The second European evidence-based Consensus on the diagnosis and management of Crohn's disease: Current management

5.2.1. Mildly active localised ileocaecal Crohn's disease 

Although the stage at which immunosuppressive and biological therapy is introduced is changing, it is important to remember that an appreciable proportion of patients with CD have a mild pattern of disease. Thus, in an inception cohort of 843 patients with CD (the IBSEN cohort), diagnosed between 1990 and 1994, only a quarter of the patients were treated with immunomodulators and 4% with anti-TNF agents during the first ten years of follow-up.3 In another cohort from Olmsted County, Minnesota, USA, 43% of patients were never treated with steroids.4 Finally, among patients diagnosed and followed up at private hospitals in Germany for a median 39months, 27% of patients had mild disease that did not need steroids.5 Despite this, the majority of patients with active CD have symptoms that merit treatment.

Budesonide 9mg daily is the favoured therapy to induce remission in mildly active, localised ileocaecal Crohn's disease, because it is superior to both placebo (relative risk (RR) 1.96, 95% CI 1.19–3.23) and mesalazine (RR 1.63; 95% CI 1.23–2.16).6 Budesonide is preferred to prednisolone for mild disease because it is associated with fewer side-effects (RR 0.64, 95% CI 0.54–0.76). However, budesonide is significantly less effective than conventional steroids for induction of remission (RR 0.86, 95% CI 0.76–0.98), particularly among patients with severe disease (CDAI>300) (RR 0.52, 95% CI 0.28–0.95). In individual studies, budesonide achieves remission in 51–60% over 8–10weeks.7, 8, 9, 10, 11, 12

Nevertheless, a recent study on budesonide (Budenofalk®) compared to mesalazine for active Crohn's disease, published in abstract form only, found no difference between the two treatments.13 Remission rates of 69.5% for budesonide and 62.1% for mesalazine were observed in the ITT population in this study. A clinically relevant and statistically insignificant CDAI drop of 100 points was observed in 89% of budesonide-treated patients and in 79% of mesalazine-treated patients. In patients with mild disease (CDAI<300 points) both treatments appeared to be equally effective in this trial. This study was presented in abstract form only after the Consensus meeting in Vienna in 2008 and these preliminary data are in contrast to a previous meta-analysis.14 This meta-analysis showed no clinically significant effect of mesalazine in the management of mild to moderately active ileocaecal Crohn's disease compared to placebo, although it found a significant reduction in the CDAI in patients with active ileocaecal CD receiving ethylcellulose-coated mesalazine 4g/day. Since the drop in CDAI was just 18 points compared to placebo (−63 vs −45, p=0.04) in 615 patients, the clinical benefit is considered marginal. Lower doses of mesalazine cannot be recommended for active CD. However, the conflicting new data implicate that mesalazine deserves further evaluation for the treatment of mildly active CD. A further study of high-dose (6g daily) mesalazine for active CD is currently under way. Future meta-analyses should incorporate more recent studies with high-dose formulations.

Antibiotics (metronidazole, ciprofloxacin), with or without mesalazine, are not recommended, because side-effects are commonplace. The same applies to nutritional therapy, which is often poorly tolerated by adults, although there are case series or small trials that have shown these treatments to be modestly effective.

5.2.2. Moderately active localised ileocaecal Crohn's disease 

For moderately active CD, either budesonide or prednisolone are appropriate initial induction therapies. Prednisolone is highly effective, but more commonly causes side-effects than budesonide.6 In a systematic (Cochrane) review of conventional corticosteroids, two studies compared corticosteroids to placebo and six studies compared corticosteroids to 5-ASA.15 Corticosteroids were found to be significantly more effective than placebo at inducing remission in CD (RR 1.99; 95% CI 1.51–2.64; p<0.00001). Interestingly, there was no difference in the proportion of patients experiencing adverse events with steroids compared to high-dose 5-ASA, and steroids did not induce more study withdrawals due to adverse events than either placebo or 5-ASA. In addition, prednisolone is less expensive than budesonide if cost is an important consideration. The dose of prednisolone is adjusted to the therapeutic response over a period of weeks (see below). A more rapid dose reduction can be associated with early relapse. The Consensus does not favour sole nutritional therapy (5.2.1, 5.4.9), antibiotics (unless septic complications are suspected), or surgery for moderately active ileal CD as first-line therapy.

Particular effort should be made to minimise corticosteroid exposure in CD, even though steroids remain (in 2009) the mainstay for treating active disease. Part of the problem is a complete lack of efficacy for maintaining remission (see Section 6.0). No more than one in four patients given corticosteroids to induce symptomatic remission will still be in remission after a year, even if patients' treatment with immunomodulators is included.16

An effective approach to minimizing steroid therapy is the early introduction of anti-TNF agents. Selection of patients appropriate for biological therapy depends on clinical characteristics, previous response to other medical therapies, phenotype and co-morbid conditions. Certain patient populations may derive greater benefit from the early introduction of biological therapy, including steroid-refractory (Section 5.3.3) or steroid-dependent patients.17 However, a study of 133 patients with active Crohn's disease who had not previously received glucocorticoids, antimetabolites, or infliximab also suggested benefit of early biological therapy in this relatively treatment naïve group. This trial randomized patients to either early combined immunosuppression or conventional treatment (commonly referred to as the Step Up/Top Down study).18 At week 26, 60.0% of 65 patients in the combined immunosuppression group were in remission without corticosteroids and without surgical resection, compared with 35.9% of 64 controls, giving an absolute difference of 24.1% (95% CI 7.3–40.8, p=0.006). It has now been established (through the SONIC study) that combination treatment with infliximab and azathioprine is more effective than infliximab alone for achieving (and maintaining) steroid-free remission in patients at an early stage of disease.1 This is addressed in the section on maintaining remission, although the distinction between induction and maintenance therapy is largely one of convenience, since there should be a seamless transition in individual patients. Evidence for the efficacy of individual anti-TNF agents is covered in Section 5.4.4.

5.2.3. Severely active localised ileocaecal Crohn's disease 

The initial treatment of severe ileal CD still includes prednisolone or intravenous hydrocortisone. A substantial change in the therapeutic approach in the past 5years has been the recognition that it is potentially possible to use clinical criteria at diagnosis to predict the subsequent course of disease (Section 5.3). This, in turn, has affected the threshold for introducing anti-TNF and immunomodulator therapy in patients with markers of poor prognosis. Given that continued treatment with either infliximab or adalimumab has been associated with a substantial reduction (about 30% at 12months) in surgery and hospitalization for CD,19, 20 the threshold is likely to decrease further. Nevertheless, there are no data that specifically apply to localised ileocaecal disease.

Anti-TNF therapy is still best reserved for patients not responding to initial therapy and for whom surgery is considered inappropriate. However, this does not mean that surgery takes precedence over adalimumab, infliximab, or certolizumab pegol (the latter is not currently licensed for CD in Europe), and the therapeutic strategy for an individual should be a joint decision between patient, physician and surgeon. Although anti-TNF therapy may reduce the need for surgical resection, the threshold for surgery in localised ileocaecal disease is lower than for disease elsewhere. Indeed, some experts still advocate surgery (especially laparoscopic-assisted resection, Section 7.2.6) in preference to anti-TNF therapy for disease in this location. Others advocate resection if medical therapy is not effective within 2–6weeks. It is now clear when starting anti-TNF therapy in patients with CD naïve to immunosuppression, that combination therapy with infliximab and azathioprine is more effective than either alone, whether for induction of remission, maintenance of remission up to 1year, or for mucosal healing.1 However, only patients with an elevated serum CRP or the presence of mucosal lesions at colonoscopy gained additional benefit from infliximab therapy. The combination of infliximab and azathioprine was not superior to infliximab alone in the subgroup of patients with active signs of inflammation. It is unknown whether combination therapy with anti-TNF agents other than infliximab would also improve outcome in patients naïve to immunosuppressives other than steroids.

It may sometimes be difficult to distinguish between active disease and a septic complication, but antibiotics should be reserved for patients with a temperature or focal tenderness, or in whom imaging has indicated an abscess. Adding ciprofloxacin and metronidazole to budesonide has shown no advantage over budesonide alone in active Crohn's disease.21

5.2.4. Colonic disease 

It is easier to confirm the activity and severity of colonic CD than it is for isolated small bowel disease, even though active ileal disease is accessible to ileocolonoscopy in the large majority of patients. This may explain why colonic disease appears to respond better to anti-TNF therapy than ileal disease.22 Systemic corticosteroids such as prednisolone or equivalent are effective,23, 24 whereas budesonide, in its current formulation, has no role in treating colonic disease, unless it primarily affects the proximal colon. Therefore steroids remain first-line therapy, with immunomodulators as steroid-sparing agents for those who have relapsed. As with disease in any location, the decision needs to take account the previous response to therapy and the pattern of disease: if there is infrequent relapse and a previous rapid response to steroids, then it is reasonable to take this conventional approach.

On the other hand, it is important that the expectations of gastroenterologists and their patients are appropriate: it is no longer acceptable for patients to be subjected to recurrent cycles of steroids when effective therapy for achieving and maintaining steroid-free remission with anti-TNF therapy exists. If symptoms persist in spite of steroids (with or without immunomodulators), disease activity should be assessed endoscopically and anti-TNF therapy commenced if activity is demonstrated (Section 5.3.2). If patients do not respond or lose response to anti-TNF therapy, then surgery is generally appropriate (Section 7.3). Occasionally colonic disease is so severe and aggressive (often in combination with perianal sepsis) that surgery to defunction the colon is necessary for symptom control before anti-TNF therapy can be used safely.

The use of sulfasalazine, metronidazole,25 or nutritional therapy 26 for adults with colonic CD has almost been consigned to history. Sulfasalazine 4g daily is modestly effective for active colonic disease,23, 24 but it cannot be recommended in view of a high incidence of side-effects. There is no evidence that mesalazine is effective for active colonic CD, but opinion still varies about the value of topical mesalazine as adjunctive therapy in left-sided colonic CD. Topical mesalazine can be considered in distal colonic CD, but a similar proportion advise or recommend it as do not use it.

5.2.5. Extensive small bowel disease 

The inflammatory burden is greater in extensive (>100cm) than in localised small bowel disease, often resulting in nutritional deficiencies. Treatment with steroids and the early introduction of concomitant immunomodulators (for their steroid-sparing effect) is considered appropriate. Nutritional support should be given as an adjunct to other treatment, and may be considered as primary therapy if disease is mild.26 However, early introduction of anti-TNF therapy should also be considered, especially in who have clinical indicators of poor prognosis (Section 5.3), as several analyses have shown that anti-TNF therapy is more effective when treatment is initiated early in the disease. Thus, in the CHARM trial with adalimumab, clinical remission rates approached 60% in patients who had CD for <2years, compared to 40% (p<0.05) in patients who had a longer duration of disease.27 A similar phenomenon was observed in patients who received infliximab as first-line treatment, for whom >90% of patients had a clinical response after the first administration.18 The same is also true for certolizumab pegol.28 However, the most compelling evidence in favour of early intervention, comes from a pilot trial in the post-operative phase of CD: 10/11 (91%) of patients treated with infliximab after ileocolic resection had no endoscopic recurrence after 1year,29 compared to 2/13 (15%, p=0.0006) treated with placebo infusions. It is important to note that these data are taken from studies that either included only a small number, or formally excluded patients with extensive disease. However, common sense indicates that the management of patients with extensive small bowel disease should be more aggressive given the well documented adverse consequences.30

In patients with extensive small bowel disease, surgical resection risks creating a short bowel. However, nutritional support with or without anti-TNF therapy prior to multiple strictureplasty is a valid strategy. Surgery, especially strictureplasty, is more appropriate for long-standing, isolated and fixed strictures. Careful consideration should be given to the optimal approach to preventing recurrence (Section 8). Anti-TNF therapy or conventional immunomodulators may be appropriate, depending on the time interval from previous surgery, or the time from diagnosis to first surgery as these factors may predict the aggressiveness of CD that post-operative therapy is designed to modify.

5.2.6. Oesophageal and gastroduodenal disease 

Upper gastrointestinal (GI) tract inflammation in Crohn's disease is increasingly diagnosed as patients more frequently undergo upper GI endoscopy. There may be no localizing symptoms and although the Montréal classification identifies upper GI involvement as a subgroup, independent of other locations, a consensus on what qualifies as significant ‘involvement’ is lacking. Reported incidence data vary considerably depending on the definitions used and the population studied. Paediatric data suggest that upper GI endoscopy is useful in differentiating CD from ulcerative colitis when inflammation is otherwise predominantly confined to the colon; however, this question has yet to be studied in adults.31 Controlled trials of individual therapies are lacking despite the finding that Crohn's disease in the proximal gut is associated with a worse prognosis.32 Evidence-based therapy is mainly derived from case series.33, 34 Most would add a proton pump inhibitor to conventional induction therapy and have a lower threshold for starting anti-TNF therapy than for disease elsewhere, given the poor prognosis.

5.3.1. Treatment of relapse compared to newly diagnosed disease 

The initial treatment of relapse should be based upon previously successful therapies. However, consideration should be given to other factors including patient preference (adverse effects, necessary speed of response, convenience, etc), the time to relapse, concurrent therapy (whether a relapse occurred during treatment with immunomodulators) and adherence to therapy.

5.3.2. Early relapse 

Any patient who has an early relapse (defined as an arbitrary period of <3months) should be started on an immunomodulator to reduce the risk of a further relapse. Opinion remains divided whether to use the same treatment to induce remission and taper more slowly or use more potent induction therapy. It is important to confirm disease activity as a cause of recurrent symptoms, although unnecessary to re-evaluate the distribution of disease unless this will alter medical or surgical management. Patients who have a relapse of moderate or severe activity should be considered for anti-TNF therapy, since infliximab is more effective than azathioprine in early (duration <2years), treatment-naïve patients with CD and there is a significant advantage in using the combination of infliximab and azathioprine.1 All anti-TNF agents are more effective when introduced at an early stage (as discussed above).

5.3.3. Steroid-refractory Crohn's disease 

For active CD that is refractory to steroids, local complications (such as an abscess) should be excluded by appropriate imaging and other causes of persistent symptoms considered. If active CD is confirmed, anti-TNF therapy is appropriate. If patients with CD are naïve to immunosuppression, treatment can follow the guidance in Section 5.2.3; see also Section 6.2.7. For patients with established CD who have active disease despite therapy with immunomodulators, post-hoc subgroup analyses of the major trials with all three anti-TNF agents have not demonstrated significant differences in efficacy between patients receiving the biologic plus concomitant immunomodulator and those treated with the biologic alone. It must, however, be remembered that these are subgroup analyses in patients that had already failed immunomodulator therapy and the studies were not designed to answer this question. Current data suggest that the effect of continuing immunomodulator therapy on reducing immunogenicity in patients receiving biologic therapy is more pronounced in patients undergoing episodic biologic therapy — a strategy that has largely been abandoned where possible, due to lesser efficacy. It is reasonable to conclude that in the majority of patients given biologics, immunosuppression should not be continued for the sole reason of decreasing antibody production, although anti-drug antibodies are not the only factor that governs the immunogenicity of a compound.40

It is also possible that the combination of steroids with an anti-TNF agent and an immunomodulator may improve outcome. In a randomized, double-blind, placebo-controlled trial, patients who had initiated corticosteroids within the last 6weeks were randomized 1:1 to receive infliximab and placebo (n=63), or infliximab and methotrexate 25mg subcutaneously each week (n=63).41 At week 14, there were no differences in the percentage of patients in steroid-free remission between the 2 groups (76% and 77%). Although this can be interpreted as a failure of methotrexate to offer additional benefit to infliximab, the very high rate of steroid-free remission (twice that seen in other studies) is notable.

The timing of surgery depends on the severity of symptoms, inflammatory burden and considerations above (Sections 5.2.3 and 5.2.4). The patient's views and extent of disease should also be taken into account. Nutritional therapy is an appropriate adjunctive, but not sole, therapy.

5.4.1. Aminosalicylates 

5.4.2. Antibiotics 

5.4.3. Corticosteroids 

5.4.4. Anti-TNF strategies 

Infliximab (Remicade®) and adalimumab (Humira®) are IgG1 anti-TNF monoclonal antibodies with potent anti-inflammatory effects, possibly dependent on apoptosis of inflammatory cells. Certolizumab Pegol (Cimzia®) is a pegylated anti-TNF Fab-antibody fragment with proven clinical efficacy despite the lack of pro-apoptotic effects. Numerous controlled trials have demonstrated efficacy of these anti-TNF agents for active Crohn's disease. Anti-TNF therapy is effective for active inflammatory CD, but should be used with care in patients with obstructive symptoms.

5.4.5. Other biological therapy 

Many new biological therapies are under development.78 The most promising novel class of agents for the treatment of Crohn's disease are selective anti-adhesion molecules. Natalizumab is a humanized monoclonal antibody against alpha4 integrin that inhibits leukocyte adhesion and migration into inflamed tissue. In ENACT-1, 905 patients were randomly assigned to receive 300mg of natalizumab or placebo at weeks 0, 4, and 8.79 The natalizumab and placebo groups had similar rates of response (56% and 49%, respectively, p=0.05) and remission (37% and 30%, respectively; p=0.12) at 10weeks. In contrast, the ENCORE trial evaluated the efficacy of natalizumab 300mg IV versus placebo at week 0–2–4 in 509 patients with moderately to severely active Crohn's disease and an increased baseline CRP. Clinical response was better in natalizumab patients (48% vs. 32%, p<0.001) as was sustained clinical remission. Of note, patients with previous exposure to infliximab responded equally well.80 Natalizumab was much more effective as maintenance therapy although the drug is only approved for the treatment of anti-TNF refractory Crohn's disease in the USA (see Section 6.2.8). Another selective anti-adhesion molecule agent, alicaforsen (anti-sense oligonucleotide to human ICAM1), has not shown benefit for active Crohn's disease at the doses used in clinical trials. Efficacy data on monoclonal antibodies against interferon-γ (Fontolizumab),81, 82 IL12/23 p40 (ABT-874, Ustekinumab)83, 84 and IL-685 have been presented (for a review, see78). Treatment by parenteral administration of IL-10 and IL-11 is ineffective, although mucosal delivery systems are being developed.86 The efficacy and safety of other novel approaches, such as stem cell transplantation87 have yet to be established.

5.4.6. Thiopurines 

Azathioprine (AZA) 1.5–2.5mg/kg/day or mercaptopurine (MP) 0.75–1.5mg/kg/day (unlicensed for use in IBD) may be used in active CD as adjunctive therapy or steroid-sparing agent. However, its slow onset of action precludes its use as a sole therapy for active disease. Purine antimetabolites inhibit ribonucleotide synthesis, but at least one mechanism of immunomodulation is to induce T-cell apoptosis by modulating cell (Rac1) signalling.87 Azathioprine is metabolised to mercaptopurine and subsequently to 6-thioguanine nucleotides. Thioguanine is discussed in the section on maintenance therapy. Since the main role of thiopurine therapy resides in maintaining remission, dose, monitoring and side effects will be discussed in the maintenance section of this paper.

5.4.7. Methotrexate 

Methotrexate 25mg/day (oral, subcutaneous or intramuscular injection — unlicensed for use in IBD) may be used in a similar fashion to thiopurines. Polyglutamated metabolites of methotrexate inhibit dihydrofolate reductase, but this cytotoxic effect does not explain its anti-inflammatory effect and inhibition of cytokine and eicosanoid synthesis with modification of adenosine levels probably contribute more.

5.4.8. Other immunomodulators 

5.4.9. Nutritional therapy 

6.1.1. General recommendations 

In view of the adverse effects of cigarette smoking on the course of Crohn's disease, smoking should be discouraged in all patients. Data from observational studies show that smoking increases the need for steroids, immunosuppressants and operations. Conversely, smoking cessation may improve the course of the disease110, 111, 112 [EL2b]. Active programs of smoking addiction should be recommended.

The absolute requirement and choice of medication for prevention of relapse in patients with medically induced remission should take into account three main factors: the course of the disease (initial presentation, frequency, and severity of flares); the extent of disease (localised or extensive — see Sections 1.112 and 1.1.13); and the effectiveness and tolerance of treatments previously used for induction of remission or maintenance. Other factors such as the presence of biological or endoscopic signs of inflammation and the potential for complications should also be considered. In addition, there may be other constraints (logistic, social, or financial) that impact on treatment choices. Finally, patients should be encouraged to participate to the decision-making process.

Patients in remission should be clinically assessed on a regular basis. Although monitoring of the C-reactive protein is frequently performed, the consequences for adjusting treatment remain unclear. Some also recommend imaging or endoscopy, but repetition of these procedures is not recommended routinely, but only in specific situations.

6.1.2. First presentation of localised disease 

There is no evidence that mesalazine is useful for maintaining medically induced remission, as the results of meta-analysis are inconsistent (see Section 6.2.1). Some consider that no maintenance treatment is an option after the first flare. Taking into account the high risk of relapse and of steroid dependence, and the higher success rate when introduced early, azathioprine is favoured if remission has been achieved with systemic steroids (see Section 6.2.4). Mercaptopurine (1–1.5mg/kg/day) can be tried in patients intolerant of azathioprine (except in cases of pancreatitis and cytopenia).113 Methotrexate is an alternative, especially for patients intolerant of thiopurines (Section 6.2.5).

6.1.3. Relapse of localised disease 

If a relapse occurs, azathioprine should be considered (see Section 6.2.4). Corticosteroids (including budesonide) are not effective for maintenance of remission, and the long-term use of corticosteroids is associated with unacceptable side effects, especially osteoporosis. Budesonide increases the time to relapse but is not effective at maintaining remission for 1year; bone loss is less, but not eliminated (see Section 6.2.3).

6.1.4. Extensive disease 

Taking into account the risks of relapse and the higher success rate when introduced early, azathioprine is recommended in patients with extensive Crohn's disease (see Section 6.2.4)

6.1.5. Steroid-dependent Crohn's disease 

Immunomodulators (azathioprine/mercaptopurine, methotrexate) are effective in steroid-dependent Crohn's disease (NNT 3).14, 93 Ileal resection is an alternative for those with localised disease depending on other disease characteristics (see Surgery for Crohn's disease Section). A very effective approach to spare steroids is the early introduction of anti-TNF agents. Selection of patients appropriate for biological therapy depends on clinical characteristics and previous response to other medical therapies. Steroid-dependent patients may derive greater benefit from the early introduction of biological therapy.17 However, a study of 133 patients with active Crohn's disease who had not previously received glucocorticoids, antimetabolites, or infliximab also suggested benefit of early biological therapy in this relatively treatment naïve group. This trial randomized patients to either early combined immunosuppression or conventional treatment (commonly referred to as the Step Up/Top Down study).18 At week 52, 61.5% of patients in the combined immunosuppression group were in remission without corticosteroids and without surgical resection compared with 42.2% in the control group (absolute difference 19.3%, 95% CI 2.4–36.3, p=0.028). It has now been established (through the SONIC study) that combination treatment with infliximab and azathioprine is more effective than infliximab alone for maintaining steroid-free remission in patients at an early stage of disease.1

6.1.5. Relapse while on azathioprine 

Patients receiving azathioprine or mercaptopurine who relapse whilst on standard maintenance doses can have their dose escalated (>2.5mg/kg/day or >1.5mg/kg respectively) until leucopenia occurs [EL3, RG D], or according to 6-TGN concentrations [EL2a, RG B] (see Section 5.4.6). Methotrexate is another option [EL1b, RG B] (see Section 6.2.5). Anti-TNF therapy has also proven to be effective in this setting [EL1a, RGA] (see Section 6.2.7).

6.1.6. Maintenance after induction of remission with Anti-TNF therapy 

Patients in a scheduled-treatment strategy with regular infliximab, appear to fare better for many (but not all) clinical endpoints, compared to patients in an episodic (on-demand) strategy [EL1b]. Concomitant immunosuppressant therapy (thiopurines, methotrexate) with anti-TNF agents is not associated with better clinical efficacy in patients who have already failed these drugs [EL1b]. However, combination of infliximab plus azathioprine is of greater efficacy in achieving and maintaining steroid-free remission than infliximab monotherapy or azathioprine monotherapy in patients naïve to both therapies [EL1b] (see Section 6.2.7).

6.1.7. Duration of maintenance treatment 

A double-blind placebo-controlled non-inferiority study comparing azathioprine withdrawal with its continuation in patients on azathioprine for more than >3.5years found that the rates of relapse after 18months were 21% and 8%, respectively117 (see Section 6.2.4). The hypothesis that azathioprine was inferior to placebo was not rejected. Long-term evaluation of these patients has been recently reported.118 The median follow-up time after azathioprine interruption was 54months; 32 of 66 patients had a relapse. The cumulative probabilities of relapse at 1, 3, and 5years were 14%, 53% and 63%. Among the 32 relapsing patients, 23 were retreated by AZA alone, all but 1 achieved successful remission. Thiopurine therapy has been associated with an increased risk of non-Hodgkin's lymphoma. 119, 120 Lewis et al. 121 conducted a decision analysis study using a Markov model. They concluded that azathioprine results in increased quality-adjusted life expectancy, especially in young patients who have the lowest baseline risk of lymphoma and the greatest life expectancy in the absence of Crohn's disease related death. The benefits of treatment exceed an increase in lymphoma risk postulated by the most extreme studies.

Long-term follow-up of Crohn's disease patients taking methotrexate does not demonstrate an increase risk of severe hepatotoxicity, as previously suggested in other diseases.97 In two series, methotrexate withdrawal in patients maintained for several years with this drug was associated with a high proportion of relapse.121, 122

The benefit of continuing an immunosuppressant such as azathioprine or methotrexate in combination with anti-TNF is discussed in Section 6.2.7. The question of whether treatment with anti-TNF agents can be safely interrupted after a period of prolonged remission is of great interest to patients and physicians. Recently, interim results of a prospective study conducted to assess the risk of relapse after infliximab discontinuation in patients on combined maintenance with immunosuppressant therapy were presented.123 115 patients with luminal Crohn's disease treated for at least one year with scheduled infliximab combined with azathioprine or methotrexate and in stable remission without steroids for at least 6months were prospectively recruited into the study. Infliximab therapy was withdrawn, and after the last infusion immunosuppressant therapy was kept at a stable dose. After a median follow-up time of 12months, 45 relapses were observed. A subgroup of patients with very low risk of relapse could be identified through a combination of biological and endoscopic markers. In relapsing patients, infliximab re-treatment was well tolerated and induced remission.

6.2.1. Aminosalicylates 

6.2.1.1. Evidence 

Randomized trials designed to evaluate the efficacy of aminosalicylates (5-ASA) for maintaining medically induced remission are shown in Table 6.1.124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134 No additional study to evaluate the efficacy of 5-ASA for this indication has been published in Crohn's disease since 2001. The five meta-analyses carried out from these trials are summarized in Table 6.2.135, 136, 137, 138, 139 The first meta-analysis by Steinhart et al. 135 shows a benefit of mesalazine (OR 0.63; CI 0.50–0.79), but not of sulfasalazine (OR 1.08; CI 0.81–1.34). The meta-analysis by Messori et al.136 also shows a benefit of mesalazine, which was associated with a reduction in the risk of clinical relapse between 0 and 6months (OR 0.56; CI 0.37–0.84; p<0.01) and between 6 and 12months (OR 0.47; CI 0.33–0.67; p<0.001). The meta-analysis by Camma et al.137 is more complete, but also includes 5 studies designed for post-operative prevention among the 15 studies analysed. A significant reduction in the relapse risk was found when all patients were included (difference between 5-ASA and placebo: −6.3%; CI −10.4% to −2.1%), but this reduction was not significant when patients treated for medically induced remission alone were considered. No dose response could be demonstrated. When the four trials with poor quality scores were excluded, no benefit from aminosalicylates was found.

Table 6.1. Placebo-controlled trials of mesalazine for maintenance of medically induced remission in Crohn's disease.

	Author [ref.]	Year	Number of patients	Dosage (g/j)	Duration (months)	Relapse rate (%)			Comment
						5-ASA	Placebo	P
	IMSG124	1990	248	1.5	12	8	31	0.053
	Bresci125	1991	38	1.6	36	80	94	NS	Not strictly randomized
	Brignola126	1992	44	2	4	52	59	NS
	Prantera127	1992	125	2.4	12	34	55	0.02
	Gendre128	1993	161	2	24	47	42	NS	Low risk
						55	71	<0.003	High riska
	Arber129	1995	59	1	12	27	55	< 0.05
	Thomson130	1995	286	3	12	27	31	NS	I+C
						40	26	NS	I
	Modigliani131	1996	129	4	12	62	64	0.05*	*For steroid weaning
	De Franchis132	1997	117	3	12	58	52	NS
	Sutherland133	1997	293	3	11.5	25	36	NS	I+C
						21	41	0.02
	Mahmud134	2001	328	2	12	48	45	NS	Olsalazine compared to placebo in ileocolonic Crohn's disease

IMSG: International Mesalazine Study Group. I: ileal C: colonic.

a Remission <3months.

Table 6.2. Meta-analysis of placebo-controlled trials of mesalazine for maintenance of medically induced remission in Crohn's disease.

	Author [ref.]	Year	Number of trials	Number of patients	Duration (months)	Result
						Odds ratio	IC95%	p
	Steinhart135	1994	10	1022	12	0.77	0.64–0.92	–
	Messori136	1994	8	941	12	0.47	0.33–0.67	<0.001
	Camma137	1997	10	1371	4 – 48	–	–	0.06
	Akobeng138	2005	7	1500	12–24	1.00	0.80–1.24	ns
	Steinhart139	2007	9	1305	4–24	0.70	0.52–0.93	0.01

A Cochrane Database systematic review on mesalazine for maintenance of medically induced remission in Crohn's disease has been published by Akobeng et al in 2005.138 The odds ratio for 6 studies where participants were followed up for 12months was 1.00 (95% CI 0.80–1.24). For the seventh study where follow-up was for 24months,96 the odds ratio was 0.98 (95% CI 0.51–1.90). When only participants who completed the study were analysed, the odds ratio (fixed effects model) for the six 12-month studies was 0.74 (95% CI 0.57–0.96), but using the random effects model, the OR was 0.68 (95% CI 0.45–1.02). The OR for the seventh study where follow-up was for 24months,96 was 0.86; 95% CI, 0.42 to 1.78. In 2007, Steinhart et al.139 published a review including a meta-analysis of 9 RCTs that investigated mesalazine for maintenance of medically induced remission in order to explore the possibility that variations in delivery of different 5-ASA formulations may explain the inconsistent results seen in the published meta-analyses. Compared to the Cochrane Database review by Akobeng et al.138 they excluded the study by Mahmud et al.134 which investigates olsalazine vs placebo, but added three studies excluded by Akobeng et al.138 because the duration of follow-up was <6months126 or because patients were randomized during a flare.126, 131 They found a clinically significant therapeutic advantage for treatment with mesalazine over control [OR=0.70 (95% CI 0.52–0.93; p=0.01], with a benefit of 6.6% (39.1 vs 32.5%) and an NNT of 16. Treatment with pH 7-dependent mesalazine significantly reduced the risk of relapse (OR 0.38; 95% CI 0.17–0.85; p=0.01) but not treatment with controlled-release mesalazine and pH 6-dependent mesalazine. The NNT to maintain medically induced remission for pH 7-dependent mesalazine was 5. They suggest that the mesalazine formulation may be a contributory factor in the results of RCT in patients maintained on mesalazine.

6.2.2. Antibiotics 

6.2.2.1. Evidence 

The results of clinical trials are summarized in Table 6.3.54, 140, 141, 142, 143, 144, 145, 146 Most are related to anti-mycobacterial agents, but these antibiotics are also potentially active against intestinal bacteria. A meta-analysis of anti-mycobacterial therapy147 includes six fully published studies. Patients in two trials144, 145 whose remission was induced by a combination of antibiotics and steroids benefited (OR 3.37; CI 1.38–8.24, whereas patients on a combination of antibiotics compared to conventional therapy141, 142, 143, 146 did not (OR 0.69; CI 0.39–1.21). A large Australian study published in 2007 confirms this54: 213 patients were randomized to clarithromycin 750mg/day, rifabutin 450mg/day, clofazimine 50mg/day or placebo, in addition to a 16-week tapering course of prednisolone. Those in remission at week 16 continued their study medications in the maintenance phase of the trial. At week 16, there were significantly more subjects in remission in the antibiotic arm (66%) than the placebo arm (50%; p=0.02). Of 122 subjects entering the maintenance phase, 39% taking antibiotics experienced at least 1 relapse between weeks 16 and 52, compared with 56% taking placebo (p=0.054). At week 104, the figures were 26% and 43%, respectively (p=0.14). During the following year, 59% of the antibiotic group and 50% of the placebo group relapsed.

Table 6.3. Placebo-controlled trials of antibiotics for maintenance of medically induced remission in Crohn's disease.

	Author [ref.]	Year	Number of patients	Antibiotics	Duration (months)	Relapse rate (%)			Concomitant therapy
						Antibiotics	Placebo	P
	Elliott140	1982	51	Sulfadoxine+Pyrimethamine	12	62	50		No
	Shaffer141	1984	27	Ethambutol+Rifampicine	24	64	36	NS	Steroids Sulfasalazine
	Basilisco142	1989	24	Rifabutine	6	71	62		Miscallenaous
	Afdhal143	1991	49	Clofazimine	12	36	50		Steroids
	Prantera144	1994	40	Ethambutol+Clofazimine+Dapsone+ifampicine	9	89	41	0.03	Steroids
	Swift145	1994	126	Ethambutol+Rifampicine+Isoniazide	24	65	62	NS	Steroids Mesalazine
	Goodgame146	2001	31	Clarithromycine+Ethambutol (3months)	12	–	–	NS	No
	Selby54	2007	213a	Clarithromycine+Rifabutin+clofazimine	24	26	43	NS	Steroids for induction

a Of 213 included in the study, 122 patients entered in the maintenance phase.

6.2.3. Corticosteroids 

6.2.3.1. Evidence 

A meta-analysis of classic corticosteroids such as prednisolone retained 3 out of 8 studies identified in the literature, including 403 patients. The population was heterogeneous: patients had medically- or surgically-induced remission and had or had not been treated with corticosteroids before. No significant difference was found between steroids and placebo after 6, 12 or 24months.56

The six randomized placebo-controlled clinical trials evaluating budesonide in ileocolic Crohn's disease for maintenance of medically induced remission are shown in Table 6.4.147, 148, 149, 150, 151, 152, 153 One study compared budesonide to 5-aminosalicylates,154 one compared budesonide to low-dose traditional systemic corticosteroids,155 one compared two doses of budesonide with no control group156 and an additional trial compared administration of a fixed dose of budesonide (6mg daily) with a flexible dose (3–9mg).157 Four meta-analyses have been published.158, 159, 160, 161 In the first,158 4 trials (449 patients) comparing the effectiveness of budesonide 3mg (n=174) or 6mg (n=90) to placebo (n=185) were considered.148, 149, 150 The one year relapse rates were 66%, 58% and 64% respectively (OR −0.8%; CI −9.9 to +8.3%; p=0.42). The frequency of corticosteroid side effects was similar between budesonide and placebo, but significant heterogeneity was noted, with two trials reporting lower rates of side effects. In the second,159 three trials were taken into account,148, 149, 150 since the fourth151 had used a different form of budesonide, and the conclusion was identical. In a further meta-analysis,160 four RCTs with identical protocols, including the Hanauer study,153 were pooled. A total of 380 patients with Crohn's in medically induced remission were randomized to receive oral budesonide 3mg, 6mg, or placebo daily for 12months. Budesonide was not effective at maintaining remission for 12months, but the median time to relapse was 268, 170, and 154days for budesonide 6mg, budesonide 3mg, and placebo groups, respectively (p=0.007). A Cochrane Database systematic review on budesonide for maintenance of surgically- or medically induced remission in Crohn's disease has been published by Benchimol et al. in 2009, after the Consensus.161 Eleven studies were included in the review, including 8 studies comparing budesonide with placebo, and the 6 studies in patients with medically induced remission. Budesonide 6mg daily was no more effective than placebo for maintenance of remission at 3months (RR 1.25; 95% CI 1.00 to 1.58; p=0.05), 6months (RR 1.15; 95% CI 0.95 to 1.39; p=0.14), or 12months (RR 1.13; 95% CI 0.94 to 1.35; p=0.19). Budesonide was not more effective than weaning doses of prednisolone for maintenance of remission at 12months (RR 0.79; 95% CI 0.55 to 1.13; p=0.20), but was better than mesalamine 3g/day (RR 2.51; 95% CI 1.03 to 6.12; p=0.04). No differences in efficacy were detected based on the different formulations of budesonide, surgically or medically induced remission, or budesonide dose. The use of budesonide 6mg resulted in slight improvements in mean time to relapse (weighted mean difference 59.93days; 95% CI 19.02 to 100.84; p=0.004). Adverse events were more frequent in patients treated with 6mg of budesonide compared with placebo but these events were relatively minor and did not result in increased rates of study withdrawal. Cortot et al.152 evaluated the possibility of switching from systemic steroids to budesonide CIR in prednisolone/prednisone dependent patients with inactive Crohn's disease affecting the ileum and/or ascending colon. After 13weeks without predniso(lo)ne, relapse rate was 32% in the budesonide group, and 65% in the placebo group (p<0.001). The number of glucocorticosteroid side effects was reduced by 50% by switching from prednisolone and was similar in the budesonide and placebo groups. Schoon et al.57 found a significant benefit for budesonide over prednisolone when assessing bone mineral density. However, in a longitudinal study of budesonide, prednisone and nonsteroid therapy, Cino et al. found that budesonide does not confer a benefit over low-dose prednisone for the preservation of bone mineral density.162 Abnormal adrenocorticoid stimulation tests were seen more frequently in patients receiving both 6mg and 3mg daily compared with placebo. The authors concluded that the modest benefits in terms of lower CDAI scores and longer time to relapse are offset by higher treatment-related adverse event rates.

Table 6.4. Placebo-controlled trials of budesonide for maintenance of medically induced remission in Crohn's disease.

	Author [ref.]	Year	No. of patients	Dosage (mg/day)	Duration (months)	Relapse rate (%)			Formulation	Comment
						Budesonide	Placebo	p
	Löfberg148	1996	90a	6	12	74	63	NS	Controlled ileal release	After induction with either prednisolone or budesonide
				3		59
	Greenberg149	1996	105a	6	12	61	67	NS	Controlled ileal release	After induction with budesonide in a RCT
				3		70
	Ferguson150	1998	75a	6	12	46	60	NS	Controlled ileal release	After induction with budesonide in a RCT
				3		48
	Gross151	1998	179	3	12	67	65	NS	pH-modified release	After induction with 6-methylprednisolone
	Cortot152	2001	120a	6	4–5	33	65	0.05 at 3months	Controlled ileal release	Steroid-dependent patients receiving 10–30mg predniso(lo)ne at entry
	Hanauer153	2005	110a	6	12	40	47	NS	Controlled ileal release	After induction with budesonide in a RCT

a Inclusion restricted to patients with ileal or proximal colonic involvement.

6.2.4. Thiopurines 

6.2.4.1. Evidence 

Clinical trials evaluating the efficacy of azathioprine for maintenance of medically induced remission in Crohn's disease are listed in Table 6.5.23, 114, 117, 164, 165, 166, 167, 168 Mercaptopurine (1–1.5mg/kg/day) which, like azathioprine in many countries except France, is unlicensed for Crohn's disease, and is considered equivalent to azathioprine. No additional placebo-controlled study has been reported in medically induced remission since 2005, albeit three trials evaluating the efficacy of azathioprine or mercaptopurine to prevent recurrence after surgery were recently published (see Section 3, Chapter 8.0). Two studies164, 165 entered steroid dependent patients and attempted to withdraw steroids after adding azathioprine or placebo. Two other studies117, 166 identified groups of patients who were in remission on azathioprine and randomized patients to receive a one year166 or 18month course of either azathioprine or placebo117.Three meta-analyses of these studies have been published by the same group, including two Cochrane Database systematic reviews.169, 170, 171 The more recent publication171 analysed six clinical trials, including 530 patients treated with azathioprine (n=231) or placebo (n=299) for medically induced remission. The Markowitz's study168 is not included in the review, presumably because it was conducted in children who were not quiescent at entry in the trial, and the study drug was mercaptopurine. The overall remission rate was 71% (95% CI 64% to 77%) for azathioprine and 52% (95% CI 36% to 66%) for placebo (OR 2.32; CI 1.55–3.49; NNT to prevent one relapse=6). There was a dose–response effect (OR1.20; CI 0.60–2.41 at 1mg/kg/day; OR 3.01; CI 1.66–5.45 at 2mg/kg/day; and OR 4.13; CI 1.59–10.71 for 2.5mg/kg/day). Two clinical trials have examined the steroid-sparing effect of thiopurines,164, 165 which was observed in 87% of patients in the azathioprine group and 53% on placebo (OR 5.22; CI 1.06–25.68). However, the risk of premature withdrawal from the study for side effects was significantly increased with azathioprine compared to placebo (OR 3.74; CI 1.48–9.45).

Table 6.5. Placebo-controlled trials of azathioprine (AZA) or mercaptopurine (MP) for maintenance of medically induced remission in Crohn's disease.

	Author [ref.]	Year	Nb of patients	Drug (mg/kg/day)	Duration (months)	Relapse rate (%)			Randomized patients
						AZA	Placebo	p or 6-MP
	Willoughby164	1971	10	AZA (2.0)	6	20	60	<0.05	Steroid-dependent patients
	Rosenberg165	1975	20	AZA (2.0)	9	20	50	<0.01	Steroid-dependent patients
	O'Donoghue166	1978	51	AZA (2.0)	12	5	41	<0.05	Patients in remission on azathioprine (withdrawal study)
	Summers23 (NCCDS) (part I, phase 2)	1979	19	AZA (2.5)	9	16	25	NS	Patients who achieved remission after 17weeks
	Summers23 (NCCDS) (part II)		151	AZA (1.0)	24	–	–	NS	Patients with inactive disease
	Candy114	1995	63	AZA (2.5)	12	58	93	<0.001	Patients with active disease. Induction of remission with a 3-month course of prednisone
	Markowitz168	2000	55	MP (1.5)	18	9	47	0.007	Children with newly diagnosed Crohn's disease; induction with prednisone
	Lémann117 (GETAID)	2005	83	AZA (1.7)	18	8	21	NS (non-inferiority design)	Patients in remission on azathioprine >42months (withdrawal study)

In three recent studies,1, 116, 170 azathioprine was used as a comparator to evaluate the efficacy of infliximab alone,1 infliximab combined with azathioprine,1, 116 or everolimus170 to induce and maintain remission. Patients initially received steroids until response which were then tapered. Steroid-free remission (CDAI<150) was the primary endpoint of the three trials. In the azathioprine groups, success rates were remarkably similar: 29% at 6months,116 30% at 6months1 and 38% at 7months.170 The apparent discrepancy between these findings and previous evidence suggesting a more pronounced beneficial effect of azathioprine on the maintenance of remission may be the more stringent criterion for success which was used in these recent studies. Another explanation could be the selection of patients with a more advanced disease, although in the study by Colombel et al.1 the median disease duration was 2years. In the Markowitz's study,168 children were enrolled before receiving any treatment, after two less than 2weeks of unsuccessful treatment with 5-ASA or if they had received less than 6weeks of prednisone. Fifty-five children were randomized to treatment with mercaptopurine or placebo within 8weeks of initial diagnosis. Both groups also received prednisone. Although remission was induced in 89% of both groups, only 9% of the remitters in the 6-MP group relapsed within the 18months following randomization, compared with 47% of controls (p=0.007). D'Haens et al.18 recently published a study comparing two strategies in patients who had been diagnosed with Crohn's disease within the past 4years (median, 2weeks from diagnosis) and who had not previously received corticosteroids, antimetabolites, or biological agents. They randomly assigned 133 patients to either early combined immunosuppression (infliximab 5mg/kg at weeks 0, 2, and 6, with azathioprine and additional treatment with infliximab and, if necessary, corticosteroids) or conventional management (corticosteroids, followed, in sequence, by corticosteroids+azathioprine in patients who experienced a relapse, and infliximab if necessary). At week 26, 39 (60%) of 65 patients in the combined immunosuppression group were in steroid free remission without surgical resection, compared with 23 (36%) of 64 controls (p=0.0062). Corresponding rates at week 52 were 40/65 (61%) and 27/64 (42%) (p=0.0278). These two studies suggest that early introduction of azathioprine combined with steroids (or infliximab) within the months following diagnosis can improve success rate.

T(h)ioguanine, the active metabolite of azathioprine and mercaptopurine, might be an alternative to these agents in intolerant patients. No controlled study is available, but in several series thioguanine appeared to be similarly effective to azathioprine or mercaptopurine.173, 174 Unfortunately, a high frequency of liver abnormalities has been reported, mostly nodular regenerative hyperplasia172, 173, 174, 175, 176, 177 which is an irreversible cause of portal hypertension. Therefore, thioguanine cannot currently be recommended for maintenance of Crohn's disease.

6.2.5. Methotrexate 

6.2.6. Other immunosuppressants 

6.2.7. Anti-TNF agents 

6.2.7.1. Evidence 

Clinical trials evaluating the efficacy of anti-TNF agents for maintenance of medically induced remission in luminal Crohn's disease are listed in Table 6.61, 28, 115, 179, 184 (for fistulating Crohn's disease — see Chapter 9.0).

Table 6.6. Randomized controlled trials of anti-TNFα including infliximab (IFX), adalimumab (ADA), certolizumab (CTZ) and CDP571 for maintenance of medically induced remission in luminal Crohn's disease.

	Author [ref.]	Year	Nb of randomized patients (initial population)	Anti-TNF	Duration (weeks)	Clinical remission rate (%)			Randomized patients
						Anti-TNF	Placebo	p
	Rutgeerts180	1999	73	IFX 10mg/kg/8weeks	44	53a	20a	0.013	Responders to an initial IFX treatment
			(73)
	Sandborn179	2001	169	CDP571 10mg/kg/8weeks	24	11	4	–	Patients with active disease randomized to CDP571 10mg/Kg or 20mg/Kg or placebo and then continued maintenance
			(169)	10mg/kg/12weeks		11	3	–
	Hanauer115 (ACCENT 1)	2002	335	IFX 5mg/kg/8weeks	54	28a	14a	0.007	Responders at week 2 to open-label IFX (single infusion of 5mg/Kg)
			(573)	10mg/kg/8weeks		38a		p<0.001
	Sandborn181	2004	396	CDP571	28	24	20	NS	Patients with active disease (induction and maintenance)
			(396)	10mg/kg
	Feagan183	2005	271	CDP571	36	29b	37b	NS	Steroid-dependent patients (in remission at entry)
				10mg/kg/8weeks
	Colombel27 (CHARM)	2007	499	ADA 40mg eow	56	36	12	<0.001	Responders to open-label ADA (80mg/40mg) at week 4
			(854)	40mg/week		41
	Sandborn184 (CLASSIC 2)	2007	55	ADA 40mg eow	56	79	44	<0.05	Patients from an induction RCT then received open-label ADA 40mg at week 0 and 2. Remitters at week 4 enrolled
			(276)	40mg/week		83
	Schreiber182 (PRECISE 2)	2007	428	CTZ 400mg/4weeks	26	48	29	<0.001	Responders to open-label induction with CTZ 400mg at weeks 0, 2, 4
			(668)
	Sandborn28 (PRECISE 1)	2007	662	CTZ 400mg/4weeks	26	10	14	0.07	Patients with active disease (induction and maintenance)
			(662)
	Van Assche192 (IMID)	2008	80	IFX+AZA or MTX	104	60	–	NS	Responders to IFX+AZA for at least 6months (in remission at entry)
			(80)	IFX		55
	Feagan41 (COMMIT)	2008	126	IFX 5mg/kg	50	56c	–	NS	Patients with active disease (induction with steroids+IFX)
			(126)	IFX+MTX		57c
	Colombel1 (SONIC)	2008	508	IFX 5mg/kg	26	44c	–	0.02d	Patients with active disease (induction and maintenance)
			(508)	IFX+AZA		57c		<0.001d
				AZA		30c

AZA = azathioprine; MTX = methotrexate,CTZ = certolizumab pegol, IFX = infliximab, ADA = adalimumab.

a Clinical remission rate was not the primary endpoint and was not reported in the article; information obtained from the authors is available in the Cochrane Database Systematic Review [Behm 2009]. b Primary endpoint was steroid withdrawal. c Primary endpoint was steroid-free remission. d Versus IFX+AZA.

Two meta-analyses of these studies have been recently published, including a Cochrane Database Systematic Review76, 185 published after the Consensus. In the first,76 because of heterogeneity, two different types of study design were analysed separately: long-term (20–52weeks) maintenance trials with randomization of responding patients to infliximab (n=2), adalimumab (n=1) or certolizumab (n=1) at weeks 2–6 after open-label induction, and long-term (24–28weeks) induction trials with randomization before induction (n=3). In overall analysis, among responders after open-label induction, anti-TNF therapy was more effective than the placebo for maintenance of remission at weeks 20–30 and 48–52 (mean difference, 23%; 95% CI, 18%–29%; P<0.001). When considering responders and nonresponders after open-label induction in 3 trials, mean difference and 95% CI were 11.6% and 5%–18%, respectively, at weeks 20–30. Three short- and long-term induction/maintenance trials evaluating certolizumab (n=1) and CDP571 (n=2) were pooled. In overall analysis, anti-TNF therapy was more effective than the placebo for maintenance of remission at weeks 20–30 but in subgroup analysis, CDP571 was not effective in maintaining remission. In the 21 studies enrolling 5356 individuals included in the meta-analysis, anti-TNF therapy did not increase the risk of death, malignancy, or serious infection.

In the Cochrane Database review,185 the authors did not combine the data from trials involving different anti-TNF agents. One study evaluating certolizumab pegol28 was excluded because it reported combined induction and maintenance data, and the authors of the review felt it was not possible to evaluate maintenance therapy in clinical responders based on the published data. In the pooled analysis, infliximab was found to be superior to placebo for the maintenance of remission (RR 2.50; 95% CI 1.64 to 3.80; p<0.0001) and clinical response (RR 2.19; 95% CI 1.27 to 3.75; p=0.005). Infliximab was also superior to placebo for corticosteroid-sparing effects (RR 3.13; 95% CI 1.25 to 7.81; p=0.01). There were no significant differences in remission rates between infliximab doses of 5mg/kg or 10mg/kg. Compared with placebo, certolizumab pegol 400mg every 4weeks was also found to be effective for maintenance of clinical remission (RR 1.68; 95% CI1.30 to 2.16; p<0.0001) and clinical response (RR 1.74; 95% CI 1.41 to 2.13; p<0.00001) to week 26 in patients who have responded to certolizumab therapy. The two studies evaluating adalimumab were evaluated separately due to heterogeneity among the participants. In CHARM,27 adalimumab was found to be superior to placebo for maintenance of clinical remission to week 54 (RR 3.28; 95% CI 2.13 to 5.06). In CLASSIC 2,184 adalimumab was also found to be superior to placebo for maintenance of clinical remission to week 54 (RR 1.82; 95% CI 1.06 to 3.13). There were no significant differences in remission rates between adalimumab 40mg weekly or every other week. There was no evidence to support the use of CDP571 for the maintenance of remission in Crohn's disease. Although differences in trial durations limit direct comparisons of all data, the authors concluded that it appears likely that infliximab, adalimumab, and certolizumab pegol have similar clinical efficacy in patients with Crohn's disease. Adverse events were observed in approximately equal frequencies in the treatment and placebo groups, but they noted that serious adverse events, including tuberculosis and lymphoma, were reported in several trials.

The ACCENT 1 study has been re-analysed185 to compare episodic and scheduled treatment strategies. This included all 573 patients (responders and non responders) and compared regularly scheduled maintenance (infliximab groups) and episodic maintenance (placebo group). Mean CDAIs were significantly better in the 10mg/kg scheduled group from weeks 10 to 54, while response and remission rates in the combined 5 and 10mg/kg scheduled-treatment were higher from weeks 10 to 30. A lower proportion of patients developed antibodies to infliximab in the scheduled-treatment groups. Perhaps most relevant was the observation that patients in scheduled strategy had fewer Crohn's disease-related hospital admissions and surgery compared to those in the episodic strategy.

The value of combining anti-TNFα with immunosuppressant agents such as azathioprine or methotrexate is still debated. Some studies have shown that the use of concomitant immunosuppressant therapy may reduce the risk of antibodies directed against infliximab and improve efficacy, but these data mostly come from studies using episodic infliximab therapy.187, 188 In the ACCENT I trial, reduced antibody formation was observed when an induction regimen is followed by maintenance treatment compared to a single dose followed by episodic treatment (8 vs. 30%; p<0.001),77, 115, 186 but concomitant immunomodulatory therapy with infliximab was not associated with better clinical outcome when 3-dose induction was followed by scheduled maintenance therapy. These findings are consistent with analyses of the impact of baseline concomitant immunosuppressants performed on data from maintenance trials with other anti-TNFα (adalimumab, certolizumab).27, 28, 182 In an open-label, randomized, controlled trial, Van Assche et al.189 have shown that during maintenance therapy with infliximab, the continuation of immunosuppressive therapy for more than 6months offers no clinical benefit over ongoing infliximab monotherapy, but is associated with higher infliximab trough levels. Feagan et al.41 have reported the results of a randomized, placebo-controlled study to evaluate the efficacy of infliximab in combination with methotrexate. Patients with active Crohn's disease who had corticosteroid therapy initiated within 6weeks were randomized to methotrexate or placebo for up to 50weeks. Both groups received induction and maintenance infliximab therapy for up to 50weeks. High rates of corticosteroid-free remission over one year were demonstrated with both regimens (56% and 57%) but triple induction therapy (prednisone+methotrexate+infliximab) followed by methotrexate+infliximab maintenance therapy was not more effective than dual induction therapy (prednisone+infliximab) followed by infliximab maintenance therapy. In contrast, the results of a randomized, double-blind study of 508 patients with active Crohn's disease who were naïve to immunomodulator and anti-TNF biologic therapies (SONIC) were recently reported.1 Combination therapy with infliximab and azathioprine was superior to infliximab and azathioprine monotherapies during the first 6months of the study; during the same period, safety results were similar among the three treatment groups. It is thus possible that there may be greater efficacy with concomitant immunomodulators in naïve patients but not in those who have already failed these drugs. 116 Available data also suggest an increased risk of hepatosplenic T-cell lymphoma when azathioprine in combination with infliximab therapy is administered.190 However there is no evidence of a higher risk of opportunistic infections with the combination of azathioprine and infliximab as compared to azathioprine or infliximab alone.1, 59, 60, 191

6.2.8. Other biologic therapies 

Natalizumab, a humanized anti-α4 integrin monoclonal antibody, was investigated for maintenance of response and remission in Crohn's disease (ENACT-2 study): 339 patients with a response (ΔCDAI≥−70) or remission after induction with natalizumab (ENACT-1, a 905 patient induction study — see Active Disease Section 5.4.5) were allocated to receive infusions of placebo or 300mg of natalizumab every 4weeks for 12months.[79] Maintenance natalizumab resulted in higher rates of sustained response (61% vs 28%, p<0.001) and remission (44% vs 26%, p=0.003) through week 36 than did switching to placebo. Despite this promising result for maintenance, treatment with natalizumab has not been approved in the European Union partly due to cases of progressive multifocal leucoencephalopathy occurred in several patients with multiple sclerosis and one patient with Crohn's disease.192

Other biologic therapies are under evaluation in Crohn's disease including anti-adhesion molecules (MLN-02, alicaforsen, CCX-282-B), anti-inflammatory cytokines (interleukin 10, interleukin 11, and interferon-beta), anti-IL12 p40 antibody (ustekinumab, ABT-874), anti-interferon-gamma (fontolizumab), anti-IL-6 (tocilizumab), anti-CD28 (abatacept) anti-CD3 (visilizumab) or anti-CD4 (cM-T412) antibodies, G-CSF (filgastrim) or GM-CSF (sargramostrim) and growth hormone (somatropin). Promising results have been reported in IBD with several of these novel biologic therapies,78, 193 but none have yet been evaluated for maintenance of remission in Crohn's disease.

6.2.9. Diet therapy 

6.2.10. Probiotics 

6.2.10.1. Evidence 

Clinical trials evaluating the efficacy of probiotics including E. coli Nissle 1917, Saccharomyces boulardii and Lactobacillus GG for maintenance of medically induced remission in luminal Crohn's disease are listed in Table 6.7.202, 203, 204, 205, 206 In a Cochrane Database Systematic Review, Rolfe et al.207 examined the role of probiotics in the maintenance of surgically-induced (2 trials) or medically induced (5 trials) remission in Crohn's disease. All of the studies included small numbers of patients and may have lacked statistical power to show differences should they exist. Compared to placebo, there was no statistically significant benefit of E. coli Nissle (RR 0.43, 95% CI 0.15 to 1.20) or Lactobacillus GG (RR 0.83, 95% CI 0.25 to 2.80) for reducing the risk of relapse there was no statistically significant benefit of probiotics for reducing the risk of relapse compared to maintenance therapy employing aminosalicylates or azathioprine (RR 0.67, 95% CI 0.13 to 3.30).

Table 6.7. Placebo-controlled trials of probiotics for maintenance of medically induced remission in Crohn's disease.

	Author [ref.]	Year	Nb of randomized patients (initial population)	Drug (mg/Kg/day)	Duration (months)	Relapse rate (%)			Randomized patients
						Probiotics	Placebo	p
	Malchow202	1997	20	E. coli 1917 (200mg/d)	12	30	70	NS	Patients with active colonic disease on a steroid-tapering regimen. ncluded when remission (CDAI<150) was achieved
			(28)
	Guslandi203	2000	32	S. boulardi 1g/d+MSZ 2g/d	6	6	–	0.08	Patients in remission for at least 3months
				MSZ 3g/d		37	–
	Zocco204	2003	35	Lactobacilli GG 18 billion/d	12	17	–	NS	Patients in remission
				MSZ 2.4g/d		25
				Lactobacilli GG+MSZ		18
	Schultz205	2004	9	Lactobacilli GG 20billion/d	6	50	60	NS	Patients with active disease treated for 2-week with antibiotics and a 3-month steroid-tapering regimen
			(11)
	Bousvaros206	2005	75	Lactobacilli GG 20billion/d	10	31	17	NS	Children in remission (PCDAI<10) on other maintenance therapies at entry (5-ASA, thiopurines, low-dose of steroids)

MSZ = mesalazine.

6.2.11. Cytapheresis and autologous stem cell transplantation 

The effectiveness of lymphapheresis was studied in 28 patients in clinical remission induced by steroids. After 18months, the rate of relapse was 83% in the lymphapheresis group and 62% in the control group (ns).208 Adacolumn® and Cellsorba® leukocyte filters have also been proposed for leucocyte apheresis, but to date, only a few case series and open studies have evaluated its efficacy in active Crohn's disease, with variable results.209 No study is available for maintenance of medically induced remission.

Oyama et al.210 conducted a phase 1 study in 12 patients with active Crohn's disease despite conventional therapies including infliximab. Peripheral blood stem cells were mobilized with cyclophosphamide and granulocyte colony-stimulating factor and CD34+ enriched. Eleven of 12 patients entered a sustained remission. After a median follow-up of 18months, only one patient has developed a recurrence. The procedure was well-tolerated. Other authors have shown similar results in a limited series.211 Randomized controlled trials are ongoing.

6.2.12. General conclusion 

Medications whose efficacy for maintaining medically induced remission in Crohn's disease is well established [EL1a] include azathioprine, infliximab and adalimumab. There is also a reasonable level of evidence [EL1b] for methotrexate, certolizumab and natalizumab [EL1b]. The efficacy of mesalazine [EL1b] and omega-3 fatty acids [EL1b] remains controversial, due to inconsistent results. There is not enough evidence to support the use of enteral nutritional supplementation, S. boulardii, E. coli Nissle 1917, cytapheresis and autologous stem cell transplantation. The available evidence shows that ciclosporin, anti-mycobacterial agents, CDP571, and Lactobacillus GG are ineffective.

7.2.1. Localised ileal or ileocaecal disease 

Patients with inflammatory Crohn's disease confined to the ileo-caecum with a maximum of 40cm affected bowel and appreciable symptoms (CDAI>220) but no imminent obstruction respond well to steroid treatment. However, this patient group will almost always require surgery during the course of their disease. Following resection, long-term studies have demonstrated that there is a 50% chance that the patient will never require a further operation (i.e. have symptoms of the same severity again).213, 214, 215, 216 In contrast there are no long-term follow-up studies (i.e. >15years) of the outcome of medical treatment. In addition, it is not known whether there are long-term differences in the quality of life of patients treated by medical as opposed to surgical therapy. Primary surgery should be considered as the first choice for patients with refractory obstructive symptoms after initial medical treatment (steroids) in ileocaecal Crohn's disease. Likewise, patients presenting with obstruction without inflammatory activity, for example assessed by C-reactive protein (CRP) levels,217, 218, 219 can also be treated with primary surgery. If, however, the patient has had previous ileocaecal resection and anastomotic stenosis has occurred, endoscopic dilatation could be tried before moving to an intestinal resection.4, 220

7.2.2. Concomitant abscess 

When active small bowel Crohn's disease is associated with a concomitant abdominal abscess, the consensus favours percutaneous drainage and delayed resection if there are obstructive symptoms. Drainage followed by medical treatment should be considered if there are no obstructive symptoms, depending on the clinical situation. Some abscesses do not lend themselves to percutaneous drainage. There are no randomized studies in the literature to clarify whether percutaneous or surgical drainage should always be followed by a delayed resection, and although most case series favour a delayed elective resection, opinions vary.221, 222, 223

7.2.3. Stricturoplasty 

Most authors limit conventional stricturoplasties to strictures <10cm in length. The majority opinion is that stricturoplasty is inadvisable for longer (>10cm) strictures. However, there are now series reported with non-conventional stricturoplasties for longer bowel segments, reporting good results.224, 225, 226, 227, 228, 229 A phlegmon in the bowel wall, carcinoma, or active bleeding with mucosal disease are contraindications to stricturoplasty. Where there are multiple strictures in a short segment and where bowel length is sufficient to avoid short bowel syndrome, resection may be preferable.

Recent systematic reviews230, 231 and patient series232 comparing stricturoplasty and resection have confirmed the safety and bowel-sparing potential of strictureplasty for small bowel Crohn's disease. The question whether resection may induce a longer recurrence-free survival has not been resolved.231, 233 There are several case reports of adenocarcinoma at strictureplasty sites,234 rendering the need for a certain caution over the long-term consequences of the procedure.

7.2.4. Anastomotic technique 

The observation that recurrent Crohn's disease almost invariably appears just proximal to the anastomosis has led to the assumption that the width of the anastomosis matters. Several studies have tried to address this.235, 236, 237, 238, 239, 240 A recent meta-analysis241 of 8 comparative studies (2 RCTs) published between 1992 and 2005 comparing end-to-end anastomosis and stapled side-to-side anastomotic configurations including 712 anastomoses in 661 patients showed that end-to-end anastomosis after ileocolonic resection for Crohn's disease was associated with increased anastomotic leak rates and overall post-operative complications. There was no significant difference with regard to peri-anastomotic recurrence rates. Thus, there is evidence to support the choice of stapled side-to-side anastomosis in this patient group. On the other hand, a recent prospective cohort study showed no difference in safety and recurrence rate between hand-sewn side-to-side and stapled side-to-side anastomosis,236 which may imply that a wide anastomotic luminal diameter is the discriminating factor, rather than the suturing technique used.

7.2.5. ‘Coincidental’ ileitis 

The finding of terminal ileitis or caecitis at laparoscopy or laparotomy for a clinical suspicion of appendicitis is non-specific, and it is virtually impossible to differentiate between Crohn's disease and infectious (e.g. Yersinia species) enteritis. Even if it were to be Crohn's ileitis, resection might not be the most appropriate strategy if the dominant symptoms relate to inflammation. Only when the patient's history indicates obstructive symptoms for more than a few days, or the proximal intestine is dilated and the inflamed bowel wall looks typical of Crohn's disease with mesenteric fat wrapping, is an experienced surgeon justified in performing a primary resection.242

7.2.6. Laparoscopic resection 

Several studies during the last few years have shown that laparoscopic resection gives substantial benefits in addition to a shorter scar. The literature previously contained mostly retrospective and non-randomized studies.243, 244, 245 However, two recent meta-analyses of 14 and 15 studies, respectively (with 10 of the studies included in both) showed benefits in the post-operative period for the laparoscopic group. Advantages included earlier recovery of normal intestinal function, shorter hospital stay and lower post-operative morbidity.246, 247 This was also confirmed in a US nationwide registry study of 49,609 resections for Crohn's disease.248 The 2826 cases (6%) done laparoscopically were associated with shorter length of stay, lower charges, a lower complication rate (8% vs. 16%), and reduced mortality (0.2% vs. 0.9%, p<0.01). There has been debate about the heterogeneity inherent in a meta-analysis, which may also apply to the registry study. However, the most important findings of reduced morbidity are similar in recent randomized trials,249, 250 which also report better results with fewer complications and shorter hospital stay compared to conventional surgery for selected patients undergoing ileocolic resection for Crohn's disease. The 10-year follow-up of a randomized trial comparing open and laparoscopic resection for ileocolic Crohn's showed equal rates of surgical recurrence.251 Moreover, better cosmesis scores and body image in the laparoscopy groups have also been reported252; important parameters to consider in this young patient group. Thus, although laparoscopic surgery for Crohn's disease is technically demanding, there is emerging evidence for significant advantages with the technique for primary ileocolonic resections.

Although extrapolated from surgery for other diagnoses, there is also a potential benefit from laparoscopy in reducing ventral hernias and adhesion formation.253

This may make it possible to perform repeat IBD surgery via the laparoscopic approach. Evidence for feasibility and safety in complex Crohn's is scarce254 with recurrent disease and intra-abdominal abscess or fistulae being important risk factors for conversion to open laparotomy.255 A high conversion rate is pertinent when dealing with complex IBD surgery to ensure patient safety. Laparoscopic surgery in complex cases should currently only be done at highly specialized centres and preferentially within clinical studies.

7.3.1. Localised colonic disease 

Limited colonic Crohn's disease treated by segmental resection results in a higher rate of recurrence than a proctocolectomy.237, 256, 257, 258, 259, 260 However, most agree that the avoidance of a permanent stoma usually outweighs the increased risk of recurrence.

7.3.2. Multi-segment colonic disease 

The consensus is less clear when it comes to a patient with macroscopic disease in two widely separated segments of the colon. Half of the experts felt that segmental resection of the macroscopic disease and two anastomoses are acceptable. Others believed that a subtotal colectomy with an ileorectal anastomosis should be performed when macroscopic disease affects the ascending and the whole of the sigmoid colon, assuming that surgery is indicated. There is some support for separate segmental resection in the literature.260 Decisions should take individual preferences of the patient and surgeon into account.

7.3.3. Dilatation of strictures 

Endoscopic dilatation is an accepted technique for the management of mild to moderate stenosing disease. Outcomes suggest a short to midterm benefit.261, 262 Most experts consider that dilatation of a stenosis in Crohn's disease should only be attempted in institutions with a 24-h surgical service. The literature does not provide any guidance on this, although perforation and other complications requiring surgical intervention can occur.263 A recent review of 13 studies enrolling 347 Crohn's disease patients treated with endoscopic dilatation for postsurgical strictures reported an 80% technical success rate. A stricture length < or=4cm was associated with a surgery-free outcome although major complications including perforation occurred in 2%.220 It was concluded that endoscopic dilatation is effective and safe, especially for recurrence after ileocolonic resections, delaying surgery by a mean of 3years.

7.3.4. Colonic stricturoplasty 

Most experts agree that stricturoplasty is not an option for strictures in the colon, although there is insufficient evidence in the literature. A particular concern is the increased chance of cancer in a colonic stricture compared to the small bowel. One retrospective report indicates that stricturoplasty for large bowel stenoses in Crohn's disease is feasible.264

7.3.5. Ileopouch-anal anastomosis 

Most IPAA series include some patients with Crohn's disease. Retrospective analyses show that these patients suffer a higher complication rate, with pouch failure reported in up to 50%.265, 266, 267, 268, 269 However, one group reports a very small increase in morbidity compared to patients with ulcerative colitis.270, 271 Some suggest this may reflect differences in pathological diagnosis. A recent meta-analysis272 of outcomes of IPAA in Crohn's disease did, however, show more anastomotic strictures and incontinence in patients with Crohn's disease. Moreover, pouch failure was 6-fold more frequent than with ulcerative colitis and indeterminate colitis. This has been confirmed in other large single-centre series.273, 274 Half the experts are prepared to recommend an IPAA for a patient with long-standing Crohn's colitis, provided there is no sign of small bowel or perianal disease, and that the patient is willing to except an increased risk of complications and pouch failure. Many would hesitate strongly to recommend this.

7.4.1. Surgery after anti-TNF therapy 

TNFα is a key player in the immune response. Therefore, inhibition by anti-TNF therapy could potentially lead to serious post-operative complications. Although the initial published literature did not support this theory,275, 276 recent studies report conflicting data. Appau et al.277 compared 30-day post-operative outcomes for Crohn's patients treated with infliximab within 3months prior to ileocolonic resection. The infliximab and non-infliximab groups had similar disease characteristics. Multivariate analysis showed infliximab use to be associated with 30-day post-operative readmission (p=0.045), sepsis (p=0.027), and intra-abdominal abscess (p=0.005). Although this study contradicts previous studies, its strength is that it concentrates on ileocolonic anastomoses after preoperative anti-TNF treatment. In another recent series of a mixed IBD population,278 anti-TNF treatment was not associated with an increased rate of post-operative complications. There is no consensus among experts as to the optimum time between treatment with an anti- TNF agent and abdominal surgery, with equal proportions suggesting one month, a longer period, or that it does not matter. There is almost no evidence from the literature. The pharmacokinetics of monoclonal antibodies is such that therapeutic concentrations generally persist after an infusion for at least 8weeks.

7.4.2. Patients on steroids 

A third of experts agree that treatment with steroids is a risk factor for post-operative complications. Uncontrolled or retrospective series indicate that patients taking ≥20mg prednisolone for >6weeks do have an increased risk for surgical complications.58, 277

7.4.3. Patients on thiopurines 

Most publications agree that azathioprine does not increase the risk of surgical complications,58, 279, 280 although some question this.281

7.5.1. Surgery and medicine are complementary 

It is important that patients with abscesses because of internal fistulas (enteric-bladder-vaginal) and/or stenotic bowel segments are considered for surgery at a relatively early stage of the disease rather than after prolonged immunosuppressive treatment that may increase the risk of sepsis and lead to compromised healing capacity. In these complex cases early surgery may in reality be the preferred top-down approach.282

Morbidity rates of long-term steroid use are well established. Biologics, particularly in combination with immune suppression, may also have long-term effects that currently are unknown. With this in mind, surgery must be viewed as a therapeutic alternative to immunosuppressive medication, rather than a last resort. The ongoing trial comparing infliximab treatment with surgery in recurrent distal ileitis in Crohn's disease283 is a welcome and important effort in an area that is currently lacking evidence

The increasing complexity and number of choices both in medical treatment and surgical options for patients with Crohn's disease makes it even more important to discuss each case in multidisciplinary clinical conferences. Correct diagnosis and assessment with optimised medical treatment are prerequisites for optimal surgical timing and good surgical outcomes. Therefore close interaction between the gastroenterologist, the colorectal surgeon and other disciplines are needed for optimal patient outcome over a life-long perspective.

7.5.2. Fitness for surgery 

An essential part of surgical management involves the selection of patients for surgery. Fitness for surgery includes nutritional, medical, social and psychological factors. Smoking is a major risk factor for post-operative recurrence (OR 2.5) and all patients with Crohn's disease should be strongly recommended to stop smoking before undergoing surgery.284 Although there is no body of evidence, nutritionally compromised patients with major weight loss (>10% in 3months) are likely to benefit from a period of preoperative nutritional support, often requiring parenteral nutrition. Patients with a low serum albumin usually have uncontrolled sepsis and may or may not be nutritionally compromised. Such patients are likely to benefit from drainage of sepsis together with nutritional support.