https://orcid.org/0009-0006-5348-4645
Abstract
The emergence of biologic therapy has coincided with a decline in surgery rates for Crohn’s disease (CD). This study aims to describe the disease course, including intra-abdominal surgery rates, biologic therapy use, and variables associated with biologic therapy initiation in a cohort of newly diagnosed CD patients.
The Inflammatory Bowel Disease in South-Eastern Norway (IBSEN) III study is a population-based inception cohort study. From 2017 to 2019, newly diagnosed inflammatory bowel disease patients were included for prospective follow-up. The present study included CD patients ≥ 18 years. Clinical, endoscopic, and demographic data were collected at diagnosis and 1-year follow-up. Data were analyzed by using the Kaplan–Meier method and regression analyses.
In total, 424 CD patients (median age 37.0 years (range 18-80), female 55.0%) were included. At diagnosis, 50.5% presented with ileal disease and 80.7% with inflammatory behavior. Within a 1-year follow-up, 39.6% of patients received their first biologic therapy and 5.2% required intra-abdominal surgery. Systemic steroid treatment, CRP ≥ 5.0 mg dL−1, Harvey–Bradshaw Index score > 4, ileocolonic disease and penetrating disease behavior at diagnosis were independently associated with increased risk of initiation of biologic therapy, while age > 40 years was associated with decreased risk.
A high proportion of patients had ileal disease and inflammatory behavior at diagnosis. Still, nearly 40% started biologic therapy within the 1-year follow-up, while only 5% required intra-abdominal surgery.
Lay Summary
In this new, population-based inception cohort (Inflammatory Bowel Disease in South-Eastern Norway III), two-fifths of incident Crohn’s disease patients started biologic therapy within 1-year follow-up, despite predominantly mild disease with ileal location and inflammatory behavior at diagnosis. Only 5.2% required intra-abdominal surgery.
The early nineties IBSEN (Inflammatory Bowel Disease in South-Eastern Norway) cohort provided valuable data on disease course of inflammatory bowel disease (IBD) patients in the pre-biologic era, and was considered the gold standard for IBD population-based cohorts for many years.
In this population-based inception cohort, including Crohn’s disease (CD) patients treated by current biologic eras treatment algorithms, a high proportion of patients had ileal location and inflammatory behavior at diagnosis as compared with the first IBSEN study. Nearly 40% started biologics within 1-year of follow-up, while 5.2% required intra-abdominal surgery.
These real-world data may inform current discussions about early treatments strategies and their potential effects on outcomes in IBD.
Introduction
Crohn’s disease (CD) is a chronic inflammatory bowel disease (IBD) that can affect the entire gastrointestinal tract with a predilection for the terminal ileum and colon. Crohn’s disease often debuts at a young age and is characterized with a chronic but unpredictable disease course with flares ranging from mild to life-threatening.
Historically, although the majority of patients present with an inflammatory phenotype at diagnosis, up to 50% develop complications such as strictures, fistulas, and abscesses, often demanding surgical intervention.1 Efforts to prevent these complications have evolved notably in recent years, focusing on early and potent therapeutic approaches aimed at achieving mucosal healing.2–4 Biologic therapies have shown short-term efficacy in selected subgroups of patients with moderate to severe disease.2,5–8 However, uncertainties persist regarding their long-term impact on reducing bowel damage, surgeries, and hospitalizations.1,3
Population-based inception cohort studies are considered the gold standard for elucidating the disease course of CD, minimizing bias, and evaluating the real-life and long-term effectiveness of current treatment paradigms.9 During the pre-biologic era in the early 1990s, the Inflammatory Bowel Disease in South-Eastern Norway (IBSEN) study provided valuable insights into the outcomes of CD patients at that time, reporting a 15.9% intra-abdominal surgery rate within the 1-year follow-up of the CD diagnosis.10 Building upon this foundation, the IBSEN III study serves as a biologic era sequel, enrolling a new unselected population-based cohort that prospectively included newly diagnosed IBD patients in an overlapping geographic area.11 The present study aims to describe disease course, surgical and medical treatment during the first year following a CD diagnosis in the IBSEN III cohort, and selected variables associated with biologic therapy initiation.
Materials and Methods
Study Design and Cohort Description
The IBSEN III study is a multicentre, population-based observational inception cohort study that prospectively follows unselected IBD patients, diagnosed between January 1, 2017 and December 31, 2019 (Clinical Trials ID: NCT02727959). Patients were included from Norway’s most populous health region (the South-Eastern Health Region), which encompasses 20 hospitals and accounted for approximately 56% (2.9 million) of the Norwegian population in 2017.11 The cohort design details have been described previously by Kristensen et al.11 The CD diagnosis was reevaluated at the 1-year follow-up, based on all available information, and reclassified accordingly. Crohn’s disease patients ≥ 18 years with fulfilled Lennard-Jones Criteria at baseline and 1-year follow-up were included in the present study. Patients were treated according to guidelines at their local hospital and their physician’s discretion throughout the study.
Data Collection
Clinical, endoscopic, and demographic data were systematically collected in accordance with the study’s standard operating procedure.11 Additional data were obtained from patient records and electronic patient questionnaires as needed. Study investigators verified registered data accuracy and completeness through cross-referencing with local hospital records.
In patients with suspected IBD, the diagnostic process included endoscopy and ruling out current intestinal bacterial infection by testing stool samples for pathogens including Clostridium difficile. Additional evaluations, such as magnetic resonance imaging (MRI) or computerized tomography (CT) scans of the small intestine, upper endoscopy, or capsule endoscopy, were performed based on the clinician’s discretion. Blood samples were drawn at local laboratories, and those obtained within ±1 month from diagnosis were included in our baseline data. Fecal calprotectin (FC) was analyzed centrally by an enzyme-linked immunoassay test (the Bühlmann Calprotectin ELISA EK-CAL, Bühlmann Laboratories AG, Switzerland).
The 1-year follow-up, originally planned within 12 ± 3 months after the date of diagnosis, was extended by up to 12 months due to the COVID-19 pandemic. Thus, all included patients in our study cohort underwent their 1-year follow-up between 9- and 24-months post-diagnosis. Fecal calprotectin, blood samples, and endoscopic examination obtained within ±3 months from the 1-year follow-up were included in the analyses. The electronic clinical research form (eCRF) system Viedoc© (PCG Solutions AB, St Persgatan 6, 753 20 Uppsala, Sweden) was used for the registration of all data.11
Classifications and Definitions
The date of diagnosis was defined as the date of the index endoscopy, except for 3 patients where the date of diagnosis was set as the date of primary surgery, preceding endoscopy, or MRI. In addition, 1 patient declined endoscopy and had the date of diagnosis determined based on the date of CT findings.
Disease location and behavior were classified according to the Montreal classification (Location: L1 = ileal, L2 = colonic, L3 = ileocolonic, L4 = upper GI (isolated or in combination with L1-L3). Behavior: B1 = inflammatory disease (nonstricturing, non-penetrating), B2 = stricturing, B3 = penetrating disease, B4 = perianal disease (isolated or in combination with B1-B3)). Progression of disease location was defined as a change from L1 to L2 to L3 or adding L4 to either of these. Progression of disease behavior was defined as change from B1 to B2 to B3 or adding B4 to either of these. If no endoscopic or radiologic procedures were performed between diagnosis and 1-year follow-up, the Montreal classification from baseline was maintained.
The Harvey–Bradshaw Index (HBI)12 was used to measure clinical disease activity, with an HBI score ≤ 4 defined as clinical remission. Biochemical remission was defined as FC ≤ 250 µg g–1 13 and CRP ≤ 5 mg L–1. The most affected segment given by endoscopic examination was scored by using a modified version of the Simple Endoscopic Score for Crohn’s Disease (SES-CD), and a score of 0 defined endoscopic remission. Steroid-free remission was defined as clinical, biochemical, and endoscopic remission without the use of systemic steroid treatment at the 1-year follow-up. Patient-evaluated disease course according to predefined visual curves from the first IBSEN study,14 was registered at the 1-year follow-up.
Inflammatory bowel disease medication during the 1-year follow-up was registered and categorized according to ascending therapeutic potency: topical treatment (topical glucocorticoid, topical 5-aminosalisylates (5-ASA) and/or topical budesonide), oral 5-ASA, oral budesonide, systemic steroid (oral and/or intravenous glucocorticoid), immunomodulators (azathioprine and/or methotrexate) and biologic therapy (infliximab (IFX), adalimumab (ADA), golimumab (GOL), vedolizumab (VDZ), and/or ustekinumab (UST)). The first biologic therapy was defined as the first biologic drug administered to a biologic-naïve patient. The decision to switch biologic therapy was categorized based on the patient’s response to the first biologic therapy, including side effects, antibody formation, or non-response without antibody formation. Crohn’s disease-related surgery was registered at the 1-year follow-up and classified as intra-abdominal surgery (ileocoecal resection, small bowel stricturoplasty, small bowel resection, or segmental colonic resection) or perianal surgery (drainage of perianal abscesses or fistulae surgery).
Anemia was defined in accordance with the World Health Organisation (WHO) as hemoglobin (Hb) levels are <12.0 g dL–1 in women and <13.0 g dL–1 in men.
Statistical Analyses
Categorical variables are presented as counts and percentages. Continuous variables are described with median and range because of skewed distribution. Group comparisons were performed using Pearson chi-square tests or Fisher’s Exact tests for categorical data and Mann–Whitney U-tests for nonnormally distributed continuous data. Missing data was omitted from the analyses when not stated otherwise.
The Kaplan–Meier method was used to estimate the cumulative incidences of first biologic therapy and first CD-related intra-abdominal surgery. The event was defined as the start of first biologic therapy or first intra-abdominal surgery, respectively, within a 12-month period following the date of diagnosis. Time to event was calculated from the date of diagnosis to the date of first biologic therapy or the date of 1-year follow-up, whichever came first. Similarly, the time to event for the first CD-related intra-abdominal surgery was calculated from the date of diagnosis to the date of first intra-abdominal surgery or the date of 1-year follow-up, whichever came first. The crude cumulative incidences of first biologic therapy were compared among the Montreal disease locations (L1-L3) using the log-rank test. The Cox proportional hazards model was used to examine possible associations between the first biologic therapy and selected covariates assessed at diagnosis, including age, sex, smoking status, university hospital affiliation, duration of symptoms before diagnosis (in months), anemia, CRP, disease location and disease behavior, HBI, and systemic steroid treatment. Covariates with a P value < 0.1 in the univariate analyses were included in a multivariate Cox proportional hazard model. Point estimates are presented as hazard ratios with 95% confidence intervals (CI). The 95% CI for both univariate and multivariate regression coefficients were estimated using bootstrapping with 10 000 repetitions and the bias-corrected and accelerated methods. Statistical significance was defined as a P value < 0.05 for all analyses. Since the analyses were considered exploratory, corrections for multiple testing were not performed. SPSS version 29 (IBM SPSS Statistics, Chicago, IL, USA) or Stata version 16 (StataCorp LLC, College Station, TX, USA) were used to perform all analyses.
Ethical Considerations
The IBSEN III study was approved by the South-Eastern Regional Committee for Medical and Health Research Ethics (Ref 2015/946-3) and performed in accordance with the Declaration of Helsinki. All patients signed an informed consent form prior to inclusion and the data were stored and analyzed in the services for sensitive data (TSD) at the University of Oslo.
Results
Study Cohort
A total of 506 patients ≥ 18 years old with a newly confirmed CD diagnosis were included in the IBSEN III study. Of these, 36 patients (7.1%) were reclassified to another diagnosis during the 1-year follow-up, 29 (5.7%) were lost to follow-up, and 16 (3.2%) were not willing to participate. One patient above 80 years of age died before the 1-year follow-up and there was no information in the records that the death was IBD-related. Thus, data from 424 (83.8%) CD patients were available for the 1-year follow-up analyses (Figure 1). There were no statistically significant differences regarding relevant baseline characteristics between those who were unable or unwilling to attend (n = 46) and the remaining study cohort (n = 424) (Table 1). The median duration from diagnosis to the 1-year follow-up was 12 months (range 9-24 months).
Baseline characteristics of the study cohort with confirmed Crohn’s disease diagnosis at the 1-year follow-up, compared with the patients who were not able to participate at the 1-year follow-up.
| Study cohort [n = 424] | No 1-year follow-up [n = 46] | P-value | |
| Age, years (median [range]) | 37.0 [18.0-80.0] | 34.5 [18.0-81.0] | 0.83 |
| Gender, females (n [%]) | 233 [55.0] | 22 [47.8] | 0.72 |
| Time with symptoms in months, median [range] | 9.0 [0.0-480.0] | 12.0 [0.0-576.0] | 0.22 |
| Missing (n [%]) | 4 [0.9] | 1 [2.2] | |
| Smoking status (n [%]) | 0.27 | ||
| Current | 39 [11.7] | 5 [23.8] | |
| Previous | 107 [32.0] | 5 [23.8] | |
| Never | 188 [56.3] | 11 [52.4] | |
| Missing (n [%]) | 90 [21.2] | 25 [54.3] | |
| Disease location at diagnosisa (n [%]): | |||
| L1 [ileal] | 214 [50.5] | 27 [58.7] | 0.35 |
| L2 [colonic] | 74 [17.5] | 9 [19.6] | |
| L3 [ileocolonic] | 136 [32.1] | 10 [21.7] | |
| L4 [upper gastrointestinal tract] + [L1/L2/L3] | 11 [2.6] | 0 [0.0] | 0.61 |
| Disease behavior at diagnosisa (n [%]) | |||
| B1 [non stricturing, non-penetrating] | 342 [80.7] | 36 [78.3] | 0.36 |
| B2 [stricturing] | 70 [16.5] | 10 [21.7] | |
| B3 [penetrating] | 12 [2.8] | 0 [0.0] | |
| B4 [perianal disease] + [B1/B2/B3] | 16 [3.8] | 3 [6.5] | 0.42 |
| SES-CDb (median [range]) | 4.0 [0.0-12.0] | 4.0 [1.0-9.0] | 0.84 |
| Missing (n [%]) | 54 [12.7] | 10 [21.7] | |
| HBIc (median [range]) | 4.0 [0.0-25.0] | 3.0 [0.0-10.0] | 0.50 |
| EIMsd (n [%]) | |||
| None | 347 [86.8] | 36 [90.0] | |
| One | 41 [10.3] | 4 [10.0] | |
| Two | 9 [2.3] | 0 [0.0] | |
| Three | 3 [0.8] | 0 [0.0] | |
| General well-beinge (median [range]) | 1 [0-4] | 1 [0-3] | |
| Abdominal painf (median [range]) | 1 [0-3] | 1 [0-3] | |
| Number of liquid stoolsg (median [range]) | 2 [0-20] | 2 [0-6] | |
| Abdominal massh (median [range]) | 0 [0-3] | 0 [0-0] | |
| Missing (n [%]) | 24 [5.6] | 6 [13.0] | |
| Hbi (g dL−1) (median [range]) | 13.6 [6.7-18.1] | 14.1 [10.8-17.3] | 0.07 |
| Missing (n [%]) | 102 [24.1] | 17 [37.0] | |
| Ferritin (μg L−1) (median [range]) | 76.5 [5.0-1788.0] | 103.0 [16.0-438.0] | 0.32 |
| Missing (n [%]) | 122 [28.8] | 18 [39.1] | |
| CRPj (mg L−1) (median [range]) | 5.0 [0.0-280.0] | 5.0 [0.9-57.0] | 0.78 |
| Missing (n [%]) | 109 [25.7] | 18 [39.1] | |
| Systemic steroidsk (n [%]) | 91 [21.5] | 8 [17.4] | 0.52 |
| Study cohort [n = 424] | No 1-year follow-up [n = 46] | P-value | |
| Age, years (median [range]) | 37.0 [18.0-80.0] | 34.5 [18.0-81.0] | 0.83 |
| Gender, females (n [%]) | 233 [55.0] | 22 [47.8] | 0.72 |
| Time with symptoms in months, median [range] | 9.0 [0.0-480.0] | 12.0 [0.0-576.0] | 0.22 |
| Missing (n [%]) | 4 [0.9] | 1 [2.2] | |
| Smoking status (n [%]) | 0.27 | ||
| Current | 39 [11.7] | 5 [23.8] | |
| Previous | 107 [32.0] | 5 [23.8] | |
| Never | 188 [56.3] | 11 [52.4] | |
| Missing (n [%]) | 90 [21.2] | 25 [54.3] | |
| Disease location at diagnosisa (n [%]): | |||
| L1 [ileal] | 214 [50.5] | 27 [58.7] | 0.35 |
| L2 [colonic] | 74 [17.5] | 9 [19.6] | |
| L3 [ileocolonic] | 136 [32.1] | 10 [21.7] | |
| L4 [upper gastrointestinal tract] + [L1/L2/L3] | 11 [2.6] | 0 [0.0] | 0.61 |
| Disease behavior at diagnosisa (n [%]) | |||
| B1 [non stricturing, non-penetrating] | 342 [80.7] | 36 [78.3] | 0.36 |
| B2 [stricturing] | 70 [16.5] | 10 [21.7] | |
| B3 [penetrating] | 12 [2.8] | 0 [0.0] | |
| B4 [perianal disease] + [B1/B2/B3] | 16 [3.8] | 3 [6.5] | 0.42 |
| SES-CDb (median [range]) | 4.0 [0.0-12.0] | 4.0 [1.0-9.0] | 0.84 |
| Missing (n [%]) | 54 [12.7] | 10 [21.7] | |
| HBIc (median [range]) | 4.0 [0.0-25.0] | 3.0 [0.0-10.0] | 0.50 |
| EIMsd (n [%]) | |||
| None | 347 [86.8] | 36 [90.0] | |
| One | 41 [10.3] | 4 [10.0] | |
| Two | 9 [2.3] | 0 [0.0] | |
| Three | 3 [0.8] | 0 [0.0] | |
| General well-beinge (median [range]) | 1 [0-4] | 1 [0-3] | |
| Abdominal painf (median [range]) | 1 [0-3] | 1 [0-3] | |
| Number of liquid stoolsg (median [range]) | 2 [0-20] | 2 [0-6] | |
| Abdominal massh (median [range]) | 0 [0-3] | 0 [0-0] | |
| Missing (n [%]) | 24 [5.6] | 6 [13.0] | |
| Hbi (g dL−1) (median [range]) | 13.6 [6.7-18.1] | 14.1 [10.8-17.3] | 0.07 |
| Missing (n [%]) | 102 [24.1] | 17 [37.0] | |
| Ferritin (μg L−1) (median [range]) | 76.5 [5.0-1788.0] | 103.0 [16.0-438.0] | 0.32 |
| Missing (n [%]) | 122 [28.8] | 18 [39.1] | |
| CRPj (mg L−1) (median [range]) | 5.0 [0.0-280.0] | 5.0 [0.9-57.0] | 0.78 |
| Missing (n [%]) | 109 [25.7] | 18 [39.1] | |
| Systemic steroidsk (n [%]) | 91 [21.5] | 8 [17.4] | 0.52 |
aDisease location/disease behavior according to the Montreal classification.
b2SES-CD, Simple Endoscopic Score for Crohn’s Disease, modified, only scoring of the most affected bowel segment by endoscopic examination.
cHBI, Harvey–Bradshaw Index.
dExtraintestinal manifestations listed as none, one, two, or three of the following: arthralgia, uveitis, erythema nodosum, aphthous ulcers, pyoderma gangrenosum, fissures, fistula, and/or abscess.
eGeneral well-being: For the previous day, as reported by the patient. Very well = score 0, slightly below par = score 1, poor = score 2, very poor = score 3, terrible = score 4.
fAbdominal pain: For the previous day, as reported by the patient. None = score 0, mild = score 1, moderate = score 2, severe = score 3.
gNumber of liquid/soft stools: For the previous day, as reported by the patient.
hAbdominal mass: none = score 0, dubious = score 1, definite = score 2, definite and tender = score 3.
iHb, hemoglobin.
jCRP, C-reactive protein.
kSystemic steroids started at diagnosis of Crohn’s disease, including intravenous and oral steroid treatment.
Baseline characteristics of the study cohort with confirmed Crohn’s disease diagnosis at the 1-year follow-up, compared with the patients who were not able to participate at the 1-year follow-up.
| Study cohort [n = 424] | No 1-year follow-up [n = 46] | P-value | |
| Age, years (median [range]) | 37.0 [18.0-80.0] | 34.5 [18.0-81.0] | 0.83 |
| Gender, females (n [%]) | 233 [55.0] | 22 [47.8] | 0.72 |
| Time with symptoms in months, median [range] | 9.0 [0.0-480.0] | 12.0 [0.0-576.0] | 0.22 |
| Missing (n [%]) | 4 [0.9] | 1 [2.2] | |
| Smoking status (n [%]) | 0.27 | ||
| Current | 39 [11.7] | 5 [23.8] | |
| Previous | 107 [32.0] | 5 [23.8] | |
| Never | 188 [56.3] | 11 [52.4] | |
| Missing (n [%]) | 90 [21.2] | 25 [54.3] | |
| Disease location at diagnosisa (n [%]): | |||
| L1 [ileal] | 214 [50.5] | 27 [58.7] | 0.35 |
| L2 [colonic] | 74 [17.5] | 9 [19.6] | |
| L3 [ileocolonic] | 136 [32.1] | 10 [21.7] | |
| L4 [upper gastrointestinal tract] + [L1/L2/L3] | 11 [2.6] | 0 [0.0] | 0.61 |
| Disease behavior at diagnosisa (n [%]) | |||
| B1 [non stricturing, non-penetrating] | 342 [80.7] | 36 [78.3] | 0.36 |
| B2 [stricturing] | 70 [16.5] | 10 [21.7] | |
| B3 [penetrating] | 12 [2.8] | 0 [0.0] | |
| B4 [perianal disease] + [B1/B2/B3] | 16 [3.8] | 3 [6.5] | 0.42 |
| SES-CDb (median [range]) | 4.0 [0.0-12.0] | 4.0 [1.0-9.0] | 0.84 |
| Missing (n [%]) | 54 [12.7] | 10 [21.7] | |
| HBIc (median [range]) | 4.0 [0.0-25.0] | 3.0 [0.0-10.0] | 0.50 |
| EIMsd (n [%]) | |||
| None | 347 [86.8] | 36 [90.0] | |
| One | 41 [10.3] | 4 [10.0] | |
| Two | 9 [2.3] | 0 [0.0] | |
| Three | 3 [0.8] | 0 [0.0] | |
| General well-beinge (median [range]) | 1 [0-4] | 1 [0-3] | |
| Abdominal painf (median [range]) | 1 [0-3] | 1 [0-3] | |
| Number of liquid stoolsg (median [range]) | 2 [0-20] | 2 [0-6] | |
| Abdominal massh (median [range]) | 0 [0-3] | 0 [0-0] | |
| Missing (n [%]) | 24 [5.6] | 6 [13.0] | |
| Hbi (g dL−1) (median [range]) | 13.6 [6.7-18.1] | 14.1 [10.8-17.3] | 0.07 |
| Missing (n [%]) | 102 [24.1] | 17 [37.0] | |
| Ferritin (μg L−1) (median [range]) | 76.5 [5.0-1788.0] | 103.0 [16.0-438.0] | 0.32 |
| Missing (n [%]) | 122 [28.8] | 18 [39.1] | |
| CRPj (mg L−1) (median [range]) | 5.0 [0.0-280.0] | 5.0 [0.9-57.0] | 0.78 |
| Missing (n [%]) | 109 [25.7] | 18 [39.1] | |
| Systemic steroidsk (n [%]) | 91 [21.5] | 8 [17.4] | 0.52 |
| Study cohort [n = 424] | No 1-year follow-up [n = 46] | P-value | |
| Age, years (median [range]) | 37.0 [18.0-80.0] | 34.5 [18.0-81.0] | 0.83 |
| Gender, females (n [%]) | 233 [55.0] | 22 [47.8] | 0.72 |
| Time with symptoms in months, median [range] | 9.0 [0.0-480.0] | 12.0 [0.0-576.0] | 0.22 |
| Missing (n [%]) | 4 [0.9] | 1 [2.2] | |
| Smoking status (n [%]) | 0.27 | ||
| Current | 39 [11.7] | 5 [23.8] | |
| Previous | 107 [32.0] | 5 [23.8] | |
| Never | 188 [56.3] | 11 [52.4] | |
| Missing (n [%]) | 90 [21.2] | 25 [54.3] | |
| Disease location at diagnosisa (n [%]): | |||
| L1 [ileal] | 214 [50.5] | 27 [58.7] | 0.35 |
| L2 [colonic] | 74 [17.5] | 9 [19.6] | |
| L3 [ileocolonic] | 136 [32.1] | 10 [21.7] | |
| L4 [upper gastrointestinal tract] + [L1/L2/L3] | 11 [2.6] | 0 [0.0] | 0.61 |
| Disease behavior at diagnosisa (n [%]) | |||
| B1 [non stricturing, non-penetrating] | 342 [80.7] | 36 [78.3] | 0.36 |
| B2 [stricturing] | 70 [16.5] | 10 [21.7] | |
| B3 [penetrating] | 12 [2.8] | 0 [0.0] | |
| B4 [perianal disease] + [B1/B2/B3] | 16 [3.8] | 3 [6.5] | 0.42 |
| SES-CDb (median [range]) | 4.0 [0.0-12.0] | 4.0 [1.0-9.0] | 0.84 |
| Missing (n [%]) | 54 [12.7] | 10 [21.7] | |
| HBIc (median [range]) | 4.0 [0.0-25.0] | 3.0 [0.0-10.0] | 0.50 |
| EIMsd (n [%]) | |||
| None | 347 [86.8] | 36 [90.0] | |
| One | 41 [10.3] | 4 [10.0] | |
| Two | 9 [2.3] | 0 [0.0] | |
| Three | 3 [0.8] | 0 [0.0] | |
| General well-beinge (median [range]) | 1 [0-4] | 1 [0-3] | |
| Abdominal painf (median [range]) | 1 [0-3] | 1 [0-3] | |
| Number of liquid stoolsg (median [range]) | 2 [0-20] | 2 [0-6] | |
| Abdominal massh (median [range]) | 0 [0-3] | 0 [0-0] | |
| Missing (n [%]) | 24 [5.6] | 6 [13.0] | |
| Hbi (g dL−1) (median [range]) | 13.6 [6.7-18.1] | 14.1 [10.8-17.3] | 0.07 |
| Missing (n [%]) | 102 [24.1] | 17 [37.0] | |
| Ferritin (μg L−1) (median [range]) | 76.5 [5.0-1788.0] | 103.0 [16.0-438.0] | 0.32 |
| Missing (n [%]) | 122 [28.8] | 18 [39.1] | |
| CRPj (mg L−1) (median [range]) | 5.0 [0.0-280.0] | 5.0 [0.9-57.0] | 0.78 |
| Missing (n [%]) | 109 [25.7] | 18 [39.1] | |
| Systemic steroidsk (n [%]) | 91 [21.5] | 8 [17.4] | 0.52 |
aDisease location/disease behavior according to the Montreal classification.
b2SES-CD, Simple Endoscopic Score for Crohn’s Disease, modified, only scoring of the most affected bowel segment by endoscopic examination.
cHBI, Harvey–Bradshaw Index.
dExtraintestinal manifestations listed as none, one, two, or three of the following: arthralgia, uveitis, erythema nodosum, aphthous ulcers, pyoderma gangrenosum, fissures, fistula, and/or abscess.
eGeneral well-being: For the previous day, as reported by the patient. Very well = score 0, slightly below par = score 1, poor = score 2, very poor = score 3, terrible = score 4.
fAbdominal pain: For the previous day, as reported by the patient. None = score 0, mild = score 1, moderate = score 2, severe = score 3.
gNumber of liquid/soft stools: For the previous day, as reported by the patient.
hAbdominal mass: none = score 0, dubious = score 1, definite = score 2, definite and tender = score 3.
iHb, hemoglobin.
jCRP, C-reactive protein.
kSystemic steroids started at diagnosis of Crohn’s disease, including intravenous and oral steroid treatment.
Flowchart of the Crohn’s disease patients included in the Inflammatory Bowel Disease in South-Eastern Norway III study.
Disease Location and Behavior (Montreal Classification)
At diagnosis, about half of the patients (n = 214) presented with ileal disease location and 342 (80.7%) had inflammatory behavior (Table 1). Colonoscopy/CT/MRI or capsule endoscopy was performed in 314 (74.1%) of the patients before their 1-year follow-up appointment. For these 314 patients, the evolution of disease location and behavior from diagnosis to the 1-year follow-up is depicted in Figure 2A and 2B respectively. At the 1-year follow-up, disease location had progressed in 28 patients (8.9%), while 20 patients (6.4%) exhibited a change in disease behavior.
Sankey diagrams showing the development of disease location (A) and behavior (B) between diagnosis and 1-year follow-up for the patients in the cohort with CT/MRI/capsule endoscopy or colonoscopy (n = 314), following the Montreal classification for disease location (L1: Ileal, L2: Colonic, L3: Ileocolonic, L4: Upper GI) and disease behavior (B1: Non-stricturing, non-penetrating, B2: Stricturing, B3: Penetrating, B4: Perianal).
Treatment
Within the 1-year follow-up, 91.7% of the patients had received medical treatment for their CD. A total of 26 (6.1%) patients had received topical treatment, 69 (16.3%) had received 5-ASA, 220 (51.9%) had received budesonide, 151 (35.6%) had received systemic steroids, 175 (41.3%) had received immunomodulators, and 168 (39.6%) had received biologic therapy. Each patient was categorized according to the patient’s highest level of medical treatment in ascending order from bottom left to top right in Figure 3. An overview of the highest level of medical treatment given within 1-year follow-up in relation to the Montreal disease location and behavior at diagnosis is shown in Tables S1A and S1B.
Overview of the patient’s highest level of medical treatment within 1-year follow-up. The medical treatment is categorized in ascending order from bottom left to top right. Each step in the figure shows the percentage of patients whose highest level of treatment falls within that category. 1Topical treatment includes 5-ASA, steroids, and/or budesonide. 2Immunomodulators include azathioprine and methotrexate. 3First biologic therapy to biologic-naïve patients.
Biologic Therapy
Of the 168 (39.6%, 95% CI [34.9-44.5]) patients that received their first biologic therapy within the 1-year follow-up, 5 patients used biologics for non-IBD indications at diagnosis and were therefore excluded from further analyses regarding biologic therapy. The median time from diagnosis to start of first biologic therapy was 3 months (range 0-21). The cumulative incidence of the start of first biologic therapy is depicted using the Kaplan–Meier curve in Figure 4A. The crude cumulative incidence of first biologic therapy was significantly lower for L1 compared with L3 (Figure S1A) and for B1 compared with B3 (Figure S1B).
The Kaplan–Meier estimate of time to first biologic therapy (A) and first intra-abdominal surgery (B) in months. Five patients were excluded from the analysis of biologic therapy since they were not biologic-naïve and used a non-IBD related biologic therapy. One-hundred-and-sixty-four patients started their first biologic therapy up to and including 12 months after diagnosis. Twenty-one patients had their first intra-abdominal Crohn’s disease-related surgery up to and including 12 months after diagnosis. Three of these patients were diagnosed during their surgery (months from diagnosis = 0). The numbers in the parenthesis below the x axis are the number of events (first biologic therapy/intra-abdominal surgery) between the month tick marks on the x axis.
Overall, 162 (96.4%) patients received anti-tumor necrosis factor (TNFi) as their first biologic (n = 101 IFX and n = 61 ADA) therapy, whereas 6 (3.6%) patients received VDZ as their first biologic therapy. Fifteen (9.2%) patients changed (11 from IFX and 4 from ADA) to a second biologic therapy, and one of these patients changed to UST as a third biologic therapy. For details see Table S2.
The univariate and multivariate regression analyses are presented in Table S3. In the multivariate analysis, treatment with systemic steroids, CRP ≥ 5.0 mg dL–1, HBI score > 4, ileocolonic disease location, and penetrating disease behavior at diagnosis remained independently associated with an increased risk of starting first biologic therapy. Age above 40 years was still associated with a reduced risk.
Surgery
Seven patients (1.7%, 95% CI [0.7-3.4]) underwent perianal surgery and 22 (5.2%, 95% CI [3.3-7.8]) patients underwent intra-abdominal surgery within the 1-year follow-up. Of these 22, 13 underwent ileocoecal resection, 1 underwent small bowel stricturoplasty, 7 underwent small bowel resection and 1 patient underwent a segmental colonic resection. Three patients were diagnosed with CD during their primary surgery. The cumulative incidence of the first intra-abdominal surgery using the Kaplan-Meier method, is depicted in Figure 4B. Of the 22 patients that underwent intra-abdominal surgery, 16 had L1, and 6 had L3 disease location at diagnosis; 5 started their first biologic therapy prior to their intra-abdominal surgery and 2 started after, while the remaining 15 did not receive any biologic therapy within their 1-year follow-up.
Disease Activity at the 1-Year Follow-Up
Disease activity at the 1-year follow-up is shown in Figure 5. At the 1-year follow-up, 371 patients (87.5%) evaluated their own disease course, using predefined curves describing their subjective assessment of disease activity over time (Figure 6). Almost two thirds (64.4%) selected Curve 1, with initial high activity followed by remission or mild severity of intestinal symptoms.
Overview of clinical (HBI), biochemical (CRP and FC), and endoscopic (SES-CD) remission status. Remission was defined using the following criteria: clinical remission (HBI ≤ 4), biochemical remission (CRP < 5 mg L−1 and FC ≤ 250 µg g−1), and endoscopic remission (SES-CD = 0). Non-remission was defined as failure to meet these remission criteria. Remission was further classified as either steroid-free remission or remission on steroids, based on the absence or presence of steroid treatment at the 1-year follow-up, respectively. HBI, Harvey–Bradshaw Index; CRP, C-reactive protein; FC, fecal calprotectin; SES-CD, Simple Endoscopic Score for Crohn’s disease.
Patient-reported evaluation of first-year disease course based on predefined curves adapted from the Inflammatory Bowel Disease in South-Eastern Norway study. Curves presented as answer options to the question “What curve best describes the progression of your disease” at the 1-year follow-up. Curve 1: initially high disease activity followed by remission or mild disease. Curve 2: initially low disease activity followed by worsening of intestinal symptoms. Curve 3: chronic continuous intestinal symptoms. Curve 4: chronic intermittent symptoms.
Discussion
This large, population-based CD inception cohort is characterized by mild disease with a predominance of ileal location (L1) and inflammatory behavior (B1) at time of diagnosis. Still, almost 2 out of 5 received biologic therapy within their 1-year follow-up, primarily with anti-TNF drugs, while only 5.2% underwent intra-abdominal surgery. Steroid use, elevated CRP, and severe disease (L3 or B3) at time of diagnosis were associated with an increased risk of starting biologic therapy, whereas patients aged over 40 years at diagnosis were less likely to start such treatment. Within 1-year follow-up, there was an 8.9% shift in disease location and a 6.4% change in disease behavior. At the 1-year follow-up, almost two thirds (64.4%) of patients self-reported remission or mild severity of intestinal symptoms, while 80.2% fulfilled the criteria for clinical steroid-free remission (HBI ≤ 4).
At diagnosis, half of the study cohort exhibited terminal ileitis (disease location L1) and 80.7% displayed inflammatory disease behavior (B1). This distribution aligns with a large Spanish inception study from 2017.15 However, with the exception of a small Israeli study,16 the proportion of patients with L1 location in our study exceeded what has been observed in several other comparable studies,17–23 which might be explained by multiple factors. High awareness of IBD, widespread use of FC among Norwegian general practitioners and outpatient clinics, along with easy access to endoscopy and MRI in publicly funded hospitals may have led to earlier diagnosis, and thus potentially more limited location and milder behavioral phenotype at time of diagnosis. However, the median patient-reported duration time of symptoms before diagnosis was 9 months in our study, surpassing the 6 months reported in the first IBSEN study.23 This could be linked to milder symptoms in our study’s predominantly ileal phenotype (50.5%), as opposed to the predominantly colonic phenotype (48.5%) in the first IBSEN study.14,24 These results align with the Swiss IBD Cohort Study, which reported a median duration of 9 months from initial symptoms to diagnosis, with more pronounced delays among patients > 40 years old and those with L1 disease at diagnosis, possibly due to milder symptoms and similarity with irritable bowel syndrome.25
There is a potential of misclassifying CD with limited extension (L1), for example, with NSAIDs-associated ulcerations or infection. The impact of the latter is expected to be limited by the exclusion of patients with positive stool pathogen tests. Use of NSAIDs before diagnosis was associated with L1 phenotype and change of diagnosis to non-IBD or lost to follow up before the 1-year follow up. Therefore, we cannot exclude some misclassification of NSAIDS-related ileal inflammation with mild L1 disease in the cohort at baseline. However, the distribution of CD versus ulcerative colitis in IBSEN III closely aligns with the first IBSEN study from the early nineties, indicating that it is unlikely that a large proportion of patients with L1 disease were misclassified as having CD.10,11 Also, it has previously been reported a tendency among clinicians to misclassify ileocecal involvement as L3 rather than L1,26 which could have resulted in an underestimation of L1 and overestimation of L3 in studies comparable to ours. Moreover, the high rates of uncomplicated disease might arise from changes in disease phenotype per se. A retrospective study in Sweden from 1963 to 2010 observed a significant increase in patients diagnosed with B1 disease (12.5% to 82.3%) over this period,27 suggesting improved early detection methods or substantial phenotypic changes.
In our study cohort, disease location and behavior progressed by 8.9% and 6.4% respectively (Figure 2A and 2B). In contrast, the EpiCom cohort from 2010 reported stable disease phenotype within 1-year follow-up.19 In a Danish cohort (2003-2004), changes in disease location and progression of disease behavior significantly increased the risk of surgical resection, while biologic therapy appeared to prevent changing of disease location, thus having the potential to improve patient outcomes.9 Previous studies have identified disease location as the strongest predictor of subsequent intestinal complications, with ileal and upper GI involvement significantly associated with stricturing or penetrating disease progression.28 Since our cohort constitutes 82.4% of patients within these risk categories at diagnosis (L1, L3, and L4), it will be crucial to investigate the impact of medical treatment and further risk of complications and surgery.
Nearly 40% of the study cohort received their first biologic therapy within the 1-year follow-up, exceeding rates reported in comparable European and Australian cohorts (ranging between 6% and 25%).15,17,19 This high utilization of biologic therapy can be attributed to several factors. Access to biologic therapies has been found to be strongly linked to the Gross Domestic Product in different countries.29 Norway has a well-funded public healthcare system, and the Norwegian Hospital Procurement Trust (NHPT) rapidly incorporates cost-effective biosimilars into its Pharmaceutical Strategy via an annual competitive bidding process, promoting affordability and accessibility.30 A nationwide Norwegian registry study observed a significant increase in the initiation of biologic therapy within 12 months post diagnosis, from 17% in 2010 to 33% in the 2016 cohort (P < .001).30 Particularly, the adoption of IFX as the first biologic surged after the approval of its biosimilar in 2013.30 Similarly, over 95% of our study cohort received IFX or ADA as first biologic therapy, following NHPT’s recommendation in the study period. In addition, the relatively small Norwegian medical IBD community enables swift adoption of novel treatments, as exemplified by early and substantial research endorsing the shift from the original TNF inhibitor to its biosimilar.31
In this study, we found that CRP ≥ 5.0 mg dL−1, HBI score > 4, L3, B3, age < 40 years, and systemic steroids at diagnosis were all independently associated with increased biologic therapy utilization. The crude cumulative incidence of first biologic therapy was significantly lower for L1 compared with L3 and for B1 compared with B3, aligning with results from the Western Hungary17 and the 2011 EpiCom cohort study.22 These outcomes underscore the utilization of biologic therapy for patients with risk factors for more severe disease in our study.
In the 1-year follow-up of the EpiCom cohort from 2010, immunomodulators were used at the highest treatment level in 29% of Western and 31% of Eastern European countries.19 In our study this percentage was lower (12.7%), but 28.3% received immunomodulators and TNFi, which is common Norwegian hospital practice for the initial 6-12 months. Despite limited evidence for its efficacy,32 16.3% of our study cohort received 5-ASA during the first year after diagnosis, while only 3.1% received it as their highest treatment step. These percentages align with Western Europe’s rates in the EpiCom cohort from 2010,19 but are lower than those observed in Eastern Europe in the same study19 and in the 2017 Spanish study.15 At our 1-year follow-up, 80.2% attained clinical steroid-free remission, aligning with the 77% 1-year remission rate (though not defined as steroid-free remission) in the EpiCom cohort from 2010.19 This is, however, somewhat higher than in a recent Israeli study of 76 newly diagnosed CD patients, of which 64.5% achieved clinical steroid-free remission, despite a higher utilization of biologics (60.5% within the 1-year follow up).16
We report a lower rate of intra-abdominal surgery compared with corresponding studies, with 5.2% of the CD patients undergoing intra-abdominal surgery within their 1-year follow-up.10,15,17,21,22,33 Within the first year after diagnosis, the 2011 EpiCom cohort noted 24% in Eastern Europe and 7% in Western Europe/Australia undergoing surgery.22 Our surgery rates align better with those reported in Western Europe/Australia (7%), where a higher percentage received biologic therapy (19%) compared with Eastern Europe (9%).22 Due to limited surgery cases in our study, however, statistical power was insufficient for univariate and multivariate Cox regression analyses to calculate intra-abdominal surgery risk factors.
Improved access to specialist care, which has been associated with decreased surgery rates in a Canadian study,34 and declining smoking rates over the last decades,35 may have contributed to the reduced surgery rates in our cohort, compared with the first IBSEN study (15.9%).10 The first IBSEN study reported 40% smokers and 16.0% ex-smokers at diagnosis,10 whereas the present study observed 11.7% smokers and 32.0% ex-smokers. However, the first IBSEN study found no elevated surgery rates among former and current smokers compared with nonsmokers,10 aligning with the EpiCom cohort from 2010,36 possibly due to their population-based study design. A recent meta-analysis revealed a more complex disease course (higher odds of first and second surgery and relapse) among smokers than nonsmokers, but most studies were conducted in referral settings.37
A nationwide Norwegian registry-based study from 2010 to 2012 found that the South-Eastern health region had the highest TNFi use and lowest cumulative surgery incidence,38 findings that are consistent with a systematic review conducted up to May 2012.39 However, a Spanish study noted no significant difference in the risk of surgery between well-equipped hospitals with higher biologic use and those with lower biologic utilization and fewer resources within 1-year follow-up.15 Similarly, the EpiCom cohort from 2010 noted higher biologic use in Western (21%) than Eastern Europe (6%), yet comparable surgery rates (16% and 12%, respectively).19 In a Danish cohort (1979-2011), reduced 5-ASA and steroid use correlated with fewer surgeries alongside the rise in use of thiopurines and TNFis, but the effect of the latter in reducing surgeries was inconclusive.40 The PROFILE study, found that top-down therapy using IFX and an immunomodulator significantly surpassed accelerated step-up treatment in preventing surgery and achieving remission at one year in newly diagnosed patients with active CD. These findings support adopting top-down treatment for these patients.8
The main strength of this study lies in its observational inception cohort design, with the prospective follow-up of incident, unselected CD patients within a well-defined geographical area. All Norwegian citizens have access to a public healthcare system providing an optimal setting for conducting such studies.11 In addition, CD diagnoses were reevaluated at the 1-year follow-up, enhancing the reliability of our findings regarding the type of diagnosis. Furthermore, the present study’s comparability with the first IBSEN study provides valuable longitudinal insights in IBD research and treatment over time.
Several limitations necessitate consideration. Our findings are based on a Norwegian healthcare system and may not be directly transferable to other healthcare systems. As a multicentre study, variations in patient inclusion and follow-up among participating hospitals may have introduced data collection bias. However, there was a substantial participation across multiple centers, and differences in proportions of included patients were primarily attributable to variations in catchment area size. All data were closely monitored by study investigators and standardization of inclusion, follow-up and data recording ensured data quality, validity, and patient comparability. However, registration of specific indication for treatments (including intra-abdominal surgeries and biologic therapy) was not included in the protocol. The risk of receiving biologic therapy was similar for patients treated at university and non-university hospitals. Some 1-year follow-ups were postponed, canceled, or performed without endoscopy mainly due to the COVID-19 pandemic, which might have introduced selection bias. Nevertheless, the median 1-year follow-up duration remained 12 months and since there were no significant differences between the study cohort who underwent 1-year follow-up and those who did not, we consider the study cohort as representative.
In conclusion, the 1-year follow-up of this population-based cohort study highlights the predominant involvement of the terminal ileum and inflammatory disease behavior at diagnosis. Utilization of biologic therapy was high and independently associated with several clinical variables indicative of disease severity. The rate of intra-abdominal surgery was markedly decreased compared to the first IBSEN cohort study, from the pre-biologic era. We propose that increased biologic usage among patients with initial primarily uncomplicated disease phenotypes contributed to the notably low surgery rates in this cohort, suggesting potential benefits of early and aggressive medical intervention. Whether this reduction reflects a successful application of treatments during the “window of opportunity” and is associated with further reduced surgery rates over time, will be assessed in future follow-ups.
Acknowledgements
The authors would like to thank all members of the IBSEN III Study Group: The study nurses, local study personnel, and the local principal investigators (Arnt Sagosen, Audun Hasund, and Tor Tønnesen) who have contributed to the inclusion and follow-up of the patients. We also want to thank Unger-Vetlesen Institute for analyzing the fecal samples and all the patients who contributed by participating in this study. The graphical abstract includes images adapted from Servier Medical Art, licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/).
Author Contributions
Study conception and design: M.L.H., V.A.K., T.E.D., R.O., G.H.H., P.R., G.P., C.L., and V.S. Collection of data: M.L.H., V.A.K., T.E.D., G.H.H., P.R., G.P., C.L., V.S., M.M.B., R.B., S.O.F., A.W.M., M.H., Ø.H., I.J., B.C.O., J.P., R.T., C.M.Y., O.H., Ø.A., S.V., and T.B.A. Statistical analyses, data interpretation, and manuscript draft: C.L., M.L.H., V.A.K., V.S., and M.H. All authors critically reviewed the manuscript and have read and approved the final version of the manuscript for important intellectual content.
Funding
The IBSEN III study is fully investigator-initiated and has received funding from the following sources: Takeda Pharmaceuticals, Pfizer, Ferring Pharmaceuticals, Tillotts Pharma, Foundation Dam (grant number 2019/FO244676), and the South-Eastern Health Authorities. The funding played no role in the design of the study, collection, analysis, or interpretation of data, or in writing of this manuscript.
Conflict of Interest
V.S.: Sponsored by funds from Takeda. T.E.D.: Served as a speaker, consultant, or advisory board member for AbbVie, Ferring, Pfizer, Pharmacosmos, Tillotts, and Vifor Pharma. Received unrestricted research grants from AbbVie and Pharmacosmos. Ø.H.: Speaker and advisory boards: Takeda, Janssen-Cilag, Bristol-Myers Squibb, and AbbVie. A.W.M.: Investigator-initiated research grant from Takeda. M.L.H.: Investigator-initiated research grants from Takeda, Pfizer, Tilllotts, Ferring, and Janssen. Speaker honoraria from Takeda, Tillotts, Ferring, AbbVie, Galapagos, and Meda. Advisory board Takeda, Galapagos, MSD, Lilly, and AbbVie. V.A.K.: Speaker honoraria Thermo Fischer Scientific, Consultant Janssen-Cilag AS, Advisory Board Tillotts Pharma AG, and Takeda AS. S.O.F. reports personal fees from Takeda, Galapagos, Jansen-Cilag, AbbVie, Pharmacosmos, Norgine, and Bristol-Myers-Squibb. The remaining authors have no conflict of interest.
Data Availability
The data underlying this article cannot be shared publicly due to the privacy of individuals who participated in the study. The data will be shared on reasonable request to the corresponding author.
Conference Presentation
Preliminary data were presented as a poster at ECCO 2023 in Copenhagen.
