Conflicts of interest: Michael Kamm has served as a consultant and speaker to Abbott, Schering-Plough, and UCB. Stephen Hanauer has received grants for clinical research and educational activities from, and has served as an advisor or consultant to, Shire, P&G Pharmaceuticals, Salix, Abbott, Centocor, UCB, and Prometheus.
Author contributions: M.A. Kamm (study concept and design; study supervision, critical revision of article), S.C. Ng (Literature search, drafting of article, revision of article), P. D'Cruz (literature search, drafting of article), PAllen (Literature search, drafting of article), S Hanauer (critical revision of article).
Anti-tumor necrosis factor (TNF) therapy to treat inflammatory bowel disease has been available for more than a decade. Although extensive data on the outcome of anti-TNF therapy from individual clinical trials and patient cohorts are available, integrated guidance on the best use of such therapy to achieve optimal clinical outcomes when managing patients with luminal Crohn's disease is lacking. This review combines published data to establish practical strategies for anti-TNF therapy with respect to effective and safe timing of introduction, use of concurrent immunosuppressive therapy, dose escalation, managing relapse, changing drugs, pregnancy and breast feeding, and stopping drug treatment. (Inflamm Bowel Dis 2011;)
Given the substantial published clinical trial and cohort evidence evaluating the benefits and limitations of anti-tumor necrosis factor (anti-TNF) therapy in luminal Crohn's disease (CD), it is timely to pull that evidence together to develop evidence-based treatment strategies. Anti-TNF therapy is often capable of effectively and quickly inducing and maintaining remission, diminishing inflammation, healing the intestinal mucosa, and improving quality of life. However, many challenges remain in obtaining optimal benefit from these drugs. A proportion of patients do not respond to, develop intolerance of, or become resistant to, anti-TNF drugs. Many patients who initially respond lose response over time. Within-class the initial response rate is similar for most drugs, although these drugs do differ in their profile and clinical use. Switching between anti-TNF drugs for loss of response to one drug is sometimes successful. Therapies that target alternative pathological processes are needed for patients who do not respond to anti-TNF drugs as well as for patients who lose response to anti-TNF agents.
While clinical trials provide essential, prospective, controlled data about single drug efficacy and safety, they do not provide all the information that is needed to guide clinical practice. In clinical trials strict inclusion criteria apply, eliminating the chance of treating the broader range of clinical scenarios. Patients leave a study if they do not meet prespecified endpoints, whereas in real life such patients may remain on a therapy or need to have their treatment adjusted. Changes in management during the course of a study are only possible in a predefined and limited manner.
There are many clinical scenarios that are not comprehensively covered by a single trial, but whose best management could be formulated by combining information from a range of studies including controlled trials, case series, experimental evidence, or deduction from studies in other diseases. For example, what should be the current best practice with regard to timing of introduction of anti-TNF therapy, what screening is required prior to their introduction, when should concurrent immunosuppressive therapy be started or stopped, when and how should drug dose be escalated, how and when should drugs be changed, when, should, or can drug treatment be ceased, the chance of a first, second, or third anti-TNF therapy being effective, how should relapses that occur while on anti-TNF therapy be managed, and what should one do about anti-TNF drugs when planning or during pregnancy, or during breast feeding?
This review aims to incorporate evidence derived from clinical trials and cohort data into a format that assists clinical decision-making in the management of patients with luminal CD, with a view to achieving best possible patient outcomes. We consider treatment goals, the optimal time to start biological therapy, the approaches available to maintain remission, and whether, and how, biological therapies can be stopped. We also incorporate safety considerations into the use of these drugs.
This article does not seek to cover the treatment of fistulizing CD,1 the entire treatment of CD with non-anti-TNF and anti-TNF therapies,2 or the use of anti-TNF therapy in ulcerative colitis.3 It is not a consensus document,4 and it will not comprehensively cover all aspects of anti-TNF drugs.5
REVIEW CRITERIA AND METHODOLOGY
The following sources were searched and relevant material included: 1) PubMed and MEDLINE were searched for all original research studies and reviews published in English-language journals, with a last cutoff publication date of July 2010, using the following keywords alone or in combination: “Crohn's disease,” “biological drugs,” “anti-TNF,” “Infliximab,” “Adalimumab,” “Certolizumab,” “clinical response,” “clinical remission,” “mucosal healing,” “immunogenicity,” and “antibody level”; 2) The reference lists from all identified full publications; 3) Abstracts presented at the United States (U.S.) Digestive Disease Week (DDW), American College of Gastroenterology Annual Scientific Meeting, the European Crohn's and Colitis Organisation (ECCO) Meeting, and United European Gastroenterology Week (UEGW) within the last 10 years.
ANTI-TNF DRUGS APPROVED FOR TREATMENT USE
The following drugs have been approved for the treatment of luminal CD: In 1998 infliximab was the first of three anti-TNF therapies to be approved by the U.S. Food and Drug Administration (FDA) in the U.S. for the treatment of adult patients with moderately to severely active luminal and fistulizing CD who have an inadequate response to conventional therapies. In 2006, this approval was extended by the FDA for the treatment of children with active CD (http://www.fda.gov/). In 2007 adalimumab was approved in the U.S. for the treatment of adult patients with moderate to severe, and in Europe for severe CD. In 2008, certolizumab pegol was approved in the U.S. for the treatment of adult patients with moderately to severely active CD who have not responded to conventional therapies. It has not been approved for use in Europe. Table 1 shows the characteristics, indications for use, modes of administration, and standard dose regimens of the three available anti-TNF drugs.
Table 1. Characteristics and Indications of Anti-TNF Drugs
Fc-free pegylated humanized anti-TNF-α Fb fragment
Approved for Use
2007 (US and Europe)
Treatment of patients with moderate to severe active luminal and fistulizing CD not responding to, or intolerant of conventional therapy
Treatment of patients with moderate to severe active CD not responding to, or intolerant of, conventional therapy
Treatment of patients with moderate to severe active luminal CD not responding to, or intolerant of conventional therapy
Mode of administration
5 mg/kg weeks 0,2,6 then every 8 weeks
16 0mg week 0 then 80 mg week 2 (USA, Europe), or 80 mg week 0 then 4 0mg week 2 (Europe), followed by 40 mg every 2 weeks (Europe offers 2 different induction regimens)
400 mg week 2 and 4, the every 4 weeks (if response occurs)
TREATMENT WITH ANTI-TNF THERAPY: DEFINITIONS OF RESPONSE AND REMISSION
Endpoints of Clinical Trials
Crohn's Disease Activity Index (CDAI): A Composite Disease Activity Score
In the clinical trial setting, disease activity in CD has been defined most commonly by the Crohn's Disease Activity Index (CDAI).6 According to the combined symptom and laboratory test-based CDAI, clinical remission has been defined as a score of less than 150. This is usually regarded as equating to an asymptomatic state. Further qualification includes whether the patient is off steroid treatment, and whether the patient was rendered asymptomatic after medical or surgical therapy. Most clinical trials use clinical response or remission as the primary endpoint and discontinuation of corticosteroids (sometimes referred to as “steroid-free remission”) as one of the secondary endpoints.
The definition of clinical response to an intervention has varied; while some trials have defined this as a 70-point or more decrease in CDAI,7 others have defined it as a 100-point or more decrease in CDAI,8 or a 70-point or more decrease in CDAI plus a 25% or more score reduction from baseline.9 These criteria do not specify whether the patient is on concomitant medications or off corticosteroids.
Primary nonresponse is defined as the lack of response or minimal, clinically insignificant response to the primary endpoint on initiation of a biological agent for the induction of remission.10 To be defined as demonstrating primary nonresponse a patient should have objective evidence of active inflammation and symptoms related to inflammatory CD. There is currently no consensus on how long a biological drug should be continued before lack of response is declared, although 58% of patients who will demonstrate a response to infliximab do so within 2 weeks.8 A substantial proportion of patients, however, will demonstrate a response up to 12 weeks after commencing treatment.11 Most responders demonstrate a clinical response by week 14 (in the case of infliximab that usually includes three induction plus one further dose). With adalimumab induction therapy it may require up to 12 weeks to determine whether response has been achieved.12Loss of response is defined as an initial response to a biological drug followed by a diminished or less durable response over time. Primary nonresponse has been defined variously as failure to achieve a defined fall of CDAI from baseline, or a CDAI which remains greater than 220. Loss of response has been defined by a CDAI of greater than 220, which correlates with at least mild disease activity. In patients with primary nonresponse, or loss of response, to an anti-TNF drug, consideration should be given to reassessing disease activity and excluding complications such as the development of stenosing disease, an abscess, or infections, including cytomegalovirus and Clostridium difficile.
Recent clinical trial and cohort data have demonstrated that when “mucosal healing” is achieved there is subsequently less clinical relapse, a decreased need for hospitalization, and less need for surgery. The goal of treatment in CD should therefore extend beyond symptomatic remission to include mucosal healing, as this endpoint has been shown to predict long-term outcome.13 The Crohn's Disease Endoscopic Index of Severity (CDEIS) assesses the degree of mucosal inflammation, and has been widely used to assess endoscopic improvement in clinical trials, but has not been used to define “mucosal healing” as a separate endpoint in clinical trials.14 The Endoscopic Score for CD (SES-CD) is a simpler index.15 The Rutgeerts score was developed to assess the extent of disease recurrence in the neoterminal ileum after a resection16, 17; the degree of endosopically identified recurrence is predictive of clinical recurrence.16 In clinical practice mucosal healing can be defined as the absence of deep ulcers, complete absence of ulcers or erosions, or no inflammation.18, 19 Alternatively, mucosal healing can be defined as complete normalization of the mucosa at endoscopy. As part of a longer-term view about disease management, mucosal healing is increasingly being considered as a factor that influences clinical decision making, including whether to change or cease anti-TNF therapy.
Response and Remission in Clinical Practice
Outside of the clinical trial setting, clinical remission is more loosely defined as the absence of any gut-related symptoms, with a normal C-reactive protein (CRP) or erythrocyte sedimentation ratio (ESR), hemoglobin, albumin, and platelet count. Clinical response is defined as the improvement of symptoms (reduced bowel frequency, reduced urgency, improved stool consistency, reduced abdominal pain, reduced rectal bleeding, improved general well-being, increased energy, and reduced lethargy) and improvement in blood test parameters (reduced CRP, ESR, and platelet count, and improved albumin and hemoglobin).
THERAPEUTIC GOALS: CLINICAL REMISSION, MUCOSAL HEALING, AND QUALITY OF LIFE
Goal 1: Inducing and Maintaining Clinical Remission
When infliximab was introduced in 1998, a single dose induction therapy resulted in 33% remission in the active group compared with 4% in the placebo group.7 In the ACCENT 1 (A Crohn's disease Clinical study Evaluating infliximab in a New long term Treatment regimen) study a single induction dose was used. Subsequently, a large Italian multicenter open-study demonstrated that three scheduled infliximab induction infusions at weeks 0, 2, and 6 were superior to a single infliximab infusion in patients with luminal and/or fistulizing CD followed up to 6 months.20 A three-dose induction regimen for infliximab is now regarded as standard.
Scheduled infliximab maintenance is more effective than episodic treatment.8 In the ACCENT 1 trial, which included patients with moderate to severe luminal CD previously treated with conventional therapies, patients who responded to a single dose of infliximab with a 70-point decrease in CDAI at week 2, were randomized to placebo or infliximab infusions (5 mg/kg at weeks 2 and 6, followed 5 mg/kg or 10 mg/kg each 8 weeks) until week 46. Fifty eight percent of patients responded at week 2. At week 30, remission rates in week-2 responders were higher in the infliximab than placebo group (45% versus 21%). Patients receiving scheduled infliximab were more likely to be in corticosteroid-free remission at week 54, were more likely to decrease and discontinue corticosteroid use, were more likely to achieve mucosal healing, had a reduced likelihood of anti-infliximab antibody development, and had improved quality of life with fewer hospitalizations compared to patients receiving episodic treatment.8, 21
Scheduled regular maintenance treatment results in less antidrug antibody production than episodic treatment.22 When treating active CD with infliximab, therefore, three-dose induction therapy followed by regular maintenance therapy every 8 weeks appears to be superior to episodic therapy. In the strictly defined environment of the clinical trial8 58% of all treated patients initially respond within 2 weeks, and 29% of these initial responders are in clinical remission at 1 year. In a single-center cohort study of 614 consecutive patients treated with infliximab, 63% of initial responder to infliximab showed sustained clinical benefit at a median follow-up of 5 years.23
The CLASSIC I and II (Clinical Assessment of Adalimumab Safety and Efficacy Study as an Induction Therapy in Crohn's Disease I and II) and CHARM (Crohn's trial of the fully Human antibody Adalimumab for Remission Maintenance) studies established the efficacy of induction and maintenance adalimumab therapy. In the CLASSIC I dose-finding induction study, remission at 4 weeks was achieved in more patients receiving the higher dose adalimumab (160 initially followed by 80 mg 2 weeks later—160/80 [36%]) than those on a lower dose (80 mg initially followed by 40 mg 2 weeks later—80/40 [24%], or 40 mg initially followed by 20 mg 2 weeks later—40/20 [18%]) or placebo.24 The CLASSIC II trial extended these findings and demonstrated that blinded treatment with adalimumab 40 mg every 2 weeks through to week 56 week was effective as a maintenance therapy with a remission rate at 1 year of 79% in induction-responsive patients.25
In the CHARM trial 854 patients with refractory CD received induction adalimumab 80 mg initially, followed 2 weeks later by 40 mg, and were then randomized to adalimumab 40 mg every week, 40 mg every other week, or placebo. Fifty percent of patients had received infliximab previously. Fifty-eight percent of patients responded to adalimumab induction. Clinical remission rates were significantly higher in the adalimumab 40 mg every other week and adalimumab weekly groups compared with placebo at week 26 (40%, 47%; 17%, respectively) and at week 56 (36%, 41%, 12%, respectively). At week 56, 23% on weekly adalimumab and 29% on adalimumab every 2 weeks were able to discontinue corticosteroids while maintaining remission, compared with 6% in the placebo group.12
In the open-label extension of the CHARM trial, the ADHERE (Additional Long-term Dosing with Humira to Evaluate Sustained Remission and Efficacy) study, adalimumab maintained long-term remission through to 2 years.26 Approximately one-third of patients taking corticosteroids at baseline who had adalimumab were able to discontinue steroids, while maintaining remission for up to 3 years.27, 28 Furthermore, among anti-TNF-naive patients who entered ADHERE, remission rates and quality of life improvement were sustained in 84% of patients on adalimumab maintenance therapy through to 3 years.29
The GAIN (Gauging Adalimumab efficacy in Infliximab Nonresponders) study evaluated the efficacy of adalimumab (160/80 mg or placebo) in inducing remission in patients with CD who had lost response to, or were intolerant of, infliximab.30 At week 4 clinical response (CDAI fall of 100) and remission were achieved in 38% and 21% of patients on adalimumab compared with 25% and 7% on placebo.
The CHARM and ADHERE studies have demonstrated that patients with moderate to severe CD can be treated effectively with adalimumab for at least 3 years. Treatment is associated with a decreased rate of hospitalization and surgery31 and sustained improvement in health-related quality of life. It is also effective in patients with CD independent of prior or concomitant use of immunosuppressive therapy.
The PRECiSE (Pegylated Antibody Fragment Evaluation in Crohn's Disease: Safety and Efficacy) 1 and 2 studies assessed the efficacy of certolizumab in inducing and maintaining remission in patients with active CD, respectively. In the PRECiSE-1 trial, 662 patients with moderate to severe CD were randomized to 400 mg of certolizumab or placebo at weeks 0, 2, and 4, and then every 4 weeks. The primary endpoints were clinical response at week 6 and week 26 in patients with a baseline CRP of 10 mg/L or higher. Response rates at weeks 6 and 26 were significantly higher in the certolizumab group (37% and 22%, respectively) than the placebo group (26% and 12%, respectively).9 The rate of remission was not different between the two groups at both timepoints.
In the PRECiSE-2 study, patients who had responded to 6 weeks of open-label induction treatment with certolizumab 400 mg were randomized to continue active treatment or receive placebo every 4 weeks. At 6 weeks, 64% of patients responded to induction therapy. The maintenance of response at week 26 was higher in the active treatment group than the placebo group (63% versus 36%).32 More patients who had a disease duration of less than 1 year maintained response and remission (90%, 68%, respectively) than patients who had a disease duration of more than 5 years (57%, 44%, respectively).33
Certolizumab also effectively maintains remission in CD for up to 18 months. Patients who completed PRECiSE 2 were eligible to enter PRECiSE 3, an open-label extension in which patients received certolizumab pegol 400 mg every 4 weeks for a further 54 weeks. At week 80 the remission rate for patients who had received continuous active treatment, and those who had received interrupted therapy (active induction, placebo, then open-label active treatment) were 62% and 63%, respectively. However, more patients in the drug-interruption group developed antibodies against certolizumab and had lower plasma concentrations of certolizumab than the continuously treated group.34
Certolizumab is effective in patients who are intolerant of, or who have lost response to, infliximab. In PRECiSE 2 the efficacy of certolizumab at week 26 was highest in patients naive to infliximab (response 69%; remission 53%), although some infliximab-experienced patients (response 44%; remission 33%) also benefited from second-line certolizumab therapy.35
In the WELCOME (26-Week open-label trial Evaluating the clinical benefit and tolerability of certoLizumab pegol induCtiOn and Maintenance in patients suffering from CD with prior loss of response or intolErance to infliximab) study, 539 patients who had failed infliximab were treated with open label certolizumab 400 mg at weeks 0, 2, and 4. Those in clinical response at week 6 were randomized to certolizumab 400 mg every 2 or every 4 weeks through to week 24. At week 6, 62% of patients responded, and 39% achieved remission. Among responders, certolizumab pegol 400 mg every 2 or 4 weeks resulted in similar maintenance of response (37% and 40%) and remission (30% and 29%).36
In the PRECiSE 4 study, the PRECiSE 2 patients (both on placebo and active groups) who were previously on active treatment and relapsed prior to week 26 were given a single extra 400-mg dose and then regular 400 mg each 4 weeks therapy. Patients in the placebo arm who relapsed were reinduced with certolizumab 400 mg at weeks 0, 2, and 4 followed by maintenance 400 mg every 4 weeks. At week 4 of PRECiSE 4 the response rates were 63% and 65%, respectively. Response was maintained in 55% and 59% of these responders, respectively, through to week 52.37
Summary of Induction and Maintenance Trials
Studies of the three licensed anti-TNF agents differed in their designs, and included different patient populations. For example, all patients in the ACCENT 1 study were anti-TNF therapy naïve, while about half in the CHARM and PRECiSE studies had been previously treated with infliximab. Comparisons of these three drugs should therefore be made with caution. There are no, and unlikely to be, head-to-head studies comparing the anti-TNF agents. In broadest terms infliximab, certolizumab, and adalimumab have comparable efficacy in the induction and maintenance of remission for up to 1 year. All three drugs should be administered on a regular basis to reduce the risk of formation of antidrug antibodies. The choice of the first anti-TNF agent will depend on a combination of patient preference with regard to the mode of administration, personal preference by the treating physician, and in some countries, the approval by third-party payers and residual patient cost. According to the trials, ≈20% of all patients treated with infliximab, adalimumab, or certolizumab will be in remission, and about 40% of initial responders will be in remission, after 1 year of treatment. Table 2 shows the response and remission rates of anti-TNF drugs from the strictly defined clinical trials; these underestimate the true response and remission figures. Serious adverse events associated with anti-TNF therapy in these studies have been infrequent, but careful screening and continuous monitoring while on therapy is required.38–42
Table 2. Clinical Response and Remission Rates of the Three Anti-TNF Agents
12-month data from cimzia studies are open label. The remainder of data in the table are from double-blind studies.
These numbers assume that patients who do not achieve early clinical response will not achieve remission at 1 year; which may not necessarily be the case. A substantial number of patients are nonresponders at week 2 or 4 according to the strict definition of the study, but will gain response by week 12 and go on to achieve remission. These figures are therefore an underestimate of the total number of patients who achieve remission.
The efficacy of anti-TNF therapy as an induction and maintenance agent has been confirmed in a number of large cohort studies. Furthermore, in steroid-dependent patients steroids can be reduced or withdrawn in many. Several studies have reported that infliximab has sustained efficacy beyond 2 years.23, 43, 44 The largest cohort of 614 patients with the most comprehensive follow-up demonstrated that over a median follow-up of 5 years, 63% of initial responders showed sustained clinical benefit from infliximab either as monotherapy or in combination with an immunosuppressant.23
Goal 2: Mucosal Healing
There is a poor correlation between clinical indices, such as the CDAI, with the endoscopic features of mucosal inflammation. Indeed, ≈20% of patients entered into the ACCENT I19 and SONIC45 studies with moderate-severe clinical activity based on the CDAI had no mucosal lesions noted at colonoscopy.46 If the course of CD is to be modified then symptomatic remission is not a sufficient therapeutic target. Mucosal healing is emerging as an important outcome measure in clinical trials because of its association with a decreased chance of relapse, reduced structural damage, and reduced need for surgery. Because of these associations, mucosal healing is believed to reflect diminished disease activity and deep tissue healing with consequent modification of the disease natural history. Although now a major focus in clinical trials, mucosal healing is yet to be fully integrated into clinical practice strategies.
Mucosal healing, regardless of the therapy used to induce it, is associated with improved outcomes in CD, at least in the short to medium term, that is, up to 2 years. “Improved outcomes” include a greater rate of clinical remission off corticosteroids, reduced hospitalizations and surgery, improved quality of life, and increased work productivity.13, 47–49
In a Norwegian, population-based, inception-cohort study patients with CD diagnosed between 1990 and 1994 (prior to the introduction of biologics) were treated with topical or systemic 5-aminosalicylate (5-ASA), corticosteroids, antibiotics, or azathioprine. Mucosal healing at 1 year after diagnosis was predictive of subsequent reduced disease activity, decreased need for corticosteroids, and reduced surgical episodes.13
A single dose of infliximab rapidly reduces the CDEIS score in patients with active CD.48 Scheduled regular maintenance infliximab therapy is associated with superior mucosal healing rates than intermittent, episodic, symptom-based therapy. In the ACCENT 1 endoscopic substudy significantly more week-2 responders on scheduled infliximab achieved mucosal healing at 1 year, compared with those on episodic infliximab (50% versus 7%, respectively).19 In a recent long term single-center observational study, 68% of patients who responded to infliximab achieved mucosal healing. More patients on scheduled infliximab had complete mucosal healing than those who received episodic infliximab, with a follow-up of up to 5 years.50 Patients who achieved mucosal healing had a much lower incidence of hospitalization and surgical intervention.19, 50 In the same study mucosal healing correlated with serum infliximab trough levels. Patients who achieved partial or complete healing on infliximab had significantly higher infliximab trough levels than those who did not achieve healing.51
Mucosal healing is more likely to be achieved using more intense early immunosuppressive therapy than gradual step-up induction therapy. In the study comparing early intense induction therapy with infliximab and azathioprine versus gradual step-up induction therapy in patients who were treatment-naïve and had a short disease duration (the “step-up top-down study”), a significantly greater proportion of the former group achieved mucosal healing at 2 years (73% versus 30% respectively).18
Achieving early, complete mucosal healing is also a predictor of sustained clinical remission in subsequent years. For instance, in the “step-up top-down study” a significantly greater proportion of the patients who achieved mucosal healing at 2 years, regardless of whether they had received initial step-up induction therapy or early combination infliximab plus azathioprine therapy, were in stable remission without corticosteroids in the subsequent 3–4 years, compared with patients without mucosal healing at 2 years.52 Combination immunosuppressive therapy, at least with infliximab and a thiopurine, achieves a greater rate of mucosal healing than either therapy alone, when treating active disease in immunosuppressive-naïve patients with less than 2 years disease duration. These findings from the SONIC (Study of Biologic and Immunomodulator Naive Patients in Crohn's Disease) study suggest that infliximab and azathioprine have a synergistic effect in relation to disease suppression and healing the mucosa.45 This improved outcome from combination therapy may relate to a synergistic immunosuppressive effect, decreased anti-infliximab antibody production, or both.
Mucosal healing is associated with long-term, steroid-free clinical remission in patients with early stage CD. In a prospective cohort of 133 newly diagnosed and treatment-naïve CD patients who were initially treated with either combination azathioprine and three doses of infliximab, or conventional therapy with corticosteroids, complete mucosal healing at 2 years predicted subsequent steroid-free remission up to 4 years after therapy was initiated. More patients with mucosal healing at 2 years were maintained in remission (71%) without further infliximab infusions or steroids in the subsequent 2 years compared with those without healing at 2 years (27%), regardless of their initial medical therapy.52
In the open-label EXTEND (Extend the Safety of and Efficacy of Adalimumab Through Endoscopic Healing) study, maintenance treatment with adalimumab resulted in a significantly greater rate of mucosal healing (defined as absence of mucosal ulceration at endoscopy) than placebo maintenance at 1 year (24% versus 0%). Patients with ileocolonic CD and a simple endoscopic score (SES-CD) (score 0–9; 0 = no mucosal inflammation) of at least 2–3 in at least one of five anatomic segments received open-label adalimumab induction 160 mg followed by 80 mg 2 weeks later. All patients were then randomized to placebo or adalimumab 40 mg each 2 weeks. At 12 weeks clinical remission was achieved in 28% of placebo and 47% of adalimumab-treated patients, and complete mucosal healing was achieved in 13% and 27% of patients, respectively. At 52 weeks clinical remission was achieved in 9% and 33%, and complete mucosal healing in 0% and 24% of patients, respectively.53
In the MUSIC (endoscopic MUcoSal Improvement in patients with active Crohn's disease treated with certolizumab pegol) study 62% of patients treated with open-label certolizumab achieved endoscopic response (defined as a reduction of CDEIS score of 5 or more points), 42% had endoscopic remission (CDEIS <6 points), and 12% had complete mucosal healing (CDEIS <3) at 10 weeks.54
Summary of Mucosal Healing
Mucosal healing appears to be a marker of effective disease suppression, or “deep remission.” Achieving mucosal healing can reduce structural damage and result in fewer long-term complications and surgical episodes in patients with CD. These goals are best achieved with early intervention with biological agents. Future studies need to address the significance of partial mucosal healing on long-term outcome.
The data on the superior outcome in the patients who achieve mucosal healing in clinical trials or cohort studies are clear. It is reasonable to assume that a proportion of patients who have not achieved mucosal healing using less intense immunosuppression (such as with a thiopurine alone) will achieve mucosal healing if treatment is intensified, using biological or combination therapy. However, it has not been tested whether intensifying treatment with the goal of achieving mucosal healing in those patients who have not achieved it while on biological monotherapy, or combination therapy, results in an improved outcome. The subgroup of patients who do not achieve mucosal healing with intense early therapy may have disease that is more “resistant” to treatment of specific inflammatory mechanisms, or mucosal healing may be achieved with more intense therapy, but at increased cost or risk of side effects. However, the benefit of increasing the intensity of therapy for patients in clinical remission on “optimal” combination therapy, with a goal of achieving mucosal healing, has not yet been tested.
Goal 3: Improvement in Quality of Life
CD can affect patients psychologically and socially, as well as physically.55, 56 The assessment of health-related quality of life (HRQoL) has been demonstrated to be a valid and reliable measure of therapeutic efficacy in the treatment of IBD.57 A variety of patient-reported outcome (PRO) instruments have been utilized to provide information about HRQoL in trials of anti-TNF therapy. The most commonly used disease-specific instrument is the Inflammatory Bowel Disease Questionnaire (IBDQ) which consists of 32 items that measure bowel, systemic, social, and emotional functions. The IBDQ is simple, reliable, consistent across cultures, and is responsive to changes in disease status.58, 59
The most commonly used generic PRO instrument is the Short Form-36 (SF-36), which comprises eight domains: physical functioning, role limitations due to physical problems, bodily pain, general health, vitality, social functioning, role limitations due to emotional problems, and mental health.60 The EuroQol-5D (EQ-5D), a generic PRO questionnaire, contains five items assessing mobility, self-care, usual activities, pain and discomfort, and anxiety and depression, and a 20-cm visual analog scale (VAS) measuring patients' perception of their current health status.
HRQoL in Active Luminal Disease
Improved HRQoL, as measured by the IBD-Q and SF-36, have been demonstrated in randomized, controlled trials in active luminal CD with infliximab,7, 61, 62 adalimumab,12, 63 and certolizumab.9, 64, 65
In the CHARM study63 remission and clinically significant improvements in HRQoL achieved among anti-TNF-naïve patients were sustained for up to 3 years of adalimumab maintenance therapy.
Screening Investigations Prior to Starting Biological Therapy
A number of comprehensive reviews have recently addressed: latent infection screening, vaccination prior to immunosuppression, and monitoring for opportunistic infections.66–69
A number of precautions are necessary before commencing anti-TNF therapy. A complete history related to risk factors for infection and immunization history needs to be obtained. Before starting biological therapy, recommended tests include a full blood count, hepatitis B virus serology, varizella zoster virus serology (in the absence of a history of chickenpox), HIV serology (after counseling), a chest radiograph, and a tuberculin skin test or interferon gamma release assay according to local or national practice guidelines. Cervical smear testing and immunization against Human Papilloma Virus (HPV)70, 71 should be encouraged in female patients who are in the appropriate age group.68, 69 The recent pandemic influenza A (H1N1) caused substantial morbidity in young patients and those who were immunocompromised. Complication of H1N1 virus infection includes pneumococcal infection. Prior to biological therapy patients should receive vaccinations against seasonal influenza, H1N1 virus, and pneumococcal polysaccharide vaccine.72
Contraindications to anti-TNF therapy include obstructive stenosis, active infection, including evidence of latent tuberculosis, congestive cardiac failure, undrained abscesses, and presence of demyelinating disorders.69
There are several deficiencies that limit the interpretation of available adverse event data. These include: 1) a lack of long-term data; 2) many patients randomized to placebo in clinical trials received initial open-label active therapy; 3) administration of open-label (rescue) or crossover to anti-TNF therapy for patients initially randomized to placebo therapy; 4) many patients in trials were on concurrent corticosteroid or immunosuppressive drugs; and 5) patients in clinical trials may not accurately reflect the broader population of patients.
Opportunistic Infections and Malignancies
The best short-term (up to 1 year) data have been derived from controlled trials of biological drugs. The infection risk associated with infliximab, adalimumab, and certolizumab therapy appear to be comparable. In a Cochrane review of adverse events of biologics in CD, which included studies of infliximab,8, 73, 74 adalimumab,12, 25 and certolizumab,32 adverse events (aside from opportunistic infections such as tuberculosis) were observed with similar frequency in the treatment and placebo
The reactivation of latent tuberculosis (TB) is a well-recognized complication associated with the use of anti-TNF therapy.12, 32, 75, 76 The estimated incidence of TB in patients treated with infliximab for rheumatoid arthritis and CD between 1998 to 2002 was 54 per 100,000. More than 50% of cases were extrapulmonary disease and most cases occurred in endemic regions.75
Following the early recognition of the high incidence of TB in patients receiving anti-TNF therapy, rigorous screening was instituted prior to commencing therapy. This had a dramatic impact on reducing the incidence of TB. In a Spanish study based on a registry of patients with rheumatoid arthritis treated with anti-TNF therapy, active TB developed in 34 patients; 32 of these events occurred prior to recommended screening, with only two cases occurring after the introduction of screening with tuberculin skin testing and chest x-ray.77 Other studies from the U.S. have reported that screening for TB with tuberculin skin test, chest x-ray, and chemoprophylaxis when indicated, prior to anti-TNF therapy, can reduce the incidence of TB by up to 90%.78 More recent interferon gamma assays may further enhance screening efficacy.
In the United States the TREAT (The Crohn's Disease Therapy, Resource, Evaluation, and Assessment Tool) registry, which prospectively records adverse events in infliximab- and nonbiological-treated patients with CD, three cases of TB have been reported in the infliximab-treated group and one in the noninfliximab group.79
Fewer data on the risk of TB are available for adalimumab and certolizumab since they were studied after recognition of the TB risks with other anti-TNF agents and the imposition of TB screening, but these agents probably carry a similar increased risk for TB as infliximab. In the CHARM trial, of 854 patients treated with adalimumab two cases of TB were reported.12 In the PRECiSE 2 study involving over 600 patients treated with certolizumab, one case of pulmonary TB was reported.80
Population- and hospital-based studies have assessed the risk of other opportunistic infections associated with the use of anti-TNF drugs. The population-based Danish Crohn's and Colitis Database demonstrated that about 10% of patients treated with infliximab suffered at least one serious infection, including abscess, Aspergillus pneumonia, or osteomyelitis.81
In a case–control study (734 IBD patients and 666 control patients) that assessed the long-term safety of infliximab in patients with IBD treated over a 14-year period, serious adverse events occurred in 13% of patients treated with infliximab and 19% of control patients (odds ratio [OR] 1.33). No difference was observed between the two groups for mortality, malignancies, and infection. Concomitant treatment with steroids was the only independent risk factor for infections in patients treated with infliximab (OR 2.69).82
The TREAT registry, which consists of more than 6000 CD patients (3401 patients treated with infliximab and 2872 patients treated with nonanti-TNF drugs) with a mean follow-up of 4.8 years has demonstrated that infliximab or prednisolone alone conferred a 2-fold increased risk of serious infections, whereas the use of narcotics increased the risk by 6-fold. The risk for serious infections was similar for infliximab alone and infliximab plus an immunomodulator.79
Up until April 2008, 3160 patients with CD had been treated with adalimumab in clinical trials, representing 3402 patient-years of adalimumab exposure. Serious infection was one of the most frequently reported serious adverse events. The safety profile of adalimumab in patients with CD was similar to that of other TNF antagonists.83
In a case-control study of 100 consecutive IBD patients, the use of corticosteroids (OR, 3.4), azathioprine or mercaptopurine (OR, 3.1), and infliximab (OR, 4.4) were associated individually with an increased risk for opportunistic infection. The use of at least two of these drugs yielded an OR of 14.5 for opportunistic infection.84
A recent meta-analysis investigated the rate of non-Hodgkin's lymphoma (NHL) in adult CD patients who had received anti-TNF therapy and in patients treated with immunomodulators.85 Among anti-TNF-treated patients, 13 cases of NHL were reported out of 8905 patients with 21,178 patient-years of follow-up (6.1 per 10,000 patient-years). The majority of these patients had previously been exposed to immunomodulators.
In a recent study of pediatric patients treated with combination infliximab and thiopurine, cases of rare, but fatal, hepatosplenic T-cell lymphoma have been described.86 This condition affects most commonly young males who have received thiopurine therapy, with or without anti-TNFs, but not with anti-TNF therapy alone.87 This complication of therapy has also occurred with combination thiopurine and adalimumab therapy. In young male patients, therefore, consideration should be given to anti-TNF monotherapy, or time-limited combination therapy.
A recent meta-analysis assessing the safety of anti-TNF therapy in 21 studies with 5356 individuals showed that there was no increased risk of death, serious infection, or malignancy in overall, subgroup, and sensitivity analyses.88
In summary, the risk of serious infections with biological drugs is small but definite. Increased caution should be taken when treating older patients or when combining biological drugs with corticosteroids or immunosuppressive therapies. The long-term risk of malignancy related to biological agents is unclear. At least in the short to medium term, the absolute risk of malignancy appears to be low. All risks should be weighed against the potential benefit in individual patients.85
WHEN TO START ANTI-TNF DRUGS IN CD
Undertreatment of CD results in ongoing active disease, poor quality of life, and the development of disease complications, whereas overtreatment leads to an unnecessary risk of infection, malignancy, other drug side effects, and cost. As with any intervention, potential benefits and risks need to be assessed on an individualized basis. The evidence of increased benefit from early, versus later, treatment should be considered. Patients with an adverse prognosis should be identified early and treated “aggressively” to optimize outcomes and minimize complications. Patients should be selected for their likelihood of gaining benefit by confirming active inflammation and excluding patients with infectious complications or significant stricturing disease.
Natural History of CD
An appreciation of the natural disease history and likelihood of having active disease or complications over time is needed in order to anticipate what proportion, and which, patients should receive potent therapy.
In an inception cohort population-based study from Denmark, in any 1 year 55% of patients were in remission and 15% had mild disease activity. Approximately 15% of patients were disabled by their disease after 15 years. Fifty percent of patients did not have any year with impaired capacity for work within 10 years of diagnosis; hence, 50% of patients were faced with impaired work capacity. The 10-year cumulative probability of surgery was 61%. Only 4% of patients had continuously active disease. In patients followed for longer than 7 years, 25% of patients had active disease each year, 22% were in remission, and 53% had fluctuating disease.89 Similar data in a 24-year population-based inception cohort from Olmsted County U.S. showed that about two-thirds of patients were in clinical or surgical remission at any one time.90
In a long-term population-based prospective study of 200 CD patients diagnosed between 1990 and 1993 in Norway, 28% of patients underwent an intestinal resection during the first 5 years after diagnosis. Forty four percent of patients reported symptom improvement during the 5-year follow-up. Overall, the 5-year disease course in this unselected cohort of CD patients was mostly mild.91
In contrast, in an American population-based study of 173 patients with CD diagnosed between 1970 and 1993, the majority of patients had an “unfavorable” outcome at long-term follow-up. Approximately one-third of patients were steroid free without an operation at 1 year. Of note, this cohort was treated before the common use of thiopurines in that area.92
Population-based cohort studies assessing the natural history of pediatric CD have shown complicated behavior in 29% of children at diagnosis and 59% at a median of 84 months follow-up. In contrast to an adult population where immunosuppressive therapy may have been administered later in the disease course,93 the introduction of immunosuppressive therapy has been shown to reduce the need for surgery in children.94, 95
In a retrospective review of 297 patients with CD, the location of disease appeared stable over the course of disease, whereas the behavior of disease changed in half the patients. Over 10 years 16% of patients had a change of disease location, and 46% had a change in disease behavior from nonstricturing and nonpenetrating to stricturing (27%) or penetrating (29%) disease. Patients with ileal CD mostly have stricturing disease, whereas those with colonic disease are more likely to develop penetrating complications.96 The majority of patients with CD will eventually develop a stricturing or a perforating complication. The initial location is the main factor in determining the type of complication.97
Factors to Consider When Deciding “When to Start” Anti-TNF Drugs
The decision to initiate biological therapy should be influenced by the patient's disease characteristics, previous response to other medical therapy, and the presence or absence of adverse prognostics factors. In contrast to the population studies mentioned previously, Loly et al98 reported that 58% of patients had disabling CD within 5 years of their initial diagnosis. Predictive markers for disabling disease include perianal disease, the need for steroids for first flare, and ileocolonic disease location. Beaugerie et al99 demonstrated that a poor outcome and disability were associated with young age at onset (<40 years old), perianal disease, the need for systemic steroids at presentation, isolated small bowel disease, and smoking. Other high-risk factors include previous poorly controlled disease requiring multiple medications and operations, high disease burden as evident by high serum CRP, and deep mucosal lesions. Smokers have been shown to have a lower response rate, and shorter duration of response, to infliximab compared to nonsmokers.100, 101 In addition, older patients, patients with isolated ileitis, stricture, and prior CD surgery showed reduced response to infliximab. Considering these factors together, in patients with concurrent perianal disease, age of onset less than 40 years, smoking history, extensive disease, and recent onset aggressive disease, early intervention with a biological agent should be considered.
Length of CD History and Response to Anti-TNF Agents
There is a higher response rate to anti-TNF therapy in patients with a shorter disease duration, a consistent finding for all three anti-TNF antibodies. In a subanalysis of the CHARM trial of patients treated with adalimumab, the remission rates at 1 year were 51% and 25% in patients with CD for less than or longer than 2 years, respectively.102 In a subanalysis of the PRECiSE trials, the clinical response to certolizumab rate at week 26, defined as a reduction in the CDAI of 100 points or greater, was greater in patients who had disease for less than 1 year compared with those who had disease for longer than 5 years (68% versus 44%, respectively).33
The CARE (Crohn's patients treated with Adalimumab: Results of a safety and Efficacy) study addressed the response and remission rates with adalimumab treatment in infliximab-naïve patients, and patients with primary nonresponse or loss of response to infliximab. In a subanalysis of this study there was a trend towards a greater remission rate in CD patients with a disease duration of less than 2 years compared with those with a disease duration of more than 5 years.103
These data suggest that the impact of biological therapy is greater earlier in the disease course. This may relate in part to biological factors, but also to the lack of transmural structural damage such as strictures and fistulae earlier in the disease course. Although patients with a long disease duration may have a slightly lower response or remission rate to anti-TNF therapy than those with a short disease history, the former should still receive treatment if indicated.
Impact of Anti-TNF Therapy in Recent Onset CD
The strength of initial treatment, that is, the induction regimen, has an impact on the subsequent disease course. In an early intervention study, 129 steroid-naïve patients with disease duration of 4 months or less were randomized to “top-down” therapy (three doses of infliximab at weeks 0, 2, and 6 together with commencement of azathioprine, followed by further, episodic infliximab or steroids as needed) versus conventional “step-up” therapy (steroids with azathioprine added as needed, and episodic infliximab in treatment failure). The steroid-free remission without surgery rate was significantly higher in the “top-down” group at 26 weeks (60% versus 36%) and 52 weeks (62% versus 42%). The “top-down” group had more rapid normalization of CRP and higher rates of remission than conventional group at 1 year. At 2 years remission was not different between the two treatment arms, although more patients in the “top-down” arm achieved mucosal healing (73% versus 30%).18 The development of fistulas, serious adverse effects, and the rates of surgery were not significantly different between the two treatment groups.
Early initial healing with infliximab may influence the natural history of CD but whether this is maintained in the long term is unknown. This “top-down versus step-up” study was conceived at a time when episodic therapy for symptom relapse was considered a satisfactory dosing regimen; it is likely that the benefit may have been greater if the currently accepted, regular, scheduled dosing regime was used. These studies measure the effect in populations of patients; such studies will therefore include patients in whom the disease course would have been less aggressive and in whom a “top-down” approach may have been unnecessary.
In the recent SONIC study, 508 CD patients with a median disease duration of 2 years and naive to thiopurines were randomized to the combination of infliximab and azathioprine, infliximab alone, or azathioprine alone.45 Steroid-free remission at week 26 was significantly more often in patients receiving combined azathioprine and infliximab compared with those who had infliximab monotherapy or azathioprine monotherapy (57%, 44%, and 30%, respectively). The highest rate of mucosal healing at week 26 was achieved in the combination therapy group (44%, 30%, and 17%, respectively). At 1 year steroid-free remission was seen in 72%, 61%, and 55% of patients, respectively. In patients with treatment-naïve, short-duration disease, combination therapy with infliximab and azathioprine is most effective in inducing and maintaining remission, and in achieving mucosal healing.
Recognition of adverse prognostic factors should lead to early use of anti-TNF therapy. Such factors include failure to achieve and maintain remission with conventional therapy, extensive small bowel disease, aggressive disease from onset, and one or more previous operations. Other high-risk factors including young age and smoking should be taken into consideration. Combination therapy is most effective. In young males, in particular, the risks associated with combination therapy need to be considered. In patients with uncontrolled luminal disease despite the use of a thiopurine, a biologic therapy can be commenced. In this circumstance it is unknown whether maintaining the azathioprine confers additional benefit to using the biological therapy alone. However, it is probably reasonable to use combined therapy, at least initially, to reduce immunogenicity to the biologic. Figure 1 shows an algorithm of factors to consider when deciding to initiate anti-TNF therapy, and Figure 2 shows the optimal, initial dose, and initial dosing interval when using anti-TNF therapy to induce and maintain remission in CD.
In patients with limited anatomical involvement and longstanding disease without progression induction for a disease flare can be achieved with either steroids or anti-TNF therapy. In patients with concurrent perianal fistulas, treatment with biological therapy, with or without a thiopurine, has been shown to be the most effective initial approach.
MAINTENANCE OF REMISSION WITH ANTI-TNF DRUGS
Following the successful induction of clinical remission with a biologic drug the maintenance of disease remission can be achieved in 30% to 80% of those patients at 1 year, depending on the biologic used, initial treatment response, and patient population.8, 24, 25 Strategies that optimize biological therapy to maintain response in the long term are critical to avoid disease complications and the need for surgery. Maintenance of mucosal healing as a specific endpoint is less well studied.
Optimizing Anti-TNF Therapy to Maintain Remission
Dosing Interval: Use a Scheduled Maintenance Strategy
Anti-TNF therapy has been generally dosed as one of three regimens: episodic (single anti-TNF antibody treatment at week 0, and retreatment on disease relapse), scheduled maintenance (regular scheduled anti-TNF therapy), and flexible regular therapy (regular treatment within a predefined time range according to the redevelopment of symptoms).
In the ACCENT I study patients who responded to an initial dose of infliximab and were maintained on scheduled regular maintenance therapy were more likely to: maintain clinical response and remission, discontinue steroids, achieve mucosal healing, and have reduced CD-related hospitalizations, compared with those on episodic treatment.8, 21 The development of antibodies to infliximab was more common in patients treated with episodic therapy (28%) compared with those on scheduled maintenance therapy (6%–9%).
In patients who have been refractory to immunosuppressants and respond to induction therapy with anti-TNF agents, scheduled maintenance therapy with anti-TNF therapy is necessary to prevent relapse.38, 104, 105
In a large observational study of 614 consecutive patients followed over a median 55 months, 46% of patients who were started on episodic infliximab lost response and had to switch to a scheduled maintenance regimen.23 Scheduled maintenance therapy is therefore superior to episodic treatment for the maintenance of remission in CD.81 However, a few centers in Europe have continued to use anti-TNF therapy episodically as a bridging strategy to immunomodulator, largely due to financial restrictions.106
Anti-TNF Therapy to Bridge to an Immunomodulator
Anti-TNF drugs have been used short-term to induce remission, which is then maintained with a standard immunosuppressive drug, usually a thiopurine. In a study from the French GETAID group, 113 patients with active CD despite at least 6 months of prednisolone, including patients who had failed to respond to a thiopurine and those who were thiopurine-naive, were randomized to three infusions of infliximab or placebo. All patients were treated with thiopurine maintained at a stable dose throughout the 1 year of study. Sixty-three percent of patients treated with infliximab achieved steroid-free remission, compared with 37% on placebo (despite thiopurine therapy). At 1 year 52% of the thiopurine-naive patients were in clinical remission compared with 27% of the thiopurine-failure patients.104
Whether the use of anti-TNF therapy as an induction agent is superior to an oral steroid has not been tested; in a rheumatoid arthritis study107, 108 anti-TNF therapy was similarly effective to corticosteroids in inducing remission, which was then maintained with methotrexate.
Combination Anti-TNF and Immunomodulator Versus Anti-TNF Monotherapy
Whether anti-TNF agents should be used as monotherapy or in combination with an immunomodulator remains controversial. The three issues central to the decision about combination therapy include immunogenicity, efficacy, and safety.2
During anti-TNF therapy the development of antibodies directed against the drug and subtherapeutic drug concentrations are associated with a reduced duration of response. In a study of 125 consecutive patients with CD, antibodies against infliximab were detected in 61% of patients. Higher antibody levels (8.0 μg/ml or greater) before an infusion correlated with 1) a shorter duration of response; 2) an increased frequency of infusion reactions and; 3) a lower infliximab concentration 4 weeks postinfusion. In patients receiving episodic infliximab concomitant immunosuppressive therapy was associated with lower titers of antibody against infliximab and higher serum concentrations of infliximab 4 weeks after an infusion.109 Combination anti-TNF therapy and immunosuppressive drug reduces immunogenicity.
In a prospective multicenter cohort study of 174 CD patients treated with episodic infliximab the concomitant use of immunosuppressive therapy (azathioprine or methotrexate) was associated with a lower incidence of antibodies to infliximab (46%) when compared with patients not taking concomitant immunosuppressive therapy (73%). Methotrexate and azathioprine appear to have similar efficacy in reducing the risk of antibody formation.22, 106 The long-term impact of combination therapy on loss of response is not completely clear.
In a subanalysis of the ACCENT I trial, reduced antibody formation and greater clinical benefit were seen in patients who had an induction regimen followed by regular scheduled maintenance infliximab treatment compared with those who had a single dose followed by episodic infliximab retreatment. Immunomodulators were associated with a decreased rate of antibody formation in patients receiving episodic treatment as well as in patients on scheduled maintenance infusions.22
There are conflicting data concerning whether concomitant use of an immunosuppressant is of therapeutic benefit in patients on scheduled anti-TNF maintenance therapy.
The SONIC study demonstrated that in thiopurine-naïve patients infliximab combined with azathioprine is more effective than infliximab monotherapy, in inducing remission and in achieving mucosal healing. This was especially marked in patients with high CRP and/or extensive mucosal lesions.45
Immunomodulator refractory patients
The role of combination therapy in patients who have failed an immunosuppressant is less clear. In the key studies involving infliximab (ACCENT 1), adalimumab (CHARM), and certolizumab (PRECiSE 1 and 2) approximately half the patients were on an oral immunosuppressive, mostly a thiopurine, but still had active disease. The outcome with anti-TNF therapy was not significantly different between those on or off an immunosuppressive therapy.
The CARE open-label study compared the short-term efficacy (week 4 and sustained at week 20) of adalimumab monotherapy versus adalimumab together with corticosteroids or immunomodulators. No significant difference was observed in the remission rate between patients on monotherapy versus combination immunomodulator therapy.110
The long-term benefits of combining scheduled infliximab with concomitant immunomodulators (azathioprine or 6-mercaptopurine) in patients with CD are unclear. In a study of 123 patients followed for 5 years after initiation of infliximab for active CD, clinical remission rates in the entire cohort were 73% at 1 year, 65% at 2 years, and 58% at 5 years. Remission rates with maintenance infliximab were significantly improved in those receiving concomitant immunomodulators at 1 year (86% versus 68%), but not at 2 years (80% versus 72%). The risk of surgery was significantly reduced in those receiving immunomodulators at the commencement of maintenance infliximab, but not in patients who continued maintenance concomitant therapy. The combination of maintenance infliximab and an immunomodulator produced modest improvements in outcomes beyond maintenance infliximab alone in this cohort.105
A similar study examined the outcome in 98 patients with CD who received infliximab and at least 6 months of immunosuppressive cotreatment with azathioprine or methotrexate.111 Disease flares, perianal complications, and switch to adalimumab were less frequently observed, and CRP was lower in periods of treatment with, compared to periods without, concomitant immunosuppression. IBD flares and perianal complications were less frequently observed in periods on azathioprine than in those with methotrexate.
A key difference between ACCENT 1 and CHARM, and these latter studies, that may account for these differences in relation to the benefit of concomitant immunosuppression is that only the former studies comprised exclusively patients who had no, or lost, response to standard immunosuppressive therapy prior to entering the study.
Baert et al109 demonstrated that combination therapy improved efficacy by reducing immunogenicity when infliximab was used episodically. This effect is less marked and less necessary in the current practice of scheduled maintenance anti-TNF therapy, although in a lesser number of patients may still be important.
In the IMID (Infliximab Maintenance Immunosuppression Discontinuation) study, 88 patients with CD who were in clinical remission after at least 6 months of combination therapy with infliximab and immunosuppressive therapy remained on infliximab but were randomized to continue at the same dose or to cease azathioprine, 6-mercaptopurine, or methotrexate. After 2 years, there was no significant difference in the clinical endpoint, defined as the need to change the infliximab dosing regimen or discontinue infliximab, between the two groups. However, while continuation of an immunosuppressant beyond 6 months offered no clear clinical benefit over 2 years of scheduled infliximab monotherapy, the former was associated with higher median infliximab trough levels, a lower frequency of antibodies to infliximab, and decreased CRP levels.112 In follow-up of the patients who stopped azathioprine the cumulative probability of infliximab failure-free survival was estimated at 85% at 12 months and 41% at both 24 and 32 months. Predictors of infliximab failure were a duration of combination therapy of less than 27 months and biological markers of ongoing inflammation (elevated CRP and platelet count).113
In a similar study, 35 patients who stopped immunosuppressive therapy after at least 6 months combination therapy were evaluated a median of 38 months after stopping the concomitant immunosuppressive therapy and continuing infliximab monotherapy. Trough infliximab levels did not change significantly after stopping immunosuppressive therapy. A low trough level prior to stopping immunosuppressive therapy predicted later infliximab clinical failure.114
Sokol et al111 showed that withdrawal of immunosuppressants in patients receiving scheduled infliximab maintenance therapy did not result in an increased disease recurrence rate or infliximab failure rate.
In summary, the benefit of dual immunosuppressive-anti-TNF therapy is probably maintained beyond 2 years, even in patients with previous thiopurine-refractory disease. In patients who have achieved sustained clinical and endoscopic remission for 6 months or more on combination therapy, the thiopurine can be withdrawn but there is a slightly increased risk of anti-TNF failure. In patients with previous recalcitrant disease who achieve remission on joint therapy, it may be prudent to maintain therapy with both drugs.
Safety of Combination Therapy: Concomitant Thiopurine Use in Young Caucasian Men
As discussed in the previous section on safety of biological therapy, combination therapy with two or more immunosuppressive agents has been associated with an increased risk of opportunistic infections and lymphoma.84, 115
The rare, but usually fatal, hepatosplenic T-cell lymphoma has occurred predominantly in young, Caucasian males who have been on combination therapy, in particular the combination of thiopurine with anti-TNF therapy.116 Given that the risk relates largely to this patient subpopulation it may be reasonable to limit combination therapy with thiopurines in this patient subgroup. Consideration should be given to the patient's prior disease course and drug history. It may be that immunosuppressive therapy is used for only the first 6 months of biological treatment, especially if clinical remission is reached by that time.
Patients with Previously Severe Disease and High-risk Disease Characteristics
A case can be made for longer-term combination therapy in patients with adverse high-risk disease features, such as aggressive disease from onset, previously poorly controlled disease with lack of response to conventional immunosuppressives, extensive small bowel disease, associated perineal fistulizing disease, and patients with more than one previous operation. There are no prospective data to demonstrate superior disease control with this approach, although the SONIC study in luminal disease and a study of combination therapy in patients with perianal fistulizing disease support indirectly the efficacy of such an approach.117 These patients with aggressive disease in whom the stakes are high if there is loss of response to biologic therapy are likely to require long-term maintenance biologic treatment. The decision needs to be made on a risk versus benefit basis.
Patients Not in Clinical Remission After 6 Months of Combination Therapy
In patients who fail to achieve clinical remission after 6 months of combination therapy, reassessment of disease activity is important to exclude complications including strictures, abscesses, and superadded infections such as cytomegalovirus. Repeat endoscopy and imaging should establish the extent of mucosal inflammation, as a subgroup will have normal endoscopy and presumably “functional” symptoms. If there is ongoing inflammation combination therapy should be continued, and dose intensification of the biological agent or the immunosuppressive considered. Alternatively, an alternative biologic agent may be needed to achieve remission.
Figure 3 summarizes the current approach to maintaining response and remission in patients with luminal CD.
LACK OR LOSS OF RESPONSE TO ANTI-TNF THERAPY
Lack of Primary Response to Anti-TNF Therapy (Fig. 4)
The initial response to anti-TNF therapy in luminal CD has been defined in studies as an initial CDAI reduction of more than 70 points8, 12, 74 or 100 points32 from baseline, a 25% or more reduction in the total CDAI score,8 or a combination of these criteria.8 Although used in clinical trials, the two main disease indices, the CDAI and the Harvey Bradshaw Index, are less relevant in clinical practice. Clinical response is based on symptom reduction and normalization of laboratory parameters such as CRP, albumin, and hemoglobin.
In clinical practice the lack of primary response to anti-TNF therapy occurs less often than has been seen in clinical trials, due to the strict index definitions used in the latter.
In the ACCENT I trial 42% of patients had a primary lack of response to infliximab induction therapy, as defined by a failure to achieve a 70 point or more reduction in the CDAI at 2 weeks.8 In two other large single-center retrospective studies, primary lack of response, assessed clinically, occurred in ≈10% of patients.23, 118
In the CHARM and PRECiSE-2 trials, 42% and 36% of patients had a primary nonresponse to open-label adalimumab,12 and certolizumab induction therapy,32 respectively. One factor leading to a higher primary lack of response in the adalimumab and certolizumab studies relates to approximately half the patients having been previously exposed to, and lost response to, infliximab.
In patients without an initial response to one anti-TNF agent, a small proportion of patients will respond to an alternative anti-TNF agent.119
Loss of response to anti-TNF therapy has been defined in some studies as an increase in the CDAI of more than 70 points from the study entry score together with a total score of greater than 175, or an increase in CDAI of 35% or more from the baseline score, or the need to introduce a new treatment for active CD.8, 21
The ACCENT 1 study demonstrated a median duration of response to a single infusion of infliximab of 19 weeks (interquartile range 10–45 weeks). In initial responders the median time to loss of response to maintenance infliximab therapy was ≈46 weeks (interquartile range, 17 to >54 weeks).8
In the ACCENT 1 and CHARM studies approximately one-third of patients on maintenance treatment lost response to infliximab and adalimumab, according to strict study definitions, over the first year of treatment.12, 74, 120, 121 In clinical practice the maintenance of clinical response is greater than these figures suggest. In a large series of 614 patients with long-term follow-up, about 60% of patients have sustained benefit, with 20% losing response to infliximab at a median follow-up of 5 years.23
A recent review assessed the frequency of loss of effect to infliximab (defined as the need for an increased infliximab dose or a decreased interval between doses) in 16 studies in which the majority of patients had received induction followed by regular maintenance therapy. The mean percentage of patients who lost response to infliximab was 37% but studies had variable follow-up times. When expressed as incidence per patient-year follow-up, the annual risk for loss of infliximab response was estimated to be 13% per patient-year.122
Factors That Influence Reduced Response or Loss of Response to Anti-TNF Therapy
Several factors, some of them interdependent variables, can influence the magnitude of the initial response and duration of response to a biological drug. These factors include the dosing interval, antibody formation, serum drug trough levels, adverse prognostic factors, and whether concomitant immunosuppressive drugs are used.
Intermittent treatment with infliximab predisposes to the formation of antibodies to infliximab, which increases the incidence of loss of response.109 In a post-hoc analysis of the ACCENT I study significantly more patients on scheduled treatment with infliximab achieved clinical remission compared to those on episodic treatment.21 Furthermore, antibodies to infliximab were found in a greater proportion of patients on episodic treatment than those on scheduled maintenance therapy.123
Infliximab reintroduction after a period of drug discontinuation (more than 4 months) in patients previously treated with a three infliximab induction regimen (at weeks 0, 2, and 6) was not associated with a higher incidence of secondary loss of response, compared to those receiving infliximab maintenance treatment.124 In this study, infliximab was restarted after a median of 13 months. Loss of response occurred in 15% of the maintenance group and 10% in the “reintroduction” group. In contrast, other studies have shown an increased rate of infusion reactions after a drug holiday. Infusion reactions have been associated with a loss of response to infliximab.109
Lastly, the STORI study, thus far published only as an abstract, demonstrated no major problem on reintroduction of infliximab for clinical disease relapse within a year of stopping it, if it was reintroduced after achieving complete endoscopic and biochemical (CRP) remission, and the patients were still on azathioprine.125
In the REACH study (Randomized, Multicenter, Open-Label Study to Evaluate the safety and Efficacy of Anti-TNF Chimeric Monoclonal Antibody in Pediatric Subjects with Moderate to Severe Crohn's Disease),120 children with CD who responded to infliximab induction were randomized to infliximab every 8 or 12 weeks. At 1 year more patients on 8-weekly maintenance therapy remained in clinical remission without dose adjustment compared with those on 12-weekly maintenance. Forty-nine percent of patients who received infliximab every 12 weeks required an intensification of their treatment schedule due to loss of response, compared to 19% who received the drug every 8 weeks. Longer dose intervals of regular scheduled infliximab are therefore associated with an increased loss of response, at least in children.
The ability of the body to mount an antibody response to biological drugs presents a therapeutic challenge. Antibody formation has been reported in 3% to 61% of patients treated with infliximab, adalimumab, and certolizumab, and is associated with decreased circulating drug levels.9, 22, 25, 32, 106, 109, 126 Since response is related to drug levels (see below), loss of response to biological therapy may relate, in some patients, to the formation of antibodies to the anti-TNF drug.
Episodic infliximab therapy, that is, retreatment on disease relapse, is associated with higher antibody levels to infliximab, and an increased rate of infusion reactions, than scheduled, regular maintenance therapy.22, 127 In the ACCENT 1 study, antibodies to infliximab were identified in 30%, 10%, and 7% of patients on episodic therapy, regular infliximab 5 mg/kg, and regular infliximab 10 mg/kg, therapy respectively.22
In a prospective cohort of 125 consecutive CD patients treated with episodic infliximab (defined as repeated infusion on relapse of disease as decided by an experienced clinician) in Leuven, a higher level of antibody to infliximab (>8 μg/mL) pre-infusion was associated with a shorter duration of clinical response to the infusion (35 versus 71 days) and higher risk of infusion reaction.109 In this study, the use of concomitant immunomodulators was associated with an increased trough concentration of infliximab 4 weeks after an infusion, and a reduced incidence of antibodies to infliximab. Patients on concomitant immunosuppressives had a significantly lower rate of antibody formation (43%) than those not on concomitant immunosuppressives (75%).
A subsequent study by the same group demonstrated that there was no difference between methotrexate and azathioprine in reducing the development of antidrug antibodies. Antibody formation appears to develop early in treatment: patients who developed antibody had lower infliximab levels 4 weeks after the first infusion than those who did not develop antibody.106
In the CLASSIC trial, antibodies to adalimumab were identified in ≈3% of patients at 1 year.24, 25 Antibodies to adalimumab were not measured in the CHARM trial.
In the PRECiSE 1 and 2 studies, antibodies to certolizumab were detected in ≈10% of patients during treatment.9, 32 Lower rates of antibody formation occurred in patients who received concomitant immunosuppressive agents (4%) than those who did not (10%).9
Trough Drug Levels
A study from Toronto demonstrated that the trough infliximab concentration correlates with subsequent clinical outcome. Compared with patients who had an undetectable trough serum infliximab level, patients with a detectable preinfusion infliximab trough level had a higher clinical remission rate, lower serum CRP level, and higher rate of endoscopic improvement at a median of 23 months follow-up.127
The infliximab concentration has been shown to correlate with the duration of response.109 However, there was not a big difference in the duration between those with high and low drug concentrations.
In a recent prospective study from Leuven and France of CD patients treated with adalimumab, 65% of patients on maintenance adalimumab therapy had to dose increase to 40 mg each week, and 39% eventually stopped adalimumab therapy due to loss of response after a median of 29 weeks. Patients who lost response to adalimumab had lower serum adalimumab trough concentrations. Patients who discontinued adalimumab due to loss of response also more frequently developed antibodies against adalimumab.126
The half-life of infliximab and adalimumab are similar—≈10 and 14 days, respectively. Given the different dosing schedules of 8-weekly and 2-weekly, respectively, the serum drug profile will differ, with greater peaks and troughs with infliximab, and a more even profile with adalimumab. Which of these profiles, if either, offers greater biological efficacy remains untested.
Combining Measurement of Serum Antibodies and Serum Trough Level (Fig. 6)
With currently available assays to evaluate antidrug antibodies the drug level has to be undetectable to be able to detect the presence of antibodies to anti-TNF biologics. Tests are in development that will allow antibody measurement in the presence of measurable drug. Conversely, drug levels can be measured in the presence of antidrug antibodies.
A combination of measurement of drug trough level and anti-drug antibody may provide useful information on which to adjust therapy. Loss of response to anti-TNF therapy in patients who have a satisfactory drug level is likely to be due to loss of biological effect of the drug, in which case an alternative drug may be needed. Patients with a low drug level and high antibody level are likely to have developed a problem related to immunogenic drug inactivation, while those with a low drug level and low antibody level may have increased drug clearance. These considerations are not absolute, as patients will differ in their response to the same drug concentrations; for example, some patients with low drug levels may still respond clinically, as seen in the SONIC study,45 while some patients with seemingly “adequate or high” levels who have unresponsive disease may well respond to even higher doses, due to their inherent biological or disease differences. Such drug adjustments according to drug and antidrug antibody levels therefore require substantial testing and validating.
Recent data from the Mayo clinic support the clinical value of antichimeric antibody and anti-TNF drug concentration measurement. In a retrospective review of the impact of drug and antidrug antibody measurement on clinical outcome in 155 patients being treated with infliximab, in patients with detected antichimeric antibodies, switching to a different anti-TNF therapy was associated with a complete or partial response in 92% of patients, while dose escalation had a response of 17%. In patients with “subtherapeutic” infliximab concentrations (measured at 4 weeks after starting therapy), dose escalation associated with a response in 86% of patients, whereas change to an alternative drug produced a response in 33%.128 Similar data are not available for adalimumab or certolizumab.
These group data on drug trough levels provide an estimate of likely response to different strategies, but need to be considered cautiously in an individual due to their poor specificity and sensitivity.127 There is substantial overlap between responders and nonresponders, and remitters and nonremitters.
Infliximab trough levels have been shown to correlate with the degree of mucosal healing. In a study published in abstract form, infliximab trough levels were measured in the serum of 210 patients with CD in whom clinical, biochemical, and endoscopic data were available before and after starting infliximab. Thirty-nine percent of patients had complete mucosal healing, 22% had partial healing, and 39% had no healing. Patients with complete or partial healing on infliximab therapy had significantly higher trough levels (median 5.00 μg/mL) than patients with no healing (median 0.95 μg/mL). These data suggest that measurement of infliximab trough levels may guide dose adjustment. For instance, in patients with a low trough level and no healing, the dose of infliximab can be increased or the dose interval reduced; in patients with a high trough level and no healing, consider switching to an alternative therapy; in those with a high trough level and mucosal healing, the drug can be continued.51
Optimal drug levels remain to be defined for the anti-TNF drugs. In particular, the relationship between different drug levels, the clinical response, and the response in relation to mucosal healing remain to be defined. Specificity and sensitivity assessment of different levels for a good outcome with dose adjustment or changing drug need to be determined. Interventions based on drug levels in relation to different therapeutic goals also need further exploration. For example, in a patient with a “therapeutic” drug level who is in clinical remission but still has mucosal disease on endoscopy, it remains to be defined whether increasing the dose leads to mucosal healing.
Group data on drug levels may be of limited value. They may be more helpful if a patient's levels are at the end of the spectrum. For example, a very high drug level and lost response suggests that biological responsiveness to the drug has been lost and the drug should be changed. A very low level should lead to checking for the presence of antidrug antibody—if the antibody level is high then an increased dose, temporarily or permanently, may be effective; if the antidrug antibody level is low then there may be a problem related to drug clearance.
A decrease in elevated baseline CRP levels to normal levels (<3 mg/L) after initiation of infliximab treatment has been shown to be associated with a better long-term outcome.23
Smoking also appears to affect the response to anti-TNF therapy, with nonsmokers demonstrating a more prolonged response than smokers (73% versus 22%).100
The current recommendation for infliximab dosing comprises induction with 5 mg/kg at 0, 2, and 6 weeks, followed by maintenance 5 mg/kg every 8 weeks. Diminished or complete loss of initial response to the drug can occur at any time after treatment begins. Regaining response can be achieved by either reducing the frequency of infusions to every 4 to 6 weeks or by increasing the dose of infliximab from 5 mg/kg to 10 mg/kg. In the ACCENT 1 study clinical response was regained in 86% of patients who had a dose increase21 (Anon., Data on file Centocor).
In one series of patients who had at least eight doses of infliximab, ≈30% and 50% of patients required dose intensification at 12 and 30 months, respectively. Most patients were able to regain response on dose intensification and continued on infliximab.129 One long-term retrospective study showed that infliximab was discontinued due to initial nonresponse in 12% and loss of response in 8% of patients. Dose-intensification was necessary in 14% of patients.44
In a large retrospective study of 614 CD patients treated with infliximab23 about half the patients treated with maintenance infliximab required at least one intervention to maintain treatment benefit at a median of 5 years. These interventions include a reduction in the interval between infusions (20%), an increase in the dose of infliximab to 10 mg/kg, and/or reinduction with infliximab infusions (26%), or an increase of the dose plus a reduction of the dose interval (4%). Overall, 72% of patients who had dose escalation and/or reinduction with infliximab, and 62% with an increase in dose and a shortening of the interval, were able to return to 8-weekly 5 mg/kg treatment. About one-third of patients with a temporary reduced interval between infusions were able to extend the treatment interval back to the original 8 weeks.23
In a subgroup analysis of the CHARM trial, during 1 year of therapy 27% of patients required dose escalation to 40 mg adalimumab every week, and approximately half of these patients maintained benefit and continued at the increased dose.130
In the open label extension certolizumab study, PRECiSE 4, the efficacy of reinduction with a single extra dose of certolizumab in patients who had experienced a significant clinical worsening of CD symptoms during continuous treatment was assessed. Patients were treated with a single additional dose of certolizumab 400 mg subcutaneously followed by standard therapy 4 weeks thereafter. At week 4, clinical response was seen in 63% of patients, which was maintained through to week 52 in 55% of patients.37
Overall, therefore, about one-third of CD patients on long-term anti-TNF therapy lose response and require dose intensification. The majority (70%–90%) can regain response with temporary or permanent dose adjustment.
In a large retrospective study of 614 patients with CD, 22% of patients lost response to infliximab and failed to respond despite dose intensification. Sixty patients were either switched to adalimumab (n = 50, 42%) or they were included in another study of treatment with certolizumab (n = 10, 8.5%).50
Adalimumab is an effective treatment in patients who have lost response to infliximab.12, 132–137 In the GAIN study, 325 patients who had lost response to, or were intolerant of, infliximab were randomized to induction treatment with adalimumab 160 mg followed 2 weeks later by 80 mg or placebo.30 Clinical remission (21% versus 7%) and response (38% versus 25%) were significantly higher in the adalimumab than placebo treatment group at 4 weeks. The observed remission and response rates were lower than rates observed in studies of anti-TNF-naïve patients.
In a long-term single-center study of 53 patients who had failed infliximab and were treated with adalimumab, the probability of maintaining a clinical response was 77%, 68%, and 51% at 26, 52, and 130 weeks, respectively. The probability of remaining free from major abdominal surgery was 82% at 6 months.
In patients with CD who have failed infliximab, therefore, about half will maintain a clinical response at 2 years when treated with adalimumab.136
In the PRECiSE 2 study, of the one-third of patients who had previously failed infliximab treatment, 44% responded to certolizumab. In the whole study population the response rate to certolizumab was higher in patients who were infliximab-naive (69%).35
In an open-label study of 539 patients who had lost response to, or become intolerant of, infliximab 62% of patients achieved a response including 39% who achieved remission at 6 weeks following three doses at weeks 0, 2, and 4 of certolizumab.36 There was no difference in the rate of response and remission at 26 weeks between patients dosed every 2 weeks or every 4 weeks.
To our knowledge there are no controlled studies on the efficacy of infliximab or certolizumab in patients who have lost response or developed intolerance to adalimumab.
There are few data on whether patients who have lost response to two anti-TNF drugs will benefit from a third anti-TNF drug. In a retrospective study 67 patients with CD who had failed two different previous anti-TNF agents patients were treated with either certolizumab or adalimumab. A short-term response was achieved in 41 of 67 (61%) and 34 of 61 (51%) patients at week 6 and week 20, respectively. At a median of 26 weeks follow-up the third anti-TNF had been stopped in 36 patients due to intolerance or failure. Two deaths were observed.138 These trials did not determine the reason for loss of response (development of antibodies or loss of response to the disease mechanism in the presence of adequate drug levels). While a third anti-TNF antibody may provide benefit for patients who develop immunogenicity, it is less likely that patients who lose response despite adequate blood levels will respond and the risks need to be carefully weighed against the benefit. Finally, it is possible that patients who develop immunogenicity to one biologic may be more likely to develop antibodies to a second agent; optimization with induction therapy and the addition of an immune suppressant (or pretreatment with corticosteroids) may be of greater value in this group of difficult-to-manage patients.
Currently, the only licensed alternative biological drug to anti-TNF drugs for the treatment of CD is the antiadhesion molecule natalizumab. It is effective as an induction and maintenance therapy for up to 2 years of therapy (ENACT 1 and 2; ENCORE).139, 140 Efficacy has been demonstrated in patients with an elevated CRP at baseline, patients with active disease despite the use of immunosuppressant, and patients who had previously been treated with an anti-TNF drug. In patients who had previously had an anti-TNF therapy more patients on natalizumab responded at week 10 compared with those on placebo (55% versus 35%).139
In the ENCORE study, 509 patients with moderate-to-severely active CD and active inflammation defined by an elevated CRP were randomized to natalizumab 300 mg or placebo intravenously at weeks 0, 4, and 8. The response rate at 8 weeks was significantly higher in patients on natalizumab compared to placebo (48% versus 32%).140
Switch to Simple Molecules
In the current climate most patients with luminal CD progress to an anti-TNF drug having failed simple immunosuppressive therapy. Some patients, however, will be faced with the need to return to immunosuppressive therapy having failed treatment with one or more anti-TNF therapies. Such simple immunosuppressive therapies have established efficacy, although few have been evaluated specifically in patients who have failed anti-TNF therapy. Such therapies include tacrolimus,141, 142 and thalidomide. In a small study of patients with CD refractory to infliximab treated with oral tacrolimus, remission was achieved in 42% of patients.143 In one retrospective study, six of eight patients treated for refractory CD to thalidomide at a median follow-up of 12 months.144
Despite the requisite intravenous infusions for infliximab therapy, up to one-third of patients may be nonadherent with maintenance therapy. Nonadherence is associated with worse outcomes and increased costs.145 Certolizumab and adalimumab can be self-administered by the patient or a caregiver. When failure with these medications occurs, compliance with drug treatment should be considered as one of the potential causes.
A small proportion of patients do not respond to a first anti-TNF agent. Use of a second antibody in some of these primary nonresponders is sometimes effective. For patients who experience a decrease or lost response to therapy, development of immunogenicity and compliance should be considered, and complications excluded. A dose intensification strategy can be employed, either as an increase in dose or a decrease in dose-interval; this may be more effective in patients who have not developed antibodies to the biologic therapy. In patients who have not benefited from dose intensification and, in particular, those who have developed antibodies to a biologic, an alternative anti-TNF therapy can be considered. A third anti-TNF agent is sometimes effective, but the proportion who benefit with each successive drug decreases. An alternative approach, available in the U.S., to loss of response is to use natalizumab. Other options are the use of simple molecule immunosuppressives and surgery.
STOPPING ANTI-TNF THERAPY
A challenging aspect in the use of anti-TNF therapy, in patients who are benefiting from therapy without major side-effects, is whether anti-TNF therapy can or should be stopped, when to stop, and how to stop.
In patients who have had multiple operations, or failed previous therapies, and are well and in remission on biological therapy, the available data on efficacy and safety support their continued use. However, a patient may wish to stop therapy for a variety of reasons, including concern about the safety of long-term therapy, women wishing to conceive, the need or desire to travel to areas of high endemic infection risk, or issues related to cost. The physician may also feel that biological therapy can be ceased due to sustained remission; such a decision will be made on the basis of evidence of the extent of the patient's sustained and complete remission as well as past history in relation to previous therapies and surgery. The decision-making approach to stopping anti-TNF therapy is summarized in Figure 7.
If the combined decision between the physician and patient is to stop biological therapy, this approach is most likely to be successful, with remission being maintained if complete clinical, endoscopic, and biochemical remission have been achieved prior to stopping (Fig. 8).
In a recent retrospective study from three Italian centers, infliximab was discontinued in 60% of patients due to prolonged remission off corticosteroids. About 78% half these patients maintained sustained clinical remission at a median follow-up of 25 months. Mucosal healing was a predictor of sustained clinical benefit after infliximab withdrawal. Ten of 11 patients who relapsed after infliximab discontinuation regained remission on retreatment with infliximab.40
In the STORI study, from the GETAID group, infliximab therapy was ceased in 115 patients treated with combined scheduled infliximab and an immunosuppressant for at least 1 year between 2006 and 2008. Patients stopped infliximab if they had been in clinical remission without the need for corticosteroids for at least 6 months and were maintained on a stable dose of immunosuppressant. The median duration of CD was 8 years. Forty-five percent of patient relapsed on withdrawal of infliximab. Factors which predicted time-to-relapse included male gender, steroid treatment in the previous 6–12 months, hemoglobin below 14.5 g/dL, raised white cell count, raised CRP, and mucosal lesions. Of the patients in whom there was endoscopic mucosal healing and a normal CRP more than 80% maintained remission after ceasing infliximab after a period of 10 months. Fecal calprotectin was a further discriminator for benefit. Fecal calprotectin level was higher in patients who relapsed than in those who sustained remission among all those who had mucosal healing and normal CRP. Mucosal healing and a normal CRP are therefore clear predictors of sustained response when stopping infliximab but remaining on a thiopurine. Fecal calprotectin provides further evidence of epithelial integrity and low epithelial inflammation. In those in whom calprotectin was still raised, presumably there was active disease proximal to the observed healed disease in the ileum and colon (although patients were only included if they had colonoscopically reachable ileal and colonic disease extent) or endoscopic mucosal normality was associated with persistent epithelial inflammatory activity.125
In the STORI study patients who relapsed on infliximab withdrawal could be reinduced and maintained with infliximab. Eighty-eight percent of these patients regained remission at 4 months.125
Patients in remission on anti-TNF therapy can be maintained on it long-term unless they lose response, develop side-effects or malignancy, want to travel, plan to become resident in a tuberculosis endemic area, or do not want to continue treatment. Patients who are in remission but who have had multiple previous operations, intolerance to multiple drugs, or difficult to control disease should probably not stop anti-TNF therapy.
Conversely, in patients who have achieved clinical remission with biological therapy, the drug can potentially be stopped and the patient transitioned to an immunosuppressive therapy if patients have endoscopic evidence of healing and normal CRP, and no other adverse prognostic factors such as extensive small bowel disease, smoking history, or perianal fistulizing disease. Elevated CRP is associated with an increased risk of relapse.146 Mucosal healing and a normal CRP should be aimed for before considering cessation of biologic therapy.
Other methods to detect mucosal healing in the small bowel have not yet been evaluated in relation to ceasing anti-TNF therapy, but might include techniques such as MR-enteroclysis.147
SPECIFIC CLINICAL SITUATIONS
Pregnancy and Breast-feeding
Studies reporting the intentional use of infliximab and adalimumab during pregnancy have reported no increased incidence of stillbirths, ectopic pregnancies, spontaneous abortion, or low birthweight infant. The rate of congenital malformations with anti-TNF agents was within the same range as nondisease controls.148, 149 Both the U.S. FDA and the ECCO have categorized anti-TNF agents as safe during pregnancy.
However, the implications of exposure of anti-TNF agents on the newborn are unknown. Infliximab and adalimumab have been shown to cross the placenta barrier during the late second and third trimesters. On this basis some have suggested delaying treatment in the last trimester till after delivery, to decrease levels in the infant.150
Animal data suggest that the pegylated antibody fragment of certolizumab does not cross the placenta.
Developmental milestones at 9 months of age in the newborn infants of patients treated with infliximab, adalimumab, or certolizumab have been shown to not be delayed.151
Infliximab has not been detected in breastmik, and is considered safe during breastfeeding.152
Anti-TNF therapy is generally regarded as contraindicated in patients with CD and strictures. The extent to which such therapy should be avoided may relate to the degree of narrowing and the extent of inflammation (versus fibrosis) associated with the stricture. Strictures that are very narrow and associated with deep ulceration may be more prone to obstruction as part of the rapid healing induced by therapy. There is no evidence that anti-TNF drugs improve fibrotic strictures.
Few studies have evaluated the efficacy of anti-TNF in patients with symptomatic strictures. In a retrospective study of 18 symptomatic patients treated with infliximab for symptomatic small bowel strictures after failure of conventional treatment, complete response, partial response, and failure were observed in 10, 7, and 1 patients, respectively, at 8 weeks. At a median follow-up of 18 months, there were five patients with complete response, 10 with partial response, and three patients failed therapy.153
In a retrospective analysis of infliximab therapy in 21 patients with inflammatory Crohn's strictures nine responded well and became asymptomatic.154
Anti-TNF therapy may be an option in patients with inflammatory strictures in CD after careful assessment of the nature of stenosis and the exclusion of sepsis. In some patients with known strictures in whom therapy is essential due to uncontrolled disease, anti-TNF therapy can be initiated in the expectation that surgery may be required if obstruction develops, with the drug continued after surgical treatment. Further prospective studies in this patient group are required to test drug efficacy and safety.
Postoperative Luminal Disease
Accumulating data suggest that anti-TNF therapy is effective in preventing postsurgical recurrence of CD. However, it is unclear whether long-term therapy is necessary.
A pilot randomized controlled trial demonstrated that infliximab is effective in preventing postoperative enoscopically identified CD recurrence. Twenty-four patients who had undergone ileocolonic resection were randomized to receive infliximab or placebo within 4 weeks of surgery and to continue infliximab every 8 weeks for 1 year. The rate of endoscopic recurrence at 1 year after surgery was significantly lower in the infliximab group (1 of 11; 9%) compared with the placebo group (11 of 13; 85%). Histologic recurrence was also lower in the infliximab group (27%) compared with the placebo group (85%).155 A higher proportion of patients were in clinical remission in the infliximab group (80%) than the placebo group (54%). In the following 12 months, 7 of 13 patients who had received placebo initially received infliximab and 71% were in remission at 2-year follow-up.156
Although these results are promising, the use of infliximab routinely to prevent recurrence in all patients after ileocecal resection is likely to result in overtreatment of some patients who are at low risk of recurrence. There are currently no studies that have assessed the use of anti-TNF agents in selected high-risk patients, including those who have failed an immunosuppressive drug. A well-defined strategy involving stratification of patients into low- or high-risk groups based on existing risk factors and endoscopic findings postsurgery may help to identify patients who would benefit most from a step-up in immunosuppressive or anti-TNF therapy, although such an approach remains to be prospectively validated.157
In a prospective long-term cohort study, 12 consecutive patients with CD who were treated with maintenance infliximab 5 mg/kg immediately after surgery and who had no clinical or endoscopic evidence of disease recurrence after 24 months were followed up for an additional 1 year. Infliximab treatment was then discontinued. None of the patients had clinical or endoscopic recurrence 3 years after surgery. After 3 years infliximab was discontinued; 4 months later 10 of 12 patients (83%) had developed endoscopic recurrence. All 10 patients were retreated with infliximab at a dose of 3 mg/kg every 8 weeks, with restored and maintained mucosal integrity for 1 year. These data suggest that long-term maintenance therapy may be required to maintain mucosal integrity in patients after surgery for CD.158
The availability of anti-TNF therapy has led to major advances in the treatment of CD. Well-designed and conducted trials and long-term population-based observational studies provide excellent data on the efficacy and safety of these drugs. Strategies should now be applied to gain maximum benefit from this therapy.
This work was supported by an unrestricted educational grant from Abbott. Abbott had no input to the manuscript.