A Crohn's Disease Patient Who Does Not Respond to Infliximab: What Is Next?

  • Nitsan Maharshak
    Department of Medicine, Division of Gastroenterology and Hepatology, Inflammatory Bowel Disease Center, University of North Carolina School of Medicine, Chapel Hill, North Carolina
    Search for articles by this author
  • Scott E. Plevy
    Reprint requests Address requests for reprints to: Scott E. Plevy, MD, CB7032, 7341c MBRB, 103 Mason Farm Road, Chapel Hill, North Carolina 27599. fax: (919) 843-2585
    Department of Medicine, Division of Gastroenterology and Hepatology, Inflammatory Bowel Disease Center, University of North Carolina School of Medicine, Chapel Hill, North Carolina
    Search for articles by this author

      Abbreviations used in this paper:

      ATI (antibodies to infliximab), CD (Crohn’s disease), CDAI (Crohn’s Disease Activity Index), CMV (cytomegalovirus), CRP (C-reactive protein), PML (progressive multifocal leukoencephalopathy), TNF (tumor necrosis factor)

      Clinical Scenario

      A 68-year-old nonsmoking woman with a history of Crohn's disease (CD) for 22 years presents with abdominal pain and diarrhea. Her course was characterized by chronic mild symptoms interspersed with significant symptomatic flares with bloody diarrhea, abdominal pain, and weight loss requiring corticosteroids approximately 3 times yearly. She would respond to prednisone, and when tapered, she was maintained on mesalamine with intermittent courses of metronidazole for milder symptoms. Eight years earlier, 6-mercaptopurine was added, but she developed pancreatitis. Endoscopically, she has had documented segmental inflammatory involvement of the rectum, ascending colon, and cecum. Eighteen months earlier she experienced a significant disease exacerbation. Stool cultures and Clostridium difficile toxin A and B assays were negative. Computed tomography enterography demonstrated segmental involvement of the colon without involvement of the small bowel. Colonoscopy and biopsies demonstrated segmental moderate to severely active colonic CD with deep ulcerations. Infliximab was initiated at 5 mg/kg intravenously with induction at weeks 0, 2 and 6, followed by every 8-week maintenance. She was in clinical remission for 7 months. During the past 5 months, she became symptomatic increasingly earlier between infliximab infusions. Three months ago when she was significantly symptomatic 4 weeks after the previous infusion, the infliximab dose was increased to 10 mg/kg. However, when first presenting to our office, she had received infliximab 2 weeks prior and had not symptomatically improved. She is currently having 6 bowel movements daily, 1 nocturnal bowel movement, most of which contain blood. She has significant lower quadrant abdominal pain and has lost 10 pounds during the past 4 months. Laboratory evaluation is remarkable for hemoglobin of 10.8 g/dL, platelets 566,000 × 106/L, C-reactive protein (CRP) 32 mg/dL, and albumin 3.4 g/dL.

      The Problem

      This patient presents with a common situation that clinicians encounter: in a CD patient whose condition exacerbates while on infliximab, how do you evaluate symptoms and how do you appropriately treat? It is important to recognize that the differential diagnosis of nonresponse/loss of response to infliximab is vast (Table 1), and many causes might in fact have little to do with infliximab efficacy. Major etiologies to consider include infectious agents such as Clostridium difficile and cytomegalovirus (CMV), postsurgical states (such as bile salt diarrhea and short bowel syndrome), irritable bowel syndrome, fibrostenosis requiring surgery, and CD disease exacerbation. Tailoring a treatment for a refractory CD patient is always a challenge, specifically when side effects prevent the use of thiopurines (pancreatitis in this case) and there is a loss of response to one of the mainstays of treatment, the anti–tumor necrosis factor (TNF) antibody infliximab.
      Table 1Basic Differential Diagnosis of Nonresponse/Loss of Response in CD Patients
      Disease complications
       New disease location
      Bowel resection related
       Bile salt diarrhea
       Small bowel bacterial overgrowth
       Short bowel syndrome
       Bacterial: C. difficile, Shigella, Salmonella, Yersinia, Campylobacter, etc
       Viral: CMV
       Fungi: ameba (history of traveling to endemic area)
       Irritable bowel syndrome
       Celiac disease


      Along with ruling out alternative etiologies, the most important aspect of a diagnostic evaluation is to verify an active inflammatory process in the gut. Confirmation of active inflammation might be supported by serum inflammatory markers such as erythrocyte sedimentation rate and CRP, but the key evaluation is imaging (computed tomography/magnetic resonance enterography) and/or endoscopy. The critical importance of obtaining objective evidence for inflammation is perhaps best exemplified in the SONIC trial of infliximab and/or azathioprine for biological and immunomodulator naive CD, which included patients with moderate to severe active CD (Crohn's Disease Activity Index [CDAI] >220). Among patients entered into this multicenter international study, 18.3% had no mucosal lesions on colonoscopy and a normal CRP. Only 33%–40% of patients with no evidence of active inflammation treated with infliximab (with or without azathioprine) achieved steroid-free clinical remission at 26 weeks of treatment compared with 57%–69% of patients who had mucosal lesions and CRP ≥0.8 mg/dL. When CD activity is documented by objective criteria, subsequent management will be further guided by whether the exacerbation is due to a primary nonresponse or secondary loss of response to infliximab.

       Primary Nonresponse

      There are no uniformly accepted definitions for defining a nonresponse to induction with infliximab; however, from the early induction trial in CD and from SONIC, the majority of patients will respond (defined in studies as a CDAI reduction of 70–100 points) after the first 2 induction infusions (weeks 0 and 2) by week 6. In clinical practice because of limited options, a third induction dose, possibly at 10 mg/kg, is sometimes contemplated before defining failure. With clinical trial definitions, 35%–40% of patients with moderate to severe CD unresponsive to conventional therapy will be categorized as primary nonresponders to infliximab. Mechanisms for primary nonresponse are obscure. Some of the suggested mechanisms of true primary nonresponse include non–TNF-dependent inflammation, smoking, and polymorphisms of the immunoglobulin G Fc receptor IIIa, which might result in rapid clearance of infliximab.

       Secondary Loss of Response

      From the large clinical trial program (ACCENT), of CD patients who initially respond to infliximab, up to 40% will lose response within the first year of treatment. If non–CD-related causes are excluded and active inflammation is present, 2 major etiologies are to be considered: antibodies to infliximab (ATI), which might appear in up to 16% of patients receiving scheduled infliximab treatment (without concomitant immunomodulators), and low serum infliximab levels, which might be a result of ATI but might have ATI-independent etiologies, likely a result of rapid antibody clearance. Patients receiving episodic infliximab treatment develop ATIs in up to 60% of cases, and therefore this treatment approach is not recommended. Most evidence supports the role of ATIs in lowering infliximab serum levels and decreasing efficacy. Serum levels of ATIs also play an independent role, because some patients with low levels of ATIs still show good clinical response to infliximab. Clinically, patients who develop ATIs are more prone to develop acute and delayed infusion-related reactions.

      Management Strategies and Supporting Evidence

      High-level evidence supporting the management of infliximab failure is scarce. Nevertheless, treatment options exist and include infliximab dose intensification (by decreasing intervals or increasing dosage). The ability to dose intensify might partly explain the discrepancy between lower rates of response to infliximab at 1 year reported in clinical trials compared with results documented in clinical practice. In addition, in clinical practice, response to treatment is usually based on the physician's global assessment, which is less stringent than the CDAI measurements used in clinical trials. In a large single center clinical practice study, Schnitzler et al reported a 63.4% response rate to infliximab after a median of more than 4 years of treatment. For 19.7% of patients a reduction in the interval between infliximab infusions was required, and 26.3% of patients needed dose increase to maintain response. Interestingly, 71.5% of patients who increased dose were able to return to 5 mg/kg, whereas only 28.7% with a shortened interval could return to every 8-week dosing. Although mechanisms for these 2 distinct phenomena were not further elucidated, one can speculate that in the former group, an inciting environmental event precipitated a flare that then can be managed short-term by dose intensification, whereas in the latter group, ATI and/or decreased serum infliximab levels mediated loss of response at the end of the infusion interval. To illustrate clinically relevant examples of these 2 phenomena, a patient in extended remission on stable dose infliximab develops an acute exacerbation after a course of antibiotics (C. difficile negative); transient dose escalation to 10 mg/kg with de-escalation once remission is again achieved has been a successful strategy in our practice. Other investigators might consider reinduction (infusions at weeks 0, 2, and 6) in similar circumstances. The second clinically appreciated phenomenon is a patient who develops symptoms 6–7 weeks after an infliximab maintenance infusion. In these patients, supported by the Schnitzler study, our experience is that interval shortening or dose escalation might be necessary long-term. As discussed below, determination of serum infliximab levels and ATI might help determine the appropriate rescue strategy.
      Another strategy to manage loss of response is to switch to another TNF inhibitor, adalimumab or certolizumab pegol. The GAIN study was a randomized, placebo-controlled trial of adalimumab in patients with moderate to severe CD (220 <CDAI <450) who either lost response or were intolerant to infliximab (the primary end point was remission at week 4 defined as CDAI <150 points after adalimumab 160 mg subcutaneously at week 0 and 80 mg at week 2). In the adalimumab group, 52% of patients achieved response (34% placebo) and 21% remission (7% placebo). An open-label trial of certolizumab pegol in patients who lost response or had an adverse event to infliximab (WELCOME) reached similar conclusions.
      A prevalent mechanism underlying loss of response to infliximab is related to the development of ATIs, with consequential reduction of serum infliximab levels. A recent small retrospective study assessed the clinical utility of ATIs and serum infliximab levels for clinical decision making when secondary loss of response is encountered. In this study, 92% of patients who developed an ATI and were switched to another anti-TNF agent had a clinical response. Increasing infliximab dosage resulted in response in only 17% of these patients. In patients with subtherapeutic concentrations (no ATI detected), infliximab dose escalation was associated with a significantly increased clinical response compared with changing to another anti-TNF. Practically, on the basis of limited data, either a 4-week postinfusion infliximab concentration of >12 μg/mL or a detectable trough infliximab concentration is associated with improved clinical remission, lower CRP, and endoscopic healing. However, therapeutic serum infliximab levels have been defined in these few retrospective analyses through correlation with clinical, biochemical (CRP), and endoscopic response parameters, and it is therefore impossible to recommend a target serum infliximab level on the basis of prevailing evidence.
      In CD patients who are naive to immunosuppression or TNF inhibitors, combination therapy with immunomodulators and infliximab decreased immunogenicity (ATIs) and increased long-term efficacy with scheduled maintenance treatment. The caveat for combination therapy has been the fear of increased episodes of rare but potentially serious events such as serious and/or opportunistic infections and malignancies. Although improved efficacy of combination therapy is clear from the SONIC study, the current literature does not support increased frequency of serious adverse events with combination therapies compared with TNF inhibitors or azathioprine/6-mercaptopurine alone. The impact of combination therapy versus monotherapy on rare events such as hepatosplenic lymphomas will be difficult to determine with precision because of the rarity of these events.
      In clinical practice when a patient loses response to TNF inhibitor monotherapy, addition of an immunomodulator (thiopurines or methotrexate) is often considered on the basis of data showing that immunomodulators increase serum levels of TNF inhibitors. However, this strategy is not evidence-based, because there are no published studies addressing the impact of immunomodulator addition on serum infliximab levels or clinical response after initiation of infliximab. In our practice, although not supported by evidence, if a patient failed or lost response to TNF inhibitor monotherapy, subsequent combination of a second anti-TNF with an immunomodulator seems to be a prudent strategy to decrease antibody formation (especially important as one moves to a second TNF inhibitor because of the diminishing therapeutic options) and improve overall efficacy (perhaps through complementary mechanisms).
      CD patients who fail to respond to infliximab and immunosuppression might be good candidates for natalizumab, which is the only biological treatment approved by the Food and Drug Administration in CD that works through a mechanism other than TNF inhibition. Natalizumab is an immunoglobulin G4 humanized anti-α4 integrin that is approved in CD patients with active inflammation who failed immunosuppressive therapy and who have inadequate responses to anti-TNF agents. The major risk is progressive multifocal leukoencephalopathy (PML), which is a result of reactivation of the JC polyoma virus. The estimated risk of PML is 1:1000 after a year of treatment. Thus, natalizumab is not used in combination with immunomodulator drugs, and careful patient selection and follow-up are necessary. With the ongoing development of diagnostic tests, antibodies to JC virus detected by enzyme-linked immunosorbent assay might be used clinically in the future to screen patients and stratify PML risk. Those who are not infected (up to 50% of the population) are at low risk to develop PML. In the large maintenance clinical trial program of natalizumab in CD (ENACT-2), in patients who were previously infliximab exposed, remission rates were similar to the entire intention-to-treat population (50% vs 55%), supporting the efficacy of natalizumab in this population. Because labeling indications for natalizumab use are quite restrictive, in patients who do not respond or lose response to natalizumab, strategies such as dose intensification or addition of an immunomodulator are prohibited; hence consideration of alternative therapies such as surgery, enrollment in clinical trials, etc is necessary.

      Published Guidelines

      The World Congress of Gastroenterology recently published guidelines regarding the use of biological therapy for inflammatory bowel disease. Loss of response to infliximab, defined as a diminished or suboptimal response to infliximab, can be managed by shortening interval between dosing or increasing dose to 10 mg/kg. Measuring infliximab trough serum concentrations might help in the assessment of response to therapy. Switching to another anti-TNF is a viable strategy, but there is a lower chance of responding to a second TNF inhibitor. When anti-TNF agents fail, switching to natalizumab might be beneficial.
      The Canadian Association of Gastroenterology published clinical guidelines in 2009 that addressed pivotal questions of management of infliximab loss of response.
      In primary nonresponse to infliximab, additional doses of infliximab are not recommended. Switching to another TNF antagonist might be considered on a case-by-case basis. A caveat to the Canadian recommendations is that natalizumab is not approved for use in Canada and therefore was not considered in the treatment guidelines. For patients who have a partial response to infliximab induction, alternative strategies (including dose escalation or switching to another TNF antagonist) are options. In cases of secondary loss of response, infliximab dosing can be intensified by increasing dosage or decreasing intervals. An alternative option is switching to another TNF inhibitor. Strategies to reduce ATIs and prevent loss of response include scheduled infliximab maintenance, combination immunomodulator therapy, and infusion pretreatment with corticosteroids.

      Areas of Uncertainty

      Many questions remain about mechanisms and management of nonresponse and loss of response to infliximab, and most strategies discussed are currently supported only by observational data.

       Unanswered Questions

      • 1
        Etiologies for nonresponse and loss of response to infliximab in CD patients. We do not have a clear understanding of mechanisms of nonresponse to infliximab and, importantly, predictors for patients who will not respond to infliximab or other anti-TNF agents. Criteria such as short duration of disease, nonsmokers, elevated serum CRP levels, and higher serum infliximab levels have been associated with better response to TNF inhibition. Nevertheless, our ability to predict who does not respond or lose response to therapy is limited. Current tools, including serum infliximab levels and antibodies to infliximab, might help guide management decisions in patients who lose response, but it is important to reiterate that these strategies are based mostly on retrospective data. Figure 1 proposes a tentative algorithm for patients who do not respond or lose response to infliximab.
        Figure thumbnail gr1
        Figure 1A practical approach toward treatment of CD patients who are either primary nonresponders or experience secondary loss of response to infliximab. Most evidence for this flow chart is based on retrospective studies and expert opinions. Some suggestions are based on stronger clinical evidence (GAIN, ENACT 2, combination therapy in SONIC). TPN, total parenteral nutrition.
      • 2
        Does the benefit-risk ratio of decreased rates of loss of response to infliximab compared with the risk of opportunistic infections and malignancies favor combination with immunomodulators? In 2009, the American College of Gastroenterology recommended anti-TNF monotherapy because of risk-benefit concerns. Despite this recommendation, it is still entirely unclear whether combination therapy is associated with a significant risk. On the basis of the SONIC trial, combination therapy should be considered standard practice because of increased efficacy and decreased immunogenicity. This recommendation has been adopted by the World Congress of Gastroenterology specifically for the first 12 months of therapy in CD patients naive to thiopurines. Exceptions might include high-risk populations for complications, such as young male patients (lymphoma) or elderly patients with comorbidities (infections). In addition, it is still unclear whether combination therapy should be administered for a limited period of time or indefinitely.
      • 3
        Can we switch from second TNF inhibitor to a third one? As more patients are treated with anti-TNF agents, the need to switch to a third TNF inhibitor is becoming a more frequent clinical scenario. A retrospective study in CD patients who lost response or developed intolerance to 2 previous TNF inhibitors demonstrated benefit in 61% of patients after 6 weeks of treatment with a third anti-TNF. However, data from large prospective cohorts and placebo-controlled trials in rheumatoid arthritis strongly suggest that this small study result is overly optimistic.
      • 4
        Can patients who fail/lose response to infliximab be “rescued” by adding an immunomodulator? Although this strategy is attempted in clinical practice, it is not supported by evidence. It has been postulated that immunomodulators might decrease infliximab clearance by mechanisms independent of antibody formation.

      Recommendations for This Patient

      Our patient developed a loss of response to infliximab. Most importantly, a compulsive evaluation to document active mucosal inflammation and rule out other etiologies for symptoms must be first performed. Three samples of stool for culture, ova, and parasites and C. difficile toxins A and B were negative. Hemoglobin level was 10 g/dL, platelets were 450 × 109/L, and white blood cell count was 15 × 109/L. Additional blood indices included a CRP of 50 mg/dL. Colonoscopy demonstrated skip lesions with severe colitis in the ascending and transverse colon along with patchy moderate involvement of the rectum. Biopsies were negative for CMV and were compatible with CD. Serum infliximab was undetected, and antibodies to infliximab were present.
      Adalimumab (160 mg subcutaneously at week 0, 80 mg at week 2, and then 40 mg every other week) and methotrexate (25 mg subcutaneously weekly) were initiated with no response. This was continued for 3 months, with adalimumab dose escalation to 40 mg weekly during the third month. With minimal improvement, this treatment regimen was stopped. On the basis of results from the large natalizumab clinical trial program, this patient was a viable candidate for natalizumab (high CRP and infliximab failure). Other alternative treatments considered at this stage included surgery, cyclosporine/tacrolimus, or recruitment to a clinical trial with an experimental treatment. After careful clinical assessment and extensive risk-benefit discussions with the patient including surgical consultation, natalizumab 300 mg intravenously every 4 weeks was initiated, with a notable clinical response within 4 weeks, clinical remission within 3 months, as expected in 40% of patients on the basis of pivotal induction studies, and sustained clinically important benefit for almost 2 years currently.