Article Text

Download PDFPDF

Original article
Natural disease course of Crohn’s disease during the first 5 years after diagnosis in a European population-based inception cohort: an Epi-IBD study
  1. Johan Burisch1,
  2. Gediminas Kiudelis2,
  3. Limas Kupcinskas2,3,
  4. Hendrika Adriana Linda Kievit4,
  5. Karina Winther Andersen5,
  6. Vibeke Andersen5,6,
  7. Riina Salupere7,
  8. Natalia Pedersen8,
  9. Jens Kjeldsen9,
  10. Renata D’Incà10,
  11. Daniela Valpiani11,
  12. Doron Schwartz12,
  13. Selwyn Odes12,
  14. Jóngerð Olsen13,
  15. Kári Rubek Nielsen13,
  16. Zsuzsanna Vegh14,
  17. Peter Laszlo Lakatos14,15,
  18. Alina Toca16,
  19. Svetlana Turcan16,
  20. Konstantinos H Katsanos17,
  21. Dimitrios K Christodoulou17,
  22. Mathurin Fumery18,
  23. Corinne Gower-Rousseau19,20,
  24. Stefania Chetcuti Zammit21,
  25. Pierre Ellul21,
  26. Carl Eriksson22,
  27. Jonas Halfvarson22,
  28. Fernando Jose Magro23,24,
  29. Dana Duricova25,
  30. Martin Bortlik25,26,
  31. Alberto Fernandez27,
  32. Vicent Hernández28,
  33. Sally Myers29,
  34. Shaji Sebastian29,
  35. Pia Oksanen30,31,
  36. Pekka Collin31,
  37. Adrian Goldis32,
  38. Ravi Misra33,
  39. Naila Arebi33,
  40. Ioannis P Kaimakliotis34,
  41. Inna Nikuina35,
  42. Elena Belousova35,
  43. Marko Brinar36,
  44. Silvija Cukovic-Cavka36,
  45. Ebbe Langholz37,
  46. Pia Munkholm1
  47. for the Epi-IBD group
    1. 1 Department of Gastroenterology, Nordsjællands Hospital, University of Copenhagen, Frederikssund, Denmark
    2. 2 Institute for Digestive Research, Medical Academy, Lithuanian University of Health Sciences, Kaunas, Lithuania
    3. 3 Department of Gastroenterology, Medical Academy, Lithuanian University of Health Sciences, Kaunas, Lithuania
    4. 4 Department of Medicine, Herning Central Hospital, Herning, Denmark
    5. 5 Medical Department, Regional Hospital of Viborg, Viborg, Midtjylland, Denmark
    6. 6 Focused research unit for Molecular Diagnostic and Clinical Research (MOK), IRS-Center Sonderjylland, Hospital of Southern Jutland, Aabenraa, Denmark
    7. 7 Division of Gastroenterology, Tartu University Hospital, University of Tarty, Tartu, Estonia
    8. 8 Gastroenterology Department, Slagelse Hospital, Slagelse, Sjaelland, Denmark
    9. 9 Gastroenterology Department, Odense University Hospital, Odense, Denmark
    10. 10 Department of Surgical, Oncological and Gastroenterological Sciences, Azienda, University of Padua, Padova, Italy
    11. 11 U.O. Gastroenterologia ed Endoscopia digestiva, Hospital Morgagni Pierantoni, Forlì, Italy
    12. 12 Department of Gastroenterology and Hepatology, Soroka Medical Center and Ben Gurion University of the Negev, Beer Sheva, Israel
    13. 13 Medical Department, The National Hospital of the Faroe Islands, Thorshavn, Faroe Islands
    14. 14 1st Department of Medicine, Semmelweis University, Budapest, Hungary
    15. 15 Division of Gastroenterology, McGill University Health Center, Montreal, Canada
    16. 16 Department of Gastroenterology, State University of Medicine and Pharmacy of the Republic of Moldova, Chisinau, Republic of Moldova
    17. 17 Department of Gastroenterology, University Hospital of Ioannina, Ioannina, Greece
    18. 18 Gastroenterology Unit, Epimad Registry, CHU Amiens Sud, Avenue Laennec-Salouel, Amiens University Hospital, Amiens, France
    19. 19 Public Health, Epidemiology and Economic Health, Registre Epimad, Lille University and Hospital, Lille, France
    20. 20 Lille Inflammation Research International Center LIRIC, Lille University, Lille, France
    21. 21 Division of Gastroenterology, Mater Dei Hospital, Msida, Malta
    22. 22 Department of Gastroenterology, Faculty of Medicine and Health, Örebro University, Örebro, Sweden
    23. 23 Department of Gastroenterology, Centro Hospitalar de São João EPE, Porto, Portugal
    24. 24 Department of Biomedicine, Institute of Pharmacology, Faculty of Medicine of Porto University, Porto, Portugal
    25. 25 IBD Clinical and Research Centre, ISCARE, Prague, Czech Republic
    26. 26 Institute of Pharmacology, 1st Faculty of Medicine, Charles University in Prague, Prague, Czech Republic
    27. 27 Department of Gastroenterology, Hospital POVISA, Vigo, Spain
    28. 28 Department of Gastroenterology, Hospital Alvaro Cunqueiro. Instituto Investigación Sanitaria Galicia Sur. EOXI de Vigo, Vigo, Spain
    29. 29 IBD Unit, Hull and East Yorkshire NHS Trust, Hull, UK
    30. 30 Department of Gastroenterology and Alimentary Tract Surgery, Tampere University Hospital, Tampere, Finland
    31. 31 University of Tampere, Tampere, Finland
    32. 32 Clinic of Gastroenterology, University of Medicine ‘Victor Babes’, Timisoara, Romania
    33. 33 IBD Department, Imperial College London, London, UK
    34. 34 Nicosia private practice, Nicosia, Cyprus
    35. 35 Department of Gastroenterology, Moscow Regional Research Clinical Institute, Moscow, Russian Federation
    36. 36 Division of Gastroenterology and Hepatology, University Hospital Center Zagreb, Zagreb, Croatia
    37. 37 Department of Gastroenterology, Herlev and Gentofte Hospital, University of Copenhagen, Herlev, Denmark
    1. Correspondence to Dr Johan Burisch, Department of Gastroenterology, North Zealand University Hospital, 3600 Frederikssund, Denmark; burisch{at}


    Objective The Epi-IBD cohort is a prospective population-based inception cohort of unselected patients with inflammatory bowel disease from 29 European centres covering a background population of almost 10 million people. The aim of this study was to assess the 5-year outcome and disease course of patients with Crohn’s disease (CD).

    Design Patients were followed up prospectively from the time of diagnosis, including collection of their clinical data, demographics, disease activity, medical therapy, surgery, cancers and deaths. Associations between outcomes and multiple covariates were analysed by Cox regression analysis.

    Results In total, 488 patients were included in the study. During follow-up, 107 (22%) patients received surgery, while 176 (36%) patients were hospitalised because of CD. A total of 49 (14%) patients diagnosed with non-stricturing, non-penetrating disease progressed to either stricturing and/or penetrating disease. These rates did not differ between patients from Western and Eastern Europe. However, significant geographic differences were noted regarding treatment: more patients in Western Europe received biological therapy (33%) and immunomodulators (66%) than did those in Eastern Europe (14% and 54%, respectively, P<0.01), while more Eastern European patients received 5-aminosalicylates (90% vs 56%, P<0.05). Treatment with immunomodulators reduced the risk of surgery (HR: 0.4, 95% CI 0.2 to 0.6) and hospitalisation (HR: 0.3, 95% CI 0.2 to 0.5).

    Conclusion Despite patients being treated early and frequently with immunomodulators and biological therapy in Western Europe, 5-year outcomes including surgery and phenotype progression in this cohort were comparable across Western and Eastern Europe. Differences in treatment strategies between Western and Eastern European centres did not affect the disease course. Treatment with immunomodulators reduced the risk of surgery and hospitalisation.

    • crohn’s disease
    • surgery for Ibd
    • epidemiology
    View Full Text

    Statistics from

    Significance of this study

    What is already known on this subject?

    • Treatment options and strategies for Crohn’s disease (CD) have changed during the past decades, including an earlier and more frequent use of immunomodulators and biological therapy.

    • Whether these changes in treatment have improved the long-term outcome of CD outside the settings of clinical trials remains unknown.

    What are the new findings?

    • Treatment of patients with CD varies across Europe, in particular Eastern European patients receive biological therapy less frequently than Western European patients.

    • Despite these geographic differences in treatment, the course of disease—including rates of patients undergoing surgery, developing stricturing or penetrating disease phenotype or being hospitalised—did not differ between Western and Eastern European patients.

    • Five-year surgery, hospitalisation and phenotype progression rates have not altered significantly in this time and are similar to cohorts from 20 years ago.

    • The use of immunomodulators reduces the risk of surgery and hospitalisation in patients with CD.

    Significance of this study

    How might it impact on clinical practice in the foreseeable future?

    • Future studies on risk stratification as well as on whether treating to target influences the disease course significantly, are needed to further improve the long-term disease course of CD.


    Crohn’s disease (CD) is a chronic inflammatory bowel disease (IBD) of unknown aetiology. Patients are often young and due to the progressive nature of the disease, CD is associated with significant disability and morbidity.1 Patients require anti-inflammatory treatment and sometimes surgery, in order to induce and maintain remission. New therapeutic approaches, including early and more aggressive intervention with immunomodulators and biological agents and treating to achieve endoscopic healing or normalisation of biomarkers, have been introduced over the last two decades.2 By aiming to block disease progression and avoid bowel damage and disability, they offer the possibility of improving the natural history of IBD,3 especially with regard to the risk of hospitalisation and surgery.

    However, it remains uncertain whether the adoption of these treatment strategies have resulted in an improved long-term outcome of CD.4 To assess their impact on disease outcome, population-based prospective cohorts of unselected patients representing the broad spectrum of the disease are necessary in order to build the most accurate picture of effectiveness and practice regarding medication and surgery in the community setting. Yet to date, few CD cohorts from the era of biological therapy and widespread use of immunomodulators have been studied.

    The Epi-IBD (formerly EpiCom) study is a prospective population-based inception cohort of unselected IBD patients investigating the occurrence, disease course and prognosis of IBD in Europe.5 The aim of the present study was to evaluate the disease course and the impact of treatment choices on disease course of CD during the first 5 years following diagnosis by analysing the Epi-IBD cohort. A further aim was to assess a possible West-East gradient in treatment strategies and disease outcomes across Europe within this cohort.


    Study population and design

    The Epi-IBD cohort is a prospective population-based inception cohort of IBD patients diagnosed in 2010 from 22 European countries and Israel.6 All participating centres were required to have a well-defined primary catchment area with up-to-date population data, including age and sex distribution. They were also required to have an established network of gastroenterologists, colorectal surgeons and general practitioners within the uptake area, who were contacted twice during the inclusion period to ensure complete coverage and recruitment of patients. Case ascertainment methods, diagnostic criteria, inclusion period and patient data were all standardised.

    Of the original cohort, five centres were unable to perform follow-up for technical or logistical reasons. Two centres (France, Malta) joined the study group at a later time, both centres having included incident patients in prospective inception cohorts in 2010 in parallel. Consequently, 29 centres from 8 Eastern and 13 Western European countries, including Israel, participated in the study (see online supplementary file). The study population consisted of 1289 patients with IBD, 488 with CD, 717 with ulcerative colitis (UC) and 84 with IBD Unclassified (IBDU), all recruited within well-described geographical areas covering a total background population of 9.7 million people (2.6 million in Eastern Europe and 7.1 million in Western Europe).

    Supplementary file 1

    Classifications and definitions

    Incident patients diagnosed with IBD according to the Copenhagen Diagnostic Criteria 7 8 between 1 January and 31 December 2010, aged ≥15 years and living in the predefined catchment areas at the time of diagnosis, were prospectively included. The date of inclusion was the date of diagnosis. Few patients experienced a change in their diagnosis during follow-up, but those who did were reclassified as having the new diagnosis from the date of the original diagnosis.

    Disease location and behaviour for CD were defined according to the Montreal Classification. 9 Complicated CD behaviour was defined as patients who developed intestinal strictures (B2) or internal fistulas or abscesses (B3). Progression in disease behaviour was defined as the development of complicated CD in patients with non-stricturing, non-penetrating phenotype (B1) at diagnosis. Patients were categorised according to the most intense level of disease behaviour observed in them during the study period. Progression in disease location was defined as an extension of the involved segments of the gastrointestinal tract, for example, from terminal ileitis to an involvement of both terminal ileum and segments of the colon. Surgery was defined as surgical resection because of intestinal inflammation or a CD-related complication such as a stenosis, fistula or perforation. Perianal surgeries included, among others, fistulotomy, seton placement and perianal abscess incision. Hospitalisation was defined as either admission for CD-related surgery (surgical hospitalisation) or other CD-related complications (medical hospitalisation). Elective admissions, for example, for endoscopy procedures or drug administration, were excluded.

    Patients were defined as users of various IBD drugs on the first day of administration, regardless of treatment duration. Treatment was categorised according to five levels of ascending therapeutic potency: 5-aminosalicylates (5-ASA) (oral and/or topical 5-ASA treatment ±topical steroids), corticosteroids (oral steroids±5 ASA or topical steroids), immunomodulators (azathioprine, 6-mercaptopurine, ciclosporin or methotrexate ±steroids), biologicals (infliximab or adalimumab in combination with any of the above) and surgery (major abdominal surgery due to CD regardless of medical treatment prior to surgery). Immunomodulators (azathioprine, 6-mercaptopurine, ciclosporin and methotrexate) were combined in one category because 95% of patients treated with immunomodulators received thiopurines. An early need for immunomodulators or biological therapy was defined as the initiation of these drugs within 6 months of diagnosis. Early need for corticosteroids was defined as patients started on ≥40 mg corticosteroids within 30 days of diagnosis.

    Disease activity of CD was measured using the Harvey Bradshaw Index (HBI).10 A HBI score of <5 was defined as remission, 5–7 as mildly active disease, 8–16 as moderately active disease and ≥16 as severely active disease.11 Causes of death and cancers were categorised according to the 10th revision of the International Classification of Diseases.12

    Data collection and validity

    Patients were followed prospectively from diagnosis for 5 years or until the date of their emigration or death. Data regarding demographics, disease activity, medical therapy including dose, date of initiation and date of cessation, surgery, hospitalisation, disease classification, cancers and deaths were collected and entered prospectively in the web-based Epi-IBD database.13 A follow-up period of 5 years (with a 3-month margin either side of the end date) was chosen in order to assess the 5-year outcome of the cohort. Measures to secure data validity have been thoroughly described elsewhere.5 In short, data validity was secured by built-in control and validation tests, locked diagnostic criteria in the database, manual data standardisation and random audits of case ascertainment and data quality.

    Statistical analysis

    Statistical analyses were performed using SAS software V.9.4 (SAS, Cary, North Carolina, USA). Differences regarding time to events were compared using the Wilcoxon two sample test. Continuous variables are expressed as median (IQR) unless otherwise stated. Groups were compared using Χ2 test or Fischer exact test, where appropriate. A P value of <0.05 was considered statistically significant.

    Possible associations between primary endpoints (progression in behaviour, progression in location, surgery or hospitalisation) and multiple covariates were analysed by stepwise Cox regression analysis using the proportional hazards assumption and associations were visualised using Kaplan-Meier plots in order to test these assumptions. As the aim was to describe the disease course following diagnosis, only events occurring after the date of diagnosis were included in the Cox regression analysis.

    The following covariates were included in the statistical model: age at diagnosis (continuous variable), sex, geographic region (Western vs Eastern Europe), disease behaviour, disease location, smoking status (never, currently, former), diagnostic delay (continuous variable), need for early corticosteroids, treatment with immunomodulators and treatment with biologicals. The reported HRs are thus all adjusted. The use of immunomodulators and biologicals were included in the regression model as time-dependent variables and with an initial lag time of 3 months for immunomodulators and 2 months for biologicals, meaning that only treatments prior to an event and lasting longer than the lag time would count as active treatment. Regarding disease location, due to the low number of such patients all those with upper GI involvement with or without involvement of the colon and/or terminal ileum were categorised as L4. We also included age and diagnostic delay as categorical variables as well as smoking as a binary variable (current vs former/never smoker), with no impact on the results.

    We created two prevalence plots for assessing the prevalence of treatment types. In the first variant, each patient was classified, on any given day since their diagnosis, as being in one of the treatment steps according to the treatment(s) received on that specific day. In the case of the patient receiving combination treatment, the level assigned was that of the most potent treatment. In the second variant (shown in the online supplementary file), in the case of combination therapy the patient was counted in each of the relevant treatment steps. In both plots, patients undergoing surgery were considered as remaining in this treatment step for 1 year.


    A total of 476 patients aged 15 years or older were initially diagnosed with CD. During follow-up, 10 patients initially diagnosed with UC and 8 diagnosed with IBDU had their diagnoses changed to CD and were included in the cohort. The changes in diagnosis occurred after a median of 23 months (IQR 5–40). Six patients initially diagnosed with CD received a new diagnosis of UC and were excluded. Therefore, the cohort consisted of 488 patients with CD. Of these, 404 (83%) were diagnosed in Western European and 84 (17%) in Eastern European centres. The median follow-up time was identical in both regions at 63 months (IQR 39–63). Patients’ sociodemographic characteristics (table 1) did not differ significantly between the two geographic regions.

    Table 1

    Characteristics of incident patients with Crohn’s disease from the Epi-IBD-cohort

    The proportion of patients undergoing various diagnostic procedures during follow-up is shown in figure 1. While all patients were investigated during the first year following their diagnosis, 217 (44%) patients were not examined again during the following 4 years. Data on disease activity during follow-up were available for analysis in 433 (89%) patients with CD. The proportion of patients in remission increased from 33% during the first year of disease to 77% in the fifth year of follow-up (figure 2).

    Figure 1

    Proportion of patients with Crohn’s disease undergoing at least one endoscopic or imaging procedure during the first and the second-to-fifth years after diagnosis.

    Figure 2

    Distribution of disease activity among patients with Crohn’s disease during 5 years of follow-up.

    Disease behaviour

    A total of 141 (29%) patients had complicated CD at diagnosis. After 5 years of follow-up, this number increased to 190 (39%). The proportion of patients with changing disease behaviour is shown in figure 3A. Overall, 49 (14%) patients diagnosed with B1 progressed to either B2 or B3 with no geographic difference observed (figure 4). Most patients had received either immunomodulators (Western Europe: 19 (45%); Eastern Europe: 2 (29%)) or biologicals (Western Europe: 11 (26%); Eastern Europe: 1 (14%)) as the highest treatment step before disease progression. The median time to progression from B1 overall was 21 months (IQR 7–39); however, this time was shorter in Western (13 months, IQR 6–36) than in Eastern Europe (40 months, IQR 11–51) (P<0.05). Predictors associated with progression in disease behaviour are shown in table 2.

    Figure 3

    Changes in disease behaviour (A) and location (B) during 5 years of follow-up among patients with Crohn’s disease. The central column represents the proportion of patients within each category that changed to one of the other categories. B1: non-stricturing, non-penetrating; B2: stricturing; B3: penetrating; L1: terminal ileum; L2: colon; L3: terminal ileum and colon; L4: upper GI tract (all patients with upper GI involvement).

    Figure 4

    Cumulative probability for progressing from non-stricturing, non-penetrating to complicated disease behaviour (structuring or penetrating disease) in patients with Crohn’s disease in a European inception cohort.

    Table 2

    Predictors associated with progression in behaviour and location, hospitalisation and surgery in patients with Crohn’s disease in the Epi-IBD-cohort

    A total of 66 (14%) patients had perianal disease (perianal fistula or abscess) after 5 years of follow-up, of whom 46 (9%) already had perianal disease at diagnosis. Most of the 20 patients, who developed perianal disease had colonic involvement (L1=8 (40%), L2=5 (24%), L3=4 (29%), L4 +L1/L2=3 (15%)) and B1 phenotype (B1=12 (60%), B2=7 (35%), B3=1 (5%)). The median time to developing perianal disease was 2 months (IQR 1–5). No geographic differences regarding perianal disease were observed.

    Disease location

    A progression in disease location was observed in 58 (12%) patients, 49 (11%) in Western and 9 (11%) in Eastern European patients (P=NS). The changes in location during follow-up are shown in figure 3b. The median time to any change in location was 13 months (IQR 8–41), which did not differ depending on the initial location or between Western and Eastern European patients. Predictors associated with progression in location in patients with L1–L3 are shown in table 2.

    Medical treatment

    Only 14 (3%) patients did not receive any treatment, including surgery, during the follow-up period. The cumulative 5-year exposures to medical treatments are shown in figure 5 and online supplementary figure 1, while the median duration of treatment is shown in table 3. Overall, geographic differences in the use of 5-ASA, immunomodulators as well as biologicals, were observed. The distribution of patients across the defined treatment steps at any given time during follow-up is shown in figure 6 and online supplementary figure 2 and the pattern of changes between treatment steps is shown in online supplementary figure 3.

    Supplementary file 2

    Table 3

    Frequency and duration of oral treatment in patients with Crohn’s disease at 5-year follow-up in the Epi-IBD cohort

    Figure 5

    Cumulative 5-year exposures for medical treatment for patients with Crohn’s disease in a European inception cohort for the whole cohort as well as in (A) Eastern Europe and (B) Western Europe. 5-ASA, 5-aminosalicylates.

    Figure 6

    The time-varying distribution of patients with Crohn’s disease receiving different potencies of treatment on any given day during follow-up for the whole cohort as well as in (A) Eastern Europe and (B) Western Europe. At any time, the distribution totals 100%. Patients receiving combination therapy are categorised according to the highest treatment step. 5-ASA, 5-aminosalicylates.

    Regarding systemic corticosteroids (excluding budesonide), 46 (9%) patients received treatment for more than six consecutive months during the follow-up period (Western Europe: 36 (9%); Eastern Europe: 10 (12%), P=NS). Most patients had B1 (n=39, 85%). Of those patients, 12 (26%) patients received immunomodulators as the highest treatment step during follow-up (Western Europe: 9 (25%); Eastern Europe: 3 (30%)), 20 (43%) were treated with biologicals (Western Europe: 15 (41%); Eastern Europe: 5 (50%)), while 7 (15%) received surgery (Western Europe: 7 (19%); Eastern Europe: 0 (0%)).

    Biologicals were used in 144 (30%) of all patients during the follow-up period; however, they were used significantly more frequently in Western Europe (n=132, 33%) than in Eastern Europe (n=12, 14%) (P<0.01). Most patients received infliximab as the initial type of biologicals (infliximab: n=110, 76%; adalimumab: n=34, 24%). Of Western European patients, 88 (67%) patients had B1, 33 (25%) had B2 and 11 (8%) had B3, while in Eastern Europe 10 (83%) patients had B1 and two (17%) had B2. Treatment before biological therapy is shown in table 4.

    Table 4

    Treatment steps reached prior to biological therapy or surgery in a European inception cohort of patients with Crohn’s disease


    A total of 175 (36%) patients were hospitalised because of CD, of which 113 (23%) were hospitalised during the first year following diagnosis. Overall, 66 (38%) hospitalisations were in relation to surgery, while 109 (62%) were for medical reasons. In Western Europe, 147 (36%) patients were hospitalised compared with 28 (33%) patients in Eastern Europe (P=NS) (figure 7). The median time to first hospitalisation was 6 months (IQR 1–22) and the median number of hospitalisations per patient was one (IQR 1–2), which was similar across Western and Eastern Europe. Treatment prior to hospitalisation is shown in table 4. Predictors associated with hospitalisation are shown in table 2.

    Figure 7

    Cumulative probability for hospitalisation (A) and surgery (B) during the first 5 years of disease in a European population-based cohort of patients with Crohn’s disease.

    When looking only at hospitalisations for medical, CD-related reasons, 113 (24%) patients were hospitalised, including 66 (17%) patients during the first year following diagnosis. In Western Europe, 94 (24%) patients were hospitalised for the first time after a median of 9 months (IQR 4–25). In Eastern Europe, a similar proportion were hospitalised (n=18, 22%; P=NS) after a median of 10 months (range: 0–62) (P=NS) (figure 7). The median number of hospitalisations per patient was one (IQR 1–2), which was similar in Western and Eastern Europe. Predictors associated with medical hospitalisation are shown in table 2.


    During 5 years of follow-up, 107 patients with CD (22%) had a resection performed (including eight hemicolectomies and eight total colectomies), of which 62 (13%) occurred during the first year following diagnosis. These resections were performed in 89 Western European patients with CD (22%) and 18 Eastern European patients with CD (21%) (figure 7). The median time to first surgery was 7 months (IQR 1–30) and did not differ between Western and Eastern Europe. A total of 37 (35%) patients had B1, 44 (41%) had B2 and 26 (24%) patients had B3 at diagnosis, while 18 (49%) patients with B1 had progressed to B2/B3 prior to surgery. A second operation was performed in six (7%) patients, of which two (33%) had B2 and three (50%) had B3. All of these patients came from Western European centres. Predictors associated with surgery are shown in table 2.

    Cancer and death

    During follow-up, eight (2%) patients were diagnosed with eight cancers, a median time of 32 months after diagnosis (IQR 10–36). All were extraintestinal cancers, including one cervical carcinoma in situ (one bile duct cancer, one breast cancer, two urinary tract cancers, one thyroid cancer, one lymphoma and one cancer in the female genital organs).

    A total of 16 (3%) patients died during follow-up at a median 25 months (IQR 9–40) after diagnosis. Two patients died because of sepsis after CD surgery, while the remaining patients died of non-CD-related causes.


    Using a European population-based inception cohort of unselected patients with IBD, we have presented the contemporary disease course as well as treatment strategies used in CD. Despite an earlier and more aggressive treatment with immunomodulators and biological therapy, rates of surgery and disease progression were not significantly different from previous population-based inception cohorts 20 years ago. Furthermore, while we observed differences in choices of treatment between patients from Western and Eastern Europe, we found no difference regarding disease course and prognosis. Finally, treatment with immunomodulators significantly reduced the risk for surgery and hospitalisations.

    Earlier diagnoses, and thus a greater number of patients diagnosed with only mild disease, along with the earlier introduction of more aggressive immunomodulators and biological therapy, has led to a more conservative approach towards surgery by physicians14 and declining surgery rates for patients with CD during the past decades.4 15–18 However, surgical rates were falling prior to the advent of biological therapy14 and to date no causal link between this increased use of immunosuppressive therapy and the decreased need for surgery has been established. In fact, surgery rates in population-based cohorts during the last decade have remained stable at approximately 20% after 5 years,4 19–21 which is similar to the findings in this study as well as to population-based cohorts from the prebiological era.22 23 Similarly, the rate of patients progressing from B1 to B2 or B3 in this cohort are consistent with other population-based cohorts since the year 2000, which have found that approximately 20% of patients progress within the first 5 years of diagnosis.4 19 24 25

    Thiopurines and biological agents have both been shown to improve the disease course in patients with IBD26–28—however, the literature on this matter has proven inconsistent. Two recent population-based cohorts demonstrated a significant reduction in surgery rates in CD over time to be associated with increased and earlier use of thiopurines,16 21 while a beneficial effect on the early disease course could not be demonstrated in two randomised, controlled trials.29 30 In the present cohort, treatment with immunomodulators, as well as biologicals, was initiated early in the disease course.31 Immunomodulators were found to significantly reduce the risk of surgery and hospitalisation while it non-significantly reduced the risk of progression. It is worth mentioning that treatment with immunomodulators was included as a time-dependant variable in the analysis with a 3 month lag time, thereby avoiding bias due to patients receiving immunomodulators, for example, in the postoperative setting. The use of biological therapy, however, was not associated with either surgery or hospitalisation. Furthermore, and similar to the findings of a recent Dutch population-based cohort,4 no regional difference in disease course was found despite significantly more frequent use of biological therapy in Western than in Eastern Europe.

    The fact that biological therapy, despite its frequent use in this cohort, did not modify the course of the disease might be explained by the population-based design of the study. The treatment choices that were made reflect the real-life decision-making of gastroenterologists in Europe. Recent studies have demonstrated the beneficial effects of treating to target and of early combined immunosuppression.32 33 However, approaches such as these are not yet established in real-life practice and are hampered by local restrictions on, for example, when to initiate biological therapy. Most patients in this cohort received biological therapy after other drugs failed them, which might have influenced their efficacy when compared with protocolled studies with strict follow-up schemes of patients.

    In our study, smoking was not significantly associated with any of the outcomes of interest, including surgery. While a recent meta-analysis found that smokers have a more complicated disease course (surgery and/or flare-ups),34 the majority of the studies they included were not population-based cohorts and this might explain why ours and other population-based cohorts have not provided any evidence of an association with smoking.22 24 35

    In this study, we provide real-world information regarding the treatment choices, as well as changes in treatment over time, in a cohort representative of the whole IBD population. Half of the patients were treated with immunomodulators and a fifth of them with biological therapy on any given day during follow-up, figures similar to those found in other recent cohorts,4 20 although patients were started on immunosuppressants earlier in this study than in previous ones.19 21 Interestingly, while current guidelines do not recommend the use of 5-ASA for CD,36 3730%–40% of patients were at any given time during the follow-up period treated with this class of drugs, including 15%–20% as a monotherapy. One previous study identified a subgroup of patients with CD who seemed to benefit from 5-ASA.38 Furthermore, 5-ASA seems to reduce the risk of colorectal cancer.39 40 Both the prevalence and duration of 5-ASA treatment were higher in Eastern than in Western Europe. However, and as discussed previously, these apparent differences in choices of treatment did not result in different surgery, hospitalisation or phenotype progression rates.

    The strength of the Epi-IBD cohort is the prospective inclusion and follow-up of incident patients with IBD diagnosed within well-defined geographic areas. Diagnostic criteria, case ascertainment methods and the recorded data were all standardised and patients thereby made comparable in observation time. Several measures previously described6 ensured that all centres performed a population-based cohort study with good data quality and validity. The patients were unselected and represent the whole spectrum of disease severity; therefore, the choices of treatment in this cohort are the results of community effectiveness rather than the requirements of a randomised controlled trial and all occurred in a real-life clinical setting.

    Limitations of the study include the heterogeneity of the participating centres in terms of healthcare systems. Choices regarding investigations, medical and surgical treatment and the availability of such treatments, are closely linked to extramedical considerations and therefore the differences observed between Western and Eastern Europe might have been caused by considerable variations between national healthcare systems. This includes the fact that most patients did not undergo imaging within the first year following their diagnosis. Furthermore, the distribution of participating centres is skewed, with more centres located in Western Europe. As the Eastern European centres are in mostly low-incident areas,1 a majority of patients in this study originate in Western Europe. However, and similar to previous findings,41 the patient populations from Western and Eastern Europe were similar in terms of socioeconomic characteristics, disease classification and diagnostic procedures used and time to diagnosis.6 Therefore, we have no reason to believe the region of origin influenced the disease course. Finally, several patients were only investigated with endoscopy and/or imaging infrequently during the observation period and this might have influenced the findings regarding disease progression. Choices regarding investigations were, however, made according to the treating physician’s discretion.

    To conclude, in this prospective population-based cohort of unselected patients with CD, patients were treated earlier and more aggressively with immunomodulators and biological therapy as compared with cohorts from the beginning of the biological era. Yet the 5-year surgery, hospitalisation and phenotype progression rates have not altered significantly in this time and are similar to cohorts from twenty years ago. Furthermore, differences in treatment choices, including the use of biological therapy, between Western and Eastern European centres did not appear to influence the disease course. However, the use of immunomodulators improved the disease course with regard to the risk of surgery and hospitalisation. Future studies on risk stratification, as well as on whether treating to target influences the disease course significantly, are needed to further improve the long-term disease course in CD.


    We are grateful to Laimas Jonaitis (Lithuania), Sigita Gelman (Lithuania), Karen Kudsk (Denmark), Claus Aalykke (Denmark), Martina Giannotta (Italy), Tommaso Gabbiani (Italy), Elena Chernin (Israel), Laszlo Lakatos (Hungary), Vitalie Turcan (Moldova), Sylvie Lanier (France), Luciano Sanromán (Spain), Romina Fernandez-Poceir (Spain) and Anastasia N. Nicolaou (Cyprus) for their contribution to the study.


    View Abstract


    • Contributors All authors participated in the study design and data acquisition, have critically reviewed the draft manuscript for content and approved the final version for publication. JB had full access to the data in the study and takes full responsibility for its veracity and statistical analysis. JB and PM analysed and interpreted the data. JB drafted the manuscript. PM supervised the study.

    • Funding Unrestricted grant support has been received from Kirsten og Freddy Johansens Fond as well as from Nordsjællands Hospital Forskningsråd.

    • Disclaimer The study sponsors have made no contributions to the study design, analysis, data interpretation or publication.

    • Competing interests JB: consulting fees from Celgene, Janssen-Cilag, AbbVie A/S and Ferring and lecture fees from Abbvie A/S, Pfizer, MSD and Takeda Pharma A/S. VA: consultancy for Janssen and MSD. RS: consulting fees and/or lecture fees from AbbVie, MSD, Takeda, Janssen-Cilag. RD’I: consulting fees from Abbvie, Biocure and lecture fees from Takeda and Mundipharma. MF: speaker/lecture fees for Abbvie, Ferring, MSD, Takeda, Boehringer and Hospira. CG-R: lecture fees from Takeda, MSD, Ferring and Tillots. CE: lecture fees from Takeda. JH: research grants from Janssen, MSD and Takeda and lecture and/or consultancy fees from Abbvie, Cellgene, Ferring, Hospira, Janssen, Medivir, MSD, Pfizer, Vifor Pharma, Takeda and Tillotts Pharma. EL: lecture or consultancy fees from MSD, Abbvie and Ferring Pharmaceuticals. DD: lecture or consultancy fees from AbbVie, Takeda and Janssen. NA: lecture fees from MSD and Jansen. VH: personal fees, non-financial support and other from MSD, AbbVie, Ferring, Faes Farma, Shire, Falk Pharma, Tillots, Otsuka, Hospira Biologicals, Takeda, Jansen and Kernpharma Biologics. AF: personal fees and non-financial support from MSD, AbbVie, Shire and Tillots. SČ-Č: lecture fees from Takeda, MSD, Abbvie.

    • Ethics approval All centres obtained approval from their respective, local ethical committees.

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

    • Collaborators Ola Niewiadomski (Australia), Sally Bell (Australia), Niksa Turk (Croatia), Silvija Cukovic-Cavka (Croatia), Ioannis Kaimakliotis (Cyprus), Anastasia Nicolaou (Cyprus), Dana Duricova (Czech Republic), Milan Lukas (Czech Republic), Martin Bortlík (Czech Republic), Olga Shonová (Czech Republic), Johan Burisch (Denmark), Pia Munkholm (Denmark), Ebbe Langholz (Denmark), Birgitte Blichfeldt (Denmark), Dorte Marker (Denmark), Katrine Carlsen (Denmark), Petra Weimers (Denmark), Natalia Pedersen (Denmark), Jens Kjeldsen (Denmark), Clays Aalykke (Denmark), Jens Frederik Dahlerup (Denmark), Karen Kudsk (Denmark), Vibeke Andersen (Denmark), Ida Vind (Denmark), Niels Thorsgaard (Denmark), Riina Salupere (Estonia), Jóngerð Olsen (Faroe Islands), Kári Rubek Nielsen (Faroe Islands), Pia Oksanen (Finland), Pekka Collin (Finland), Konstantinnos H Katsanos (Greece), Dimitrios K Christodoulou (Greece), Alexandros Skamnelos (Greece), Dimitrios Politis (Greece), Karin Ladefoged (Greenland), Peter Laszlo Lakatos (Hungary), Zsuzsanna Vegh (Hungary), Laszlo Lakatos (Hungary), Peterne Demenyi (Hungary), Szabina Nemethne Kramli (Hungary), Colm O’Morain (Ireland), Giualia Dal Piaz (Italy), Alessia Santini (Italy), Giulia Girardin (Italy), Renata D’Inca (Italy), Doron Schwartz (Israel), Selwyn Odes (Israel), Limas Kupcinskas (Lithuania), Laimas Jonaitis (Lithuania), Gediminas Kiudelis (Lithuania), Irena Valantiene (Lithuania), Romanas Zykus (Lithuania).

    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.