Article Text

Download PDFPDF

Early intervention in Crohn’s disease: towards disease modification trials
  1. Silvio Danese1,2,
  2. Gionata Fiorino1,
  3. Laurent Peyrin-Biroulet3
  1. 1Department of Gastroenterology, IBD Center, Humanitas Research Hospital, Milan, Italy
  2. 2Department of Biomedical Sciences, Humanitas University, Milan, Italy
  3. 3Department of Hepato-Gastroenterology and Inserm U954, University Hospital of Nancy, Lorraine University, Vandoeuvre-lès-Nancy, France
  1. Correspondence to Dr Silvio Danese, Department of Biomedical Sciences, Humanitas University, Via Manzoni 113, 20089 Rozzano, Milan, Italy; sdanese{at}


Crohn’s disease (CD) is a chronic progressive destructive inflammatory bowel disease. As in rheumatoid arthritis, there is increasing evidence that early treatment initiation with disease-modifying agents, such as biological drugs, may lead to complete disease control, prevention of disease progression thus protecting against irreversible damage and restoration of normal quality of life. Data from randomised clinical trials with immunosuppressants and biologics suggest that treating patients with a disease duration of <2 years and an absence of complications may significantly reduce the risk for complications and increase time in remission in patients with CD. Moreover, rapid disease control may effectively prevent disease progression and allow dose reduction or even withdrawal of treatment, reducing the risk of long-term adverse events and healthcare costs. However, prospective disease modification trials are needed to confirm these initial results. Here we review the literature regarding early intervention in adult patients with CD and propose criteria for future disease modification trials.

  • Crohn’s disease
  • pharmacotherapy
  • immunoregulation

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Crohn’s disease (CD) is a progressive, chronic disease, with a destructive and disabling course over time.1 2 The cycle of inflammation and tissue repair can cause fibrosis and consequent obstructive intestinal strictures, and depending on severity, transmural lesions can become fistulas and abscesses. These complications are all associated with cumulative bowel damage, frequently requiring surgery and associated further bowel damage.3 4 Population-based studies suggest that 20% of patients have bowel damage at diagnosis,5 defined as the presence of complications, such as strictures, fistulas and/or abscesses,3 and we have recently shown, using cross-sectional imaging, that about 40% of patients already have bowel damage within the first 2 years after diagnosis.5 6 About half of all patients with CD will develop bowel damage over 20 years5 (online supplementary figure 1).

Antitumour necrosis factor α (anti-TNF) therapy has led to a significant reduction in hospitalisation and in surgery rates in patients with CD in the short term, compared with conventional immunomodulators such as azathioprine.7 Evidence is still lacking with ustekinumab and vedolizumab. However, current therapeutic strategies alone fail to alter disease course in the long term.8–10 A recent extensive review of the literature including randomised controlled trials (RCTs) and population-based studies suggests that the anti-TNF agents may provide a protective effect for undergoing surgery, although the risk of surgery has decreased significantly in the past 60 years, which may be related to factors other than anti-TNF alone, such as changes in care from surgeons to gastroenterologists, earlier diagnosis and implementation of practice guidelines.11 12 It is important to remember that surgery is not curative, as disease recurs in more than 50% of patients at year 3, despite effective preventive therapeutic strategies.13 14

Recent data from pivotal trials in patients with CD suggest that anti-TNF therapy is more effective in early disease,15–17 whereas early introduction of azathioprine led to disappointing results.18 19 Two main reasons may contribute to greater efficacy of anti-TNF therapy in early CD: immunology20 21 and the high percentage of patients with pure inflammatory, uncomplicated disease.5 6

Evidence from early-CD paediatric populations20 22 (from 0 to 6 months from the onset of symptoms to diagnosis) compared with late CD in paediatric populations (5–8 years since diagnosis) showed that, similarly to acute infectious colitis, in early CD, interleukin (IL) 12 induced a stronger polarisation T-helper (Th)−1 response, characterised by higher IFN-γ production and IL12Rβ2 chain expression than in late disease. Furthermore, mucosal levels of IL12p40 and IL12Rβ2 messenger RNA were significantly higher in children with early than late CD.20 The RISK study identified that older age, African-American race and ASCA and CBir1 seropositivity were associated with disease complications. CBir1 seropositivity to bacterial flagellin had the strongest effect on the disease progression and might be implicated also in the pathogenesis of the disease.22

Data from population-based cohort studies5 show that early non-complicated CD may evolve to complications within 90 days with a cumulative risk of 18.6%, suggesting that early intervention is crucial for changing the natural history of the disease.

Over the last decade, several trials included an early CD population.17 23–26 However, no controlled trials specifically addressed this issue. Recently, a worldwide expert consensus defined early CD as a disease diagnosed within 18 months, with no complications and no previous treatment with disease-modifying agents (disease-modifying antirheumatic drugs (DMARDs), mainly thiopurines, methotrexate (MTX) and/or biologics).27 However, such definition needs to be validated in prospective RCTs.

Herein, we review the current evidence on early intervention in adult patients with CD, based on currently available data from the most important studies on therapeutics in CD, with the aim of proposing a suitable study design for future trials in early CD.

Lessons from rheumatoid arthritis

The concept of early intervention is well established in other immune-mediated diseases, such as rheumatoid arthritis (RA)28 (figure 1). Several definitions of early RA have been proposed29: either as a duration of disease/symptoms of less than 6 months30 or no more than 1 year of disease duration together with no previous use of DMARD.31 Clinical studies investigating whether early introduction of immunosuppressants and/or biologics at the earliest stage of the disease show that between 70% and 90% of patients treated with infliximab or etanercept achieve disease remission in the short term (table 1). The BeSt study enrolled 508 patients with early RA (defined as disease duration of 2 years or less)32; comparing four different treatment strategies showed that early combination of treatments was associated to lower rates of joint damage (p<0.05).32 The COMET trial showed that patients naïve to MTX and who had short disease duration (≤2 years) treated with the combination of MTX and etanercept achieved disease remission in 50% of cases vs 28% with monotherapy, and 80% of them had no radiological progression compared with 59% in the control group.33 Very early intervention (≤4 months) was effective in up to 70% of patients treated with combination therapy.28 Consistent data were found in previous studies with infliximab, suggesting that the first 2 years of disease may be a valid window of opportunity for effective therapy in RA.28 Another study34 involving 711 patients with early RA diagnosed from over 1795 patients with symptoms of arthritis for less than 12 months showed that patients starting any treatment within 12 months since the onset of first symptoms or receiving disease-modifying agents (DMARDs) within 3 months from disease onset were significantly more likely to achieve disease remission (OR 2.03 (1.25–3.30) and 1.65 (1.06 to 2.55), respectively). The TICORA study35 showed the potential disease-modifying effect of intensive management versus routine care (mean disease duration: 16 months). Patients under tight control showed improved disease activity (p<0.0001), radiographic disease progression (p=0.02), disability (p=0.0025) and quality of life at no additional cost.

Figure 1

Early use of DMARDs is associated with significantly higher benefit for long-term outcomes in immune-mediated disease, such as joint damage in rheumatoid arthritis (A) and disability in multiple sclerosis (MS) (B).  EDSS, Expanded Disability Status Scale; RA, rheumatoid arthritis.

Table 1

Relevant trials on early intervention in RA

Early intervention in CD: what is the level of evidence?

Randomised controlled trials

Currently, no specific controlled trials addressing early intervention in CD compared with late CD have been conducted (table 2). There is evidence that the increase in use of conventional immunosuppressants in all patients with CD at diagnosis is not able to change the natural history of CD, in terms of complications and risk of surgery.36 The RCT AZTEC involving 53 newly diagnosed patients with CD (<8 weeks) did not show significant differences in terms of steroid-free remission after 76 weeks of treatment between azathioprine and placebo (p=0.48). Similarly, the open-label RAPID trial in adult patients with CD with a diagnosis of CD <6 months showed no differences in terms of clinical remission (p=0.69), but patients receiving azathioprine earlier were at significantly less risk for perianal surgery at month 36, demonstrating that early treatment may play a role in preventing CD-related damage in the long term,19 even if there is no advantage in terms of symptom control.

Table 2

List of relevant RCT data on early intervention in CD

Several post hoc analyses on trial populations have been conducted, results of which may suggest that early intervention with monoclonal antibodies could be more effective in reversing long-term disease progression to bowel damage than the conventional step-up approach.

A post hoc analysis of the SONIC trial23 made by stratifying patients according to disease duration defined by the Paris consensus demonstrated higher efficacy in terms of clinical remission, clinical remission+mucosal healing and clinical remission+mucosal healing+normal C-reactive protein (CRP), compared with non-early-CD population.16 Similarly, the EXTEND trial24 reported 33% of deep remission in early CD (diagnosed within 2 years), 20% in patients with 2–5 years of disease duration and 16% or patients with disease duration >5 years.

Patients enrolled in the CHARM trial25 37 with shorter disease duration at baseline were more likely to be in clinical remission at week 56 (OR 0.97, p=0.046) and maintained higher remission rates than patients with longer disease duration through 3 years of treatment, with significantly lower rate of serious adverse events compared with the other disease duration groups.37 Similar findings come from the STEP-UP/TOP-DOWN trial (SUTD)17 (average disease duration of 6 months): patients treated with TD strategy were more likely to achieve corticosteroid-free remission at week 26 (p=0.0062) and week 52 (p=0.0278), with more rapid drop of CRP (p=0.0244), and absence of ulcerations on endoscopy at week 104 (p=<0.001).

Cohort studies

There are also several real-life retrospective cohort studies investigating the effects of early treatment in patients with CD (table 3). Recently, data from the Swiss Cohort showed that early treatment was associated with reduced risk of bowel strictures (p<0.05), intestinal surgery (HR 0.322, p=0.005), perianal surgery (HR 0.361, p=0.042) and development of any complication (HR 0.567, p=0.006)38 One study on 168 newly diagnosed patients with CD in South Korea39 showed that early use of thiopurines (within 6 months of diagnosis) was associated with higher rates of clinical and steroid-free remission (p=0.043 and p=0.035, respectively). Similar data were found in the DELTA Cohort between patients starting anti-TNF within 16 months since diagnosis and those receiving the first anti-TNF later.40 The CONNECT study from South Korea,41 instead, showed that early use of thiopurines (median time 1.5±1.9 months since diagnosis) was associated with lower rates of CD-related surgery (24.1 vs 36.4%; p<0.001) and longer time to surgery (12 vs 9 months, p=0.017) than in patients treated with a conventional step-up approach. Other retrospective studies9 38 42 show that early use of immunosuppressants and/or biologics in the first 18–24 months since diagnosis is more likely to induce mucosal healing and reduce the time for surgery. Although these studies are strongly limited by the retrospective design and the highly heterogeneous study population, they seem to support the rationale for early treatment in patients with CD.

Table 3

List of relevant cohort studies on early intervention in CD

Early intervention in CD: where should we go?

There is an urgent need for prospective controlled trials clearly demonstrating that appropriate early treatment strategies can change the natural course of CD and restore bowel integrity and normal quality of life, by preventing the progression to irreversible bowel damage. This new study design in CD needs clear definitions and outcomes that remain to be validated. However, there is increasing evidence supporting that early diagnosis and treatment in patients with no bowel damage may achieve long-term remission and avoid complications over time. Pariente et al showed that patients with less than 2 years since diagnosis have less bowel damage compared with those with longer disease duration.4 More recently, Fiorino et al demonstrated that bowel damage at diagnosis is associated with a higher risk of complications and surgery and hospitalisations in newly diagnosed patients with CD.6 Therefore, disease modification trials need to consider appropriate patient selection. The Paris definition of early disease27 is currently the most accurate definition of patients with early CD, since disease duration less than 18 months and absence of disease complications and no previous use of DMARDs may identify patients with the highest chance of achieving all treatment goals. However, this definition does not include objective parameters of disease activity (such as biomarkers or mucosal lesions), which are important to identify patients more likely to respond to medical therapy, as demonstrated by the SONIC trial.16 23 Those treatment goals can differ concerning short-term and long-term outcomes that could be objective indicators for treatment success. In the short term, it is demonstrated that deep remission, defined as the combination of clinical and endoscopic remission, is associated with longer duration of remission, less risk for hospitalisation and surgery and better quality of life.24 43 In the long term, prevention of bowel damage progression,3 4 defined as the development of CD-related complications, such as strictures, fistulas and abscesses, leading to major surgery, intestinal dysfunction and prevention from disability,1 2 44 45 should be the main goals. This approach has been recently investigated in the REACT trial,46 an open-label cluster randomised trial randomising 1084 patients to receive early treatment with adalimumab+azathioprine or MTX after one ineffective course of steroids (no response or relapse within 12 weeks after withdrawal) or conventional step-up approach. Of these patients, 921 patients and 729 completed 12 months and 24 months follow-up in early treatment compared with 806 and 675 completing these follow-up times in the conventional group. The median disease duration was 149 months vs 158 in the ET and control group, respectively, despite no differences in terms of the primary outcome (clinical remission, defined as a CD activity index (CDAI)<150) found at months 12 or 24 in the ET and control groups (66.0% vs 61.9%, p=0.5169, and 73.1% and 65.1%, p=0.0829, respectively). In terms of secondary outcomes of this trial, the ET strategy was associated with a significant lower risk of surgery (HR=0.69, p=0.0314), serious disease complications (HR=0.73, p=0.0005) and combined hospital admission, disease complications and hospitalisation rate (HR=0.73, p=0.0003), suggesting that early therapy may change the disease course and avoid further complications even in the short term. Although not all aspects of the Paris criteria for early disease were met in this trial, those data support that early intervention and a tight control approach can reduce the risk of bowel damage progression and may change the natural course of CD. However, spontaneous remission rates expected in patients with CD,47 heterogeneous probability to progress during the course of the disease5 48 49 and different baseline characteristics that may impact on the natural history of the disease (eg, age, disease location, smoking, etc)50 51 should be taken into account in order to select the best patients’ population who may benefit of early treatment, excluding those at risk of overtreatment in a more long-term benign course. This is also true for patients with RA, since the selection of patients with certain baseline characteristics at diagnosis (such as CRP and erythrocyte sedimentation rate levels) may benefit from different therapeutic strategies.52 Besides evidence that supports early treatment (even within 3 months since diagnosis) may positively impact on the disease course and joint damage progression, a considerable number of patients may benefit from conventional therapy and more conservative step-up approach (up to 59% in the COMET trial).33 Similar data on predictive factors for disease progression at diagnosis are needed also in patients with CD.

More recently, in the phase 3 randomised controlled CALM trial,53 244 patients naïve to immunosuppressants and biologics with a mean disease duration of 0.95±1.98 years were randomised to be followed up and eventually receive stepwise therapy with adalimumab, based on biomarkers of inflammation, such as CRP and faecal calprotectin (active group) or on clinical symptoms (control group). At week 48, a significant greater proportion of patients treated with the treat-to-target approach achieved mucosal healing (45.9% vs 30.3, p<0.01), deep remission (36.9 vs 23.0, p=0.014), normalisation of CRP, faecal calprotectin and endoscopy (29.5% vs 15.6%, p=0.006) and endoscopic remission (45.9% vs 30.3%, p=0.01), whereas no differences were found for Crohn’s Disease Endoscopic Index of Severity=0 (18.0% vs 16.4%, p=0.72). This is the first prospective RCT supporting the treat-to-target approach in patients with early CD based on objective signs of inflammation.

Another key point for a disease modification trial is the appropriate follow-up period. The most important RCTs are usually conducted for 1 year to test short-term outcomes for efficacy. However, to achieve clear evidence on the short-term and long term outcomes, a longer trial duration is required. Both the SUTD and the REACT trial were conducted over at least 2 years to detect the impact of treatment strategies in the study population. The REACT trial showed that long-term early treatment strategies combined with tight control is able to reduce disease complications and the need of surgery at month 24. Even if disease progression may take longer, several studies show that bowel damage may occur quite early,5 6 54 supporting the hypothesis that a 2-year period of follow-up may be sensitive enough to detect the impact of disease modification strategies in patients with CD.

Finally, the identification and selection of patients who may benefit from early intervention is crucial. Recently, Nuij et al40 found no differences in terms of inflammatory bowel disease (IBD)-related complications and mucosal healing in the Delta cohort, evaluating patients treated with anti-TNF within 16 months from diagnosis (57% of the total study population) compared with those receiving anti-TNF later. Additionally, no significant difference was shown when more stringent criteria of 12 months since diagnosis were applied. Given the limitations of this study (retrospective subgroup analysis, underpowered sample size), the authors concluded that the main explanation for the lack of benefit of early treatment in this cohort lies in the absence of clear criteria to select patients who could be expected to benefit from early anti-TNF therapy. Kugathasan et al recently showed that early anti-TNF use is associated with less penetrating complications (HR 0.30, 95% CI 0.10 to 0.89; p=0.0296) but not less stricturing complications (HR 1.13, 95% CI 0.51 to 2.51; p=0.76).22 These data suggest that early intervention may not be appropriate in all patients, even though we do not have clear predictors to discriminate the right patient for the right strategy. These data should be interpreted cautiously because of the uncontrolled study design and because of the limited study population on whom these analyses were performed, since it may be underpowered. Further data from prospective randomised trials are needed. The data from the RISK study22 may suggest that the combination of demographic, genetic, epigenetic and other parameters may identify patients more at risk of complicated evolution of CD who will benefit of an early treatment with anti-TNF or other effective agents, although these data need confirmation by a prospective randomised clinical trial.

There are two best options for a disease modification prospective trial design. One may be similar to the SUTD trial,17 including all patients with early CD according to the Paris definition, randomised to receive early treatment with combined biological agents (not limited to anti-TNF) and immunomodulators or conventional step-up approach (steroids, then immunomodulators, then biological agents) by using deep remission (clinical+biomarker+ endoscopic definitions of remission) at week 52 as a short-term outcome and prevention of bowel damage and disability at month 24 as a long-term outcome. The second proposed study design may be more similar to the REACT46 and CALM trials55 by evaluating the same treatment goals by randomising patients to be treated earlier with effective therapy (steroids, then biologics+immunosuppressants in case of disease progression or lack of control of inflammation) or to be treated by conventional step-up approach. In both scenarios, outcome measures should be different from those used in previous trials. First, the CDAI should be replaced by patient-reported outcomes, as recently requested by regulatory agencies,56 57 that should include the reporting of symptoms related to the disease and other outcome measures impacting on normal life, such as mood disorders, work impairment, fatigue and quality of life.58 Second, objective assessment of inflammation and bowel damage should include specific biomarkers (such as CRP and faecal calprotectin), endoscopy and imaging techniques, such as MRI.59 Third, the long-term outcomes should be measured by indexes for bowel damage (structural and functional), such as the Lémann Index4 and the recently developed disability index.1 This comprehensive measure of the impact of CD on a patient’s life would ultimately demonstrate the real power of early intervention and therapeutic strategies in changing the natural history of CD in the long term.

Early intervention may allow dose decreases or drug withdrawal

Another relevant key point for disease modification trials may be to identify the right scenario for dose reduction or withdrawal of therapy. DMARDs are indeed effective in maintaining remission in the long term, but monoclonal antibodies and new biological agents are limited by high costs and potential side effects. A recent systematic review of 45 original studies of DMARD dose reduction in patients with RA60 (15 RCTs and 30 observational studies) from four research databases showed that DMARD dose reduction can be considered in all patients with RA who achieve stable (eg, ≥6 months) low disease activity or remission. The best strategies may be disease-activity-guided dose optimisation and fixed dose reduction, since direct DMARD discontinuation (without restarting) results in a high flare rate, worse physical functioning and more joint damage. Disease activity should also be monitored closely during DMARD dose reduction, and the dose should be increased to the lowest effective dose, as soon as disease relapses. No clear predictors have so far been identified to understand which patients would benefit most from this strategy.60 Kuijper et al recently investigated the possibility of dose reduction of DMARDs after successful early treatment in a large cohort of patients with RA. Patients from the Rotterdam Cohort were randomised to initial treatment with triple DMARD therapy with steroid bridging or MTX monotherapy with steroid bridging; patients were evaluated every 3 months. Sustained clinical remission was achieved in 57% of patients during 2 years of follow-up, and 21% achieved drug-free remission. Among those who flared after DMARD dose reduction (41% and 37% within 1 year in the two study groups, respectively), 65% of patients who tapered DMARDs reachieved remission within 6 months after treatment intensification.61

Initial data have been coming from IBD populations,62–68 showing that a therapeutic drug-monitoring-driven dose reduction with tight control for the early treatment of further relapses during dose reduction may be an effective strategy to efficiently treat patients with CD, saving unnecessary costs for maintenance of disease control. In particular, early combined immunosuppression with anti-TNF and immunomodulators can be more effective than monotherapy in achieving disease control and clinical and endoscopic remission23 but may also be effective in further therapy dose reduction. Drobne et al69 retrospectively analysed 223 patients (65 received infliximab monotherapy, 158 received infliximab and an immunomodulator). They found higher trough levels of infliximab (p=0.02) and less anti-drug antibody (ATI) formation (p=0.01) in those receiving combined immunosuppression. Interestingly, among patients receiving combined therapy, levels of infliximab remained stable after immunomodulators were withdrawn (p=0.70), and after withdrawal of immunomodulators, just 38% of patients required increasing doses of infliximab, whereas 18% discontinued infliximab. At the time of immunomodulator withdrawal, trough levels of infliximab and CRP were most strongly associated with response to infliximab thereafter. The authors found that withdrawal of immunomodulators after at least 6 months of combined therapy may be enough to carry over the benefits observed in this study and may be associated with effective anti-TNF discontinuation. These data need to be confirmed by further prospective disease-modifying trials as an extension of early treatment studies in patients who achieve full disease control. At the moment, one study promoted by the GETAID, the CURE study, is ongoing.70 This study will evaluate early treatment with adalimumab in patients with CD, the rate of deep remission in those patients and potential predictors for treatment withdrawal after full disease control is achieved.70 The results of this study may open the way to larger prospective studies finally demonstrating whether this approach should be used in clinical practice.

Conclusion and perspectives

Achieving sustained deep remission in early CD represents the main goal in the management of patients with CD. Using the most effective strategy in the therapeutic window of opportunity is key for avoiding disease progression and for restoring normal quality of life in patients with CD, although evidence supporting this statement is lacking. Moreover, early therapeutic intervention with complete disease control may allow dose reduction or even treatment withdrawal in the maintenance phase, reducing side effects, costs and also improving quality of life (figure 2). The arrival of new drugs such as small molecules (JAK inhibitors and S1P modulators) may represent a unique opportunity to initiate disease modification trials in early CD to identify the right strategy for the right patient at the right time. Identification of the right patients for such trials may be challenging, so the Red Flags Index71 allowing early referral may facilitate that. RA data strongly suggest very early disease (<4 months from the onset of symptoms) seems to be highly responsive to therapy,33 34 and effective treatment strategies may dramatically change the natural history of disease, increasing the chance to recover from inflammation and avoid progression to irreversible joint damage. As in RA, targeting very early CD might be the next step and the ultimate goal in the near future.

Figure 2

Pros and cons of early intervention in CD. Early treatment may positively impact on several disease outcomes and prevent disease progression. However, data are needed to demonstrate that these benefits overcome higher costs, risk for adverse events and the risk of overtreatment in up to 20% of patients. Anti-TNF, antitumour necrosis factor α; CD, Chron’s disease; RCT, randomised controlled trial.

Supplementary Material

Supplementary Figure 1

Supplementary Material

Supplementary Figure 2


We gratefully thank Marie Cheeseman for English revision.



  • Contributors All the authors drafted the manuscript and critically revised and accepted the final version.

  • Competing interests SD has served as a speaker, consultant and advisory board member for Schering-Plough, Abbott Laboratories, Merck, UCB-pharma, Ferring, Cellerix, Millenium Takeda, Nycomed, Pharmacosmos, Actelion, Danone, Alpha Wasserman, Genentech, Grunenthal, Pfizer, Astra Zeneca, Novo Nordisk, Cosmo Pharmaceuticals, Vifor and Johnson & Johnson; GF served as a consultant and advisory board member for MSD, AbbVie, Takeda, Janssen, Mundipharma, Celltrion, Alfa Wassermann, Sandoz and Pfizer; LP-B: Honoraria from Merck, Abbvie, Janssen, Genentech, Mitsubishi, Ferring, Norgine, Tillots, Vifor, Hospira/Pfizer, Celltrion, Takeda, Biogaran, Boerhinger-Ingelheim, Lilly, HAC-Pharma, Index Pharmaceuticals, Amgen, Sandoz, Forward Pharma GmbH, Celgene, Biogen, Lycera and Samsung Bioepis.

  • Provenance and peer review Not commissioned; internally peer reviewed.