately, evidence-based medical treatment strategies for indeterminate colitis are lacking, as randomized controlled trials have included patients with a well-characterized type of inflammatory bowel disease, e.g. Crohn's disease or ulcerative colitis. As many patients with indeterminate colitis will eventually be found to have either ulcerative colitis or Crohn's disease, any therapy that is effective in both diseases may potentially prove to be useful for the treatment of indeterminate colitis as well.

Infliximab, a chimeric monoclonal antibody against human tumour necrosis factor-α, has been successfully used for the treatment of inflammatory and fistulizing Crohn's disease.11, 12 The impressive positive results of infliximab in randomized controlled trials for Crohn's disease have also been observed in clinical practice and reported in retrospective studies.13–16 In addition, recent data from the ACCENT I and II trials have suggested that repeated infliximab infusions can maintain remission in both inflammatory and fistulizing Crohn's disease.17, 18 The role of infliximab for the treatment of ulcerative colitis remains controversial. In a small randomized trial, 50% of patients with steroid-resistant ulcerative colitis responded to treatment with infliximab, compared with none of those who were treated with placebo.19 Several small observational series have suggested a benefit of infliximab in the treatment of medically refractory ulcerative colitis.20–22 Interestingly, one of these studies found that patients with steroid-responsive disease were more likely to respond to infliximab than those with steroid-resistant disease.22 Another randomized trial, however, showed no clinical benefit of infliximab in the treatment of steroid-resistant ulcerative colitis compared with placebo.23

The use of infliximab for the treatment of indeterminate colitis has not been reported previously. In the current study, we evaluated the efficacy of infliximab therapy in patients with indeterminate colitis who were referred to the Inflammatory Bowel Disease Center, Cedars-Sinai Medical Center, Los Angeles, CA, USA for treatment-resistant disease.

Materials and methods

Patients and definitions

Indeterminate colitis was defined as colitis that, despite extensive work-up, could not be classified with certainty as either ulcerative colitis or Crohn's disease based on established endoscopic and histopathological criteria.24, 25 Briefly, patients with chronic idiopathic colitis were classified as having indeterminate colitis if they did not have small-bowel involvement on ileo-colonoscopy or small-bowel follow-through study, endoscopy was inconclusive for either Crohn's disease or ulcerative colitis, and the biopsy specimens revealed active and patchy transmucosal chronic inflammation in the absence of granuloma.26

On the basis of this definition, we identified 20 patients with indeterminate colitis, referred to the Inflammatory Bowel Disease Center for steroid-resistant or steroid-dependent colitis, who received infliximab treatment. The medical records of all patients were reviewed and the following data were obtained: demographic information, disease duration, concurrent medications, number of infliximab infusions, clinical response to treatment, inflammatory markers (erythrocyte sedimentation rate and C-reactive protein when available) and surgical/pathological diagnosis on follow-up evaluation. The Institutional Review Committee of the Cedars-Sinai Medical Center approved the review of the patients' medical records for the purpose of this study (Ex-1235).

Infliximab was initiated at a dose of 5 mg/kg of body weight as a 2-h intravenous infusion. The number and interval of infusions received by each patient were determined by the treating physician based on the extent and duration of the clinical response. Some patients who had incomplete or short-duration responses on initial or repeat infusions were subsequently treated with a higher dose (10 mg/kg) of infliximab. Because of the indeterminate features of the colitis and an inability to reassure patients that they had ulcerative colitis, they were offered infliximab as a ‘last’ effort to control their disease before surgical intervention was undertaken. All patients were considered to have failed maximal medical therapy, including 5-aminosalicylates, corticosteroids and immunomodulatory agents. Patients were initially continued on concomitant medications. Patients who were receiving corticosteroids were gradually tapered according to their response to infliximab treatment. Patients were monitored during and after the infusion for adverse events.

The determination of serum immune markers, perinuclear antineutrophil cytoplasmic antibodies (pANCA), Saccharomyces cerevisiae antibodies (ASCA) and outer-membrane porin C (OmpC) antibodies was performed as part of the diagnostic work-up of colitis by Prometheus Laboratories (San Diego, CA, USA), as described previously.27–29 Inflammatory markers, such as the erythrocyte sedimentation rate and C-reactive protein, were also analysed before and after infliximab treatment when available.

Response to infliximab treatment

Due to the retrospective nature of this study, a simplified clinical response evaluation was employed. Clinical response was classified into three categories: complete response, partial response and non-response. A complete response was defined as the cessation of colitis-related symptoms, such as abdominal cramping, diarrhoea and bleeding. A partial response was defined as a reduction, but not resolution, of abdominal cramping, diarrhoea and/or bleeding. All other outcomes were defined as a non-response. Clinical response and delayed adverse events were assessed in regular follow-up clinic visits and by patient-initiated telephone calls.

Results

Patient characteristics

The characteristics of the 20 patients with indeterminate colitis who were treated with infliximab are shown in Table 1. Ten were male and 10 were female, with an age range of 9–45 years. The median duration of the disease was 3 years (range, 0.5–11 years). The indication for infliximab treatment is also shown in Table 1. All patients had failed treatment with 5-aminosalicylates, 15 were steroid dependent or refractory, 14 were refractory or intolerant to thiopurines and three had failed treatment with ciclosporin. At the initiation of infliximab therapy, 11 of 20 patients (55%) were receiving steroids. Eleven (55%) were receiving immunomodulator therapy with mercaptopurine, azathioprine or tioguanine (thioguanine). Six of the 20 patients (30%) were being treated with a combination of steroids and thiopurines. Thirteen of the 20 patients (65%) were on 5-aminosalicylates. One additional patient was being treated with ciclosporin. The analysis of serum markers showed that 14 of the 20 patients (70%) were pANCA positive, three (15%) were ASCA positive (all also expressed pANCA), four (20%) were OmpC positive (three were also pANCA positive) and four (20%) were negative for all serological markers.

Table 1.  Baseline characteristics of patients with indeterminate colitis treated with infliximab
Characteristic
  1. 5-ASA, 5-aminosalicylates; ASCA, Saccharomyces cerevisiae antibodies; AZA, azathioprine; MP, mercaptopurine; OmpC, outer-membrane porin C antibodies; pANCA, perinuclear antineutrophil cytoplasmic antibodies; TG, tioguanine.

Age (years) (mean, range)22.2, 9–45
Sex (female/male)10/10
Disease duration (years) (median, range)3, 0.5–11
Indication for infliximab therapy (n)
 5-ASA refractory20
 Steroid-dependent/refractory15
 Thiopurine refractory/intolerant14
 Ciclosporin failure 3
Concurrent medications (n)
 Steroids11
 MP/AZA/TG11
 Both 6
 5-ASA13
 Ciclosporin 1
Serologies (n)
 pANCA14
 ASCA 3
 OmpC 4
 Negative 4

Clinical response and adverse events

The patients received a mean of 5.4 and a median of three (range, 1–16) infliximab infusions (Figure 1). The duration of treatment ranged from one infusion to 3 years. All patients initially received infliximab at 5 mg/kg. Fourteen (70%) were classified as complete responders to infliximab treatment and two (10%) as partial responders. Four (20%) of the 20 treated patients were non-responders and underwent ileal pouch-anal anastomosis. Because of attenuation of the response to the 5 mg/kg dose, four of the 14 responders were subsequently treated with infliximab, 10 mg/kg, with re-achievement of response. Inflammatory parameters (erythrocyte sedimentation rate and/or C-reactive protein) were available pre- and post-initial infliximab infusion in 15 of the 20 patients. The mean pre-treatment erythrocyte sedimentation rate was 36 mm/h (range, 6–96 mm/h), and decreased to 20 mm/h (range, 1–111 mm/h) following infliximab infusion (Figure 2). Two of three non-responders showed an elevation of the erythrocyte sedimentation rate post-infusion, and all but one patient who clinically responded showed a reduction in the erythrocyte sedimentation rate post-infusion. Similar changes were observed for C-reactive protein. The mean pre-treatment C-reactive protein was 4.78 mg/dL, which decreased to 1.02 mg/dL following infliximab infusion. Although the number of patients was small for the correlation of serological markers with clinical response to infliximab, all five patients (100%) who were pANCA negative showed a complete response to infliximab therapy, compared with eight of 14 (57%) who were pANCA positive. One pANCA-positive patient was a partial responder. Seven of the 11 patients on concurrent steroid treatment responded to treatment with infliximab. Four of these seven patients gradually tapered and eventually discontinued steroid treatment, one failed and two had incomplete follow-up evaluations. Six of the 11 patients receiving concomitant therapy with thiopurines discontinued treatment. Three were subsequently treated with methotrexate. The patient who was receiving ciclosporin at the commencement of infliximab treatment also discontinued the medication.

Figure 1.

Number of infliximab infusions received by patients with indeterminate colitis. Four patients received a total of 20 infusions at 10 mg/kg.

Figure 2.

Determination of erythrocyte sedimentation rate (ESR) pre- and post-infliximab treatment in responders and non-responders. Bars show mean values.

The infliximab infusions were tolerated well in all but three patients. One developed dizziness and fever following the infusion. Another patient developed a serum sickness-like reaction, 1 week following his third infliximab infusion. One additional patient developed what was initially felt to be a lupus-like drug reaction, but which over time has proven to be true lupus. All three patients discontinued infliximab therapy and two experienced exacerbation of their disease.

Patient follow-up

During follow-up evaluation, 10 patients were re-classified as having Crohn's disease and two as having ulcerative colitis. In the remaining eight patients, the colitis remained ‘indeterminate’ (Figure 3). The diagnosis of Crohn's disease was based on the development of fistulae and/or small-bowel inflammation or characteristic upper tract Crohn's disease or granulomas found on follow-up biopsies. Eight (80%) of the 10 patients who were later re-classified as having Crohn's disease responded to infliximab treatment vs. none of the two patients found to have ulcerative colitis and six (75%) of the eight who continued to have features of indeterminate colitis. The remaining two patients with indeterminate colitis were partial responders.

Figure 3.

Outcome of patients with indeterminate colitis treated with infliximab. *All colon resection specimens were consistent with ulcerative colitis. **During follow-up evaluation, one patient underwent a diverting ileostomy with improvement of colitis and one underwent a subtotal colectomy with the surgical specimen consistent with Crohn's disease. CD, Crohn's disease; IC, indeterminate colitis; IPAA, ileal pouch-anal anastomosis; UC, ulcerative colitis.

Overall, six (30%) of the 20 patients underwent surgery during follow-up evaluation. Four underwent an ileal pouch-anal anastomosis and, in all cases, the colectomy specimen was consistent with ulcerative colitis. However, two of these patients were later re-classified as having Crohn's disease based on the development of inflammation and ulcers above the pouch. One patient who had initially responded to infliximab treatment and was later found to have Crohn's disease underwent a diverting ileostomy with improvement of his colitis. One additional patient, who initially responded to infliximab, underwent a subtotal colectomy and the surgical specimen was consistent with Crohn's disease.

Discussion

Our study shows that infliximab is effective for the treatment of medically refractory indeterminate colitis. All patients in our series had severe colitis refractory to medical interventions. Most remained symptomatic despite corticosteroids, several were steroid dependent and many failed to respond adequately to therapy with the immunomodulators mercaptopurine/azathioprine or tioguanine. Despite their disease severity, 14 of the 20 patients (70%) responded to infliximab treatment. Despite the fact that 10 of the 20 patients (50%) were eventually re-classified as having Crohn's disease, the efficacy of infliximab treatment was similar in those patients to that in patients with colitis that still had indeterminate features in follow-up evaluation. When patients with refractory disease can be definitively classified as having ulcerative colitis, colectomy with ileal pouch-anal anastomosis is considered to be an effective long-term surgical treatment. Patients with Crohn's disease, however, have a high risk of complications following ileal pouch-anal anastomosis and many require long-term immunomodulation.5, 30 Thus, patients with established or suspected Crohn's disease are often treated more aggressively with medical interventions, as surgical options cannot offer the same likelihood of symptom-free or medication-free outcomes as in ulcerative colitis. Similarly, patients with indeterminate colitis are more likely to undergo surgical resection for medically refractory disease and are at higher risk for the development of pouch-related complications following ileal pouch-anal anastomosis.9, 31–33 In our study, half of the patients who responded to infliximab treatment were eventually re-classified as having Crohn's disease, although longer follow-up may be necessary to determine whether, in the remaining infliximab responders, colitis will remain ‘indeterminate’. It is interesting that the response to infliximab in many of these patients and thus the longer duration of colonic disease may have enabled the disease to be fully manifested as Crohn's disease, or alternatively more aggressive evaluations have increased the diagnostic yield for Crohn's disease.

It is also interesting that the resected colectomy specimens of all patients who failed to respond to infliximab and underwent ileal pouch-anal anastomosis were consistent with ulcerative colitis, but two were subsequently re-classified as having Crohn's disease during longer term follow-up evaluation. The results of this study are consistent with the generally lower response rate of patients with ulcerative colitis to infliximab treatment in randomized trials.19, 23 It is also conceivable that ulcerative colitis and ulcerative colitis-like Crohn's disease may share common immunological and genetic aspects, which could render them less responsive to tumour necrosis factor blockade. Certainly, the identification of such immunological and genetic aspects of these diseases, which may predict a response to infliximab, would be of significant clinical importance.

The number of indeterminate colitis patients treated with infliximab was too small to analyse factors that may have predicted a response to treatment. In Crohn's disease, Parsi et al. have shown that non-smoking and concurrent immunosuppressive treatment are associated with higher rates and a longer duration of response to infliximab in patients with inflammatory Crohn's disease. Non-smoking was associated with a longer duration but not rate of response to infliximab in fistulizing Crohn's disease.34 Vermeire et al. have shown that young age, colonic involvement and concomitant immunosuppressive treatment, but not smoking, are independent predictors of response to infliximab.35

In our study, we could not analyse the effect of these parameters in the response to infliximab. Despite the small number of patients, the absence of pANCA serum marker, however, predicted a 100% response rate to infliximab, compared with an approximately 60% response rate of patients who were pANCA positive. This observation is consistent with the lower response rate to infliximab of Crohn's disease patients who are pANCA positive, carry a specific lymphotoxin haplotype and have a phenotype of ulcerative colitis-like Crohn's disease.36

The central role of tumour necrosis factor in mucosal inflammation has been clearly shown in animal models of ileitis or colitis.37–40 In most animal models of inflammatory bowel disease, anti-tumour necrosis factor antibody administration ameliorates the disease, and this observation has been translated into the effective treatment of patients with Crohn's disease. The mechanism by which infliximab treatment exerts its therapeutic benefit in Crohn's disease or indeterminate colitis is likely to be multifactorial. Recent studies have shown that infliximab induces apoptosis of human monocytes and lamina propria T cells, which may explain the prolonged disease remission observed in treated patients.41–44 Neutralization of tumour necrosis factor-α may reverse its pro-inflammatory effects, such as the induction of chemokines and adhesion molecules, maturation of dendritic cells, augmentation of interferon-γ production by lamina propria T cells and induction of intestinal epithelial cell apoptosis.45

In summary, our study shows that infliximab is an effective therapy for medically refractory indeterminate colitis, and may be considered in clinical practice for the treatment of patients who do not respond to conventional medical therapy and are contemplating surgery. In approximately one-half of indeterminate colitis cases treated with infliximab, longer follow-up evaluation also enabled the treating physician to re-classify their disease.

Ancillary