Advertisement

Disease course and treatment outcomes of Crohn's disease patients with early or late surgery – A Danish nationwide cohort study from 1997 to 2015

  • M.Z. Sarikaya
    Correspondence
    Corresponding author at: Kettegaard Alle 30, 2650 Hvidovre, Denmark.
    Affiliations
    Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark

    Gastro Unit-Medical Division, Hvidovre University Hospital, Hvidovre, Denmark

    Copenhagen Centre for Inflammatory Bowel Disease in Children, Adolescents and Adults, Hvidovre University Hospital, Hvidovre, Denmark
    Search for articles by this author
  • M. Zhao
    Affiliations
    Gastro Unit-Medical Division, Hvidovre University Hospital, Hvidovre, Denmark

    Copenhagen Centre for Inflammatory Bowel Disease in Children, Adolescents and Adults, Hvidovre University Hospital, Hvidovre, Denmark
    Search for articles by this author
  • B. Lo
    Affiliations
    Gastro Unit-Medical Division, Hvidovre University Hospital, Hvidovre, Denmark

    Copenhagen Centre for Inflammatory Bowel Disease in Children, Adolescents and Adults, Hvidovre University Hospital, Hvidovre, Denmark
    Search for articles by this author
  • F. Bendtsen
    Affiliations
    Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark

    Gastro Unit-Medical Division, Hvidovre University Hospital, Hvidovre, Denmark

    Copenhagen Centre for Inflammatory Bowel Disease in Children, Adolescents and Adults, Hvidovre University Hospital, Hvidovre, Denmark
    Search for articles by this author
  • J. Burisch
    Affiliations
    Gastro Unit-Medical Division, Hvidovre University Hospital, Hvidovre, Denmark

    Copenhagen Centre for Inflammatory Bowel Disease in Children, Adolescents and Adults, Hvidovre University Hospital, Hvidovre, Denmark
    Search for articles by this author
Open AccessPublished:October 17, 2022DOI:https://doi.org/10.1016/j.dld.2022.09.016

      Abstract

      Background

      Studies on early surgery among Crohn's disease patients are few and focus on ileocolonic resections.

      Aim

      The aim of this nationwide cohort study was to investigate the disease course in all Crohn's disease patients who underwent early and late major abdominal surgery.

      Methods

      In a Danish nationwide cohort of Crohn's disease patients from 1997 to 2015 we included 493 patients (group 1) resected within 29 days, 472 patients (group 2) resected between 30 and 180 days, and 1,518 patients (group 3) resected after 180 days of diagnosis. Re-operation, hospitalisations and medications were analysed.

      Results

      The cumulative risk of re-operation was lower among patients from group 1 (five-year risk: 16.5% vs. group 2: 18.2% and group 3: 21.2%, p = 0.004). Fewer patients from group 2 and 3 required hospitalisations (269 (56.5%) and 803 (52.8%) vs. group 1: 329 (66.8%) p<0.001). Patients from group 3 had a higher cumulative use of immunomodulators in the first three years after initial surgery (one-year risk: 24.6% vs. group 1: 19.4% and group 2: 17.0%, p<0.001).

      Conclusion

      Crohn's disease patients resected within 29 days of diagnosis had a lower cumulative risk of re-operation and a lower cumulative exposure to immunomodulators in the initial years after surgery.

      Keywords

      1. Introduction

      The prevalence of Crohn's disease (CD) is increasing worldwide and is expected to increase further [
      • Alatab S.
      • Sepanlou S.G.
      • Ikuta K.
      • et al.
      The global, regional, and national burden of inflammatory bowel disease in 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017.
      ]. In the past two decades, managing CD has improved following the development of novel treatments such as biological therapies and, more recently, small molecules. While the use of biological therapies and immunomodulators has increased significantly, data about their impact on the natural disease course of CD are conflicting as several studies have failed to demonstrate significant reductions in surgery rates over time [
      • Dittrich A.E.
      • Sutton R.T.
      • Haynes K.
      • et al.
      Incidence Rates for Surgery in Crohn's Disease Have Decreased: a Population-based Time-trend Analysis.
      ,
      • Mao E.J.
      • Hazlewood G.S.
      • Kaplan G.G.
      • et al.
      Systematic review with meta-analysis: comparative efficacy of immunosuppressants and biologics for reducing hospitalisation and surgery in Crohn's disease and ulcerative colitis.
      ,
      • Burisch J.
      • Kiudelis G.
      • Kupcinskas L.
      • et al.
      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.
      ,
      • Murthy S.K.
      • Begum J.
      • Benchimol E.I.
      • et al.
      Introduction of anti-TNF therapy has not yielded expected declines in hospitalisation and intestinal resection rates in inflammatory bowel diseases: a population-based interrupted time series study.
      ].
      As such, surgery remains an important treatment option for CD patients, especially for those who are refractory to medical therapies or experiencing complications such as ileus, strictures or penetrating disease that cannot be treated otherwise. Previous studies have found that between 70% and 80% of CD patients undergo surgery within 20 years of their diagnosis [
      • Cosnes J.
      • Gower–Rousseau C.
      • Seksik P.
      • et al.
      Epidemiology and Natural History of Inflammatory Bowel Diseases.
      ], but the rate of primary surgery has decreased significantly in recent years [
      • Kalman T.D.
      • Everhov Å.H.
      • Nordenvall C.
      • et al.
      Decrease in primary but not in secondary abdominal surgery for Crohn's disease: nationwide cohort study, 1990–2014.
      ]. Surgery is not curative and post-operative recurrence is common in CD. While post-surgical remission may last longer than drug-induced remission, data from the biological era are lacking [
      • Silverstein M.D.
      • Loftus E.V.
      • Sandborn W.J.
      • et al.
      Clinical course and costs of care for Crohn's disease: markov model analysis of a population-based cohort.
      ]. Furthermore, surgery has been shown to exert a significant positive effect on the quality of life and disability of CD patients [
      • Ha F.J.
      • Thong L.
      • Khalil H.
      Quality of life after intestinal resection in patients with crohn disease: a systematic review.
      ,
      • D'Ugo S.
      • Romano F.
      • Sibio S.
      • et al.
      Impact of surgery on quality of life in Crohn's disease: short- and mid-term follow-up.
      ,
      • Wright E.K.
      • Kamm M.A.
      • De Cruz P.
      • et al.
      Effect of Intestinal Resection on Quality of Life in Crohn's Disease.
      ,
      • Lo B.
      • Prosberg M.V .V..
      • Gluud L.L.L.
      • et al.
      Systematic review and meta-analysis: assessment of factors affecting disability in inflammatory bowel disease and the reliability of the inflammatory bowel disease disability index.
      ].
      Previous retrospective studies and a recent trial suggest that some patients with ileocolonic CD might benefit from early surgery, with fewer hospital admissions, a lighter burden of medical treatment, and a reduced risk of re-operation [
      • An V.
      • Cohen L.
      • Lawrence M.
      • et al.
      Early surgery in Crohn's disease a benefit in selected cases.
      ,
      • ARATARI A.
      • PAPI C.
      • LEANDRO G.
      • et al.
      Early versus late surgery for ileo-caecal Crohn's disease.
      ,
      • Stevens T.W.
      • Haasnoot M.L.
      • D'Haens G.R.
      • et al.
      Laparoscopic ileocaecal resection versus infliximab for terminal ileitis in Crohn's disease: retrospective long-term follow-up of the LIR!C trial.
      ,
      • Windsor A.C.J.
      Ileal Crohn's disease is best treated by surgery.
      ]. Therefore, the timing of surgical resection is an important question that influences the overall clinical outcome and disease course in CD patients. Moreover, existing studies of early surgery have mainly focused on ileocecal resected patients. These studies are not population-based, which is why focusing on early surgery in the general CD population is important [
      • ARATARI A.
      • PAPI C.
      • LEANDRO G.
      • et al.
      Early versus late surgery for ileo-caecal Crohn's disease.
      ,
      • Stevens T.W.
      • Haasnoot M.L.
      • D'Haens G.R.
      • et al.
      Laparoscopic ileocaecal resection versus infliximab for terminal ileitis in Crohn's disease: retrospective long-term follow-up of the LIR!C trial.
      ,
      • Windsor A.C.J.
      Ileal Crohn's disease is best treated by surgery.
      ].
      This study aimed to provide real-world evidence for the disease course of CD patients undergoing early and late surgery in a nationwide, population-based setting.

      2. Materials and methods

      2.1 Data sources

      This study used the Danish National Patient Registry and the Danish National Prescription Registry. All Danish citizens are given a ten-digit personal identification number at birth or upon their immigration [
      • Bøcker Pedersen C.
      • Gøtzsche H.
      • Ostrup Møller J.
      • et al.
      The Danish Civil Registration System. A cohort of eight million persons.
      ] that can be used to associate them with the national registries to obtain individual-level healthcare data. The National Patient Registry contains individualised data about inpatient contacts since 1977 and outpatient contacts since 1995, including diagnoses, date of diagnoses, date of contacta, diagnostic and surgical procedures performed and medications administered or prescribed through hospitals (such as biologics) [
      • Andersen T.F.
      • Madsen M.
      • Jørgensen J.
      • et al.
      The Danish National Hospital Register. A valuable source of data for modern health sciences.
      ]. The National Prescription Registry was established in 1995 and contains information about types of medication, dosage, packaging and date of issue for all medications that have been collected [
      • Wallach Kildemoes H.
      • Toft Sørensen H.
      • Hallas J
      The Danish national prescription registry.
      ].

      2.2 Study population

      From the Danish National Patient Registry, we identified all patients aged 18 years or older diagnosed with CD in Denmark between January 1, 1997 and December 31, 2015. Patients were identified using the CD diagnosis code K50.x, from the International Classification of Diseases 10th revision (ICD-10) [
      • Andersen T.F.
      • Madsen M.
      • Jørgensen J.
      • et al.
      The Danish National Hospital Register. A valuable source of data for modern health sciences.
      ]. This code has previously been validated to have 97% agreement [
      • Lo B.
      • Vind I.
      • Vester-Andersen M.K.
      • et al.
      Validation of ulcerative colitis and Crohn's disease and their phenotypes in the Danish National Patient Registry using a population-based cohort.
      ]. For patients to be included in this study, they were required to have at least two entries of a CD diagnosis on different dates, which in previous studies has been shown to have the highest validity [
      • Lo B.
      • Vind I.
      • Vester-Andersen M.K.
      • et al.
      Validation of ulcerative colitis and Crohn's disease and their phenotypes in the Danish National Patient Registry using a population-based cohort.
      ,
      • Lophaven S.N.
      • Lynge E.
      • Burisch J.
      The incidence of inflammatory bowel disease in Denmark 1980-2013: a nationwide cohort study.
      ,
      • Fonager K.
      • Sørensen H.T.
      • Rasmussen S.N.
      • et al.
      Assessment of the diagnoses of Crohn's disease and ulcerative colitis in a Danish hospital information system.
      ]. The date of diagnosis was defined as the first date on which a CD diagnosis had been registered. Patients registered with both CD and ulcerative colitis were classified as having inflammatory bowel disease unclassified (IBDU). If an IBDU patient had only CD diagnosis codes or CD surgery procedures registered in the preceding five years, they were included as CD patients. CD patients with a diagnosis of colorectal cancer within 30 days of their surgeries were not included in the study.
      From the cohort of incident CD patients, we identified those who underwent an intestinal resection due to their CD based on Nomesco Classification of Surgical Procedure codes (Supplementary Table 1). The population was divided into three groups based on days between date of diagnosis and initial surgery. Groups 1, 2 and 3 were defined as CD-related surgery within 29 days (or within 30 days before diagnosis), between 30 and 180 days, and more than 180 days after the date of diagnosis, respectively. Early surgery was defined as surgery performed within six months of diagnosis and therefore included both groups 1 and 2. To avoid the inclusion of early-resected patients who received IBD medication before their initial CD surgery, patients in group 1 treated with IBD medicine prior to their initial surgery were excluded, except for patients who received corticosteroids eight months or more prior, since corticosteroids might have been prescribed for diseases other than CD.

      2.3 Outcomes

      Re-operation, hospitalization and IBD medication were analysed for each subgroup of CD patients. We categorised IBD-related medications into four levels of ascending potency: (1) topical 5-aminosalicylates (5-ASA), oral 5-ASA, rectal steroids; (2) oral or intravenous corticosteroids: prednisolone, methylprednisolone and budesonide; (3) immunomodulators: azathioprine, methotrexate, 6-mercaptopurine; (4) biologic therapy. A course of systemic steroids was defined as a 14-day daily dose of 50 mg, tapering by 5 mg every week. Additionally, a course of budesonide was defined as a daily dose of 9 mg, tapering by 3 mg every fourth week. The definition of immunomodulator and biologic therapy use was having received at least one treatment.
      Surgery was categorised as (1) colon resections (including rectum), (2) small bowel resections, (3) ileocecal resections, and (4) simultaneous colon and small bowel resection. Major abdominal surgeries were identified and classified based on Nomesco Classification of Surgical Procedure codes. Hospitalisations were defined as inpatient hospital contacts due to CD of at least one overnight stay. hospitalization due to surgery, as well as diagnostic procedures, were not included in the analysis. In our sensitivity analysis, no statistical difference was found in hospital admissions of at least two overnight stays. Hospitalisations and re-operations within 30 days of the initial surgery were attributed to surgical complications and were not included in the analysis.
      Specific values for comorbidities and age (Supplementary Tables 2 and 3) were combined to calculate the Charlson comorbidity index (CCI) at the date of the initial CD-related surgery. Furthermore, we compared outcomes before and after January 1, 2005, to evaluate any impact of the increasing use of biologic therapy and identify any general trends, as biologicals only became widely used in Denmark around this time.

      2.4 Statistical analysis

      Medians and interquartile ranges (IQR) were used for continuous data. Categorical data were presented with numbers and percentages. Multivariable Cox regression analysis was used when investigating time to surgery, sex, age, whether the initial surgery was performed before or after 2005, CCI and the initial surgery type. Kaplan-Meier survival analysis was used to estimate the cumulative risk of re-operations, medications, and hospitalisations after the initial surgery. Log-rank test and chi-square test were used to compare categorical variables. P-values were considered statistically significant if they were below 0.05. All statistical analyses were performed using R version 4. 0. 2 ‘Core Team (2019)’.

      3. Results

      A total of 2483 out of 9739 CD patients met the inclusion criteria. They were followed for a median of 11 years (IQR: 6–15). A total of 493 (19.9%) patients were included in group 1, 472 (19.0) were included in group 2 and 1518 (61.1%) were included in group 3. Table 1 provides further characteristics of the study groups. The cumulative risk of the main outcomes overall is presented in Supplementary Fig. 1.
      Table 1Characteristics of Crohn's disease (CD) patients in a Danish nationwide cohort with an initial resection within 29 days (group 1) of their diagnosis, between 30 and 180 days (group 2) after their diagnosis, and more than 180 days (group 3) after their diagnosis.
      Group 1 n = 493Group 2 n = 472Group 3 n = 1518p
      Female

      n (%)
      243

      (49.3)
      273

      (57.8)
      874

      (57.6)
      0.004
      Age at diagnosis

      median years (IQR)
      49

      (33–62)
      37

      (26–54)
      35

      (24–49)
      <0.001
      Duration of follow-up

      median years (IQR)
      9

      (4–13)
      8

      (4–12)
      12

      (7–16)
      <0.001
      Disease duration until surgery

      median days (IQR)
      1

      (0–9)
      86

      (51–126)
      1078

      (472–2311)
      <0.001
      Charlson comorbidity index

      median (IQR)
      1 (0–2)0 (0–2)0 (0–1)<0.001
      First surgery type

      Ileocecal resection

      Colon resection

      Small bowel resection

      Colon and small bowel resection

      Uncategorized resection


      225 (45.6)

      150 (30.4)

      93 (18.9)

      22 (4.5)

      3 (0.6)


      257 (54.4)

      138 (29.2)

      58 (12.3)

      18 (3.8)

      1 (0.2)


      709 (46.7)

      480 (31.6)

      204 (13.41)

      107 (7.0)

      18 (1.2)
      <0.001
      Second surgery typen = 114n = 105n = 3880.074
       ○ Ileocecal resection17 (14.9)14 (13.3)45 (11.6)
       ○ Colon resection47 (41.2)40 (38.1)149 (38.4)
       ○ Small bowel resection26 (22.8)27 (25.7)110 (28.4)
       ○ Colon and small bowel resection22 (19.3)18 (17.1)65 (16.8)
       ○ Uncategorized resection2 (1.8)6 (5.7)19 (4.9)
      Test used: chi-sqare test for categorical data and t-test for continuous data.

      3.1 Re-operations

      In total, 114 (23.1%) patients from group 1, 105 (22.2%) patients from group 2 and 388 (25.6%, p = 0.256) patients from group 3 required at least one reoperation a median of 26, 22 and 28 months after the initial surgery, respectively (IQR: 7–85, 9–71, 9–68, p<0.001, Supplementary Table 4). We observed 142 (28.7%) re-operations in group 1, 145 (30.5%) in group 2 and 521 (34.2%, p = 0.274) in group 3. Fewer patients in group 1 had more than one reoperation than patients from groups 2 and 3 (n = 22 (19.3%), n = 31 (29.5%), n = 97 (25.0%), p = 0.212). Group 1 had a lower cumulative risk of re-operation within five years than groups 2 and 3 (p = 0.004) (Fig. 1A and Supplementary Table 5). After 2005, group 3 patients had a higher cumulative risk of re-operation, but this was only significant from around three years after the initial surgery (p = 0.049, Fig. 1C). When comparing time to re-operation, patients who underwent their initial surgery before 2005 had a much longer time to re-operation than patients with their initial surgery performed after 2005 (median of 57 months (IQR 18–101) vs. median of 13 months (IQR 6–43), p<0.001). The cumulative risk of re-operation before and after 2005 showed neither significant differences overall nor in each group separately (overall: p = 0.094, Supplementary Fig. 2A and group 1: p = 0.160, group 2: p = 0.410, group 3: p = 0.150, Fig. 1B-C).
      Fig 1:
      Fig. 1Cumulative risk of re-operation among Crohn's disease patients in a Danish nationwide cohort with an initial resection within 29 days (group 1) of their diagnosis, between 30 and 180 days after (group 2) their diagnosis, and more than 180 days after (group 3) their diagnosis.
      In group 1, patients who initially underwent a small bowel resection had the highest re-operation rate, while this was the case for colon and small bowel resected patients in group 2 and colon resections in group 3 (n = 27 (29.0%), 10 (55.6%), n = 154 (32.1%)). In groups 2 and 3, the cumulative risk of re-operation was lower among those who initially underwent an ileocecal resection compared to other types of surgery (p<0.001 and p<0.001, respectively) (Supplementary Fig. 3B-C).

      3.2 Hospitalization

      In group 1, 329 (66.8%) patients had at least one hospital admission, while this was the case for 269 (56.5%) patients from group 2 and 803 (52.8%, p<0.001) patients from group 3 within a median of three, six, and eight months of their surgery (IQR: 2–13, 3–27, 3–28, p = 0.040, Supplementary Table 4). Group 1 had fewer hospitalisations per patient than groups 2 and 3 (208 (63.2%) vs. 193 (71.8%) and 564 (70.2%), p = 0.033) However, the cumulative risk of hospitalization was higher among patients from group 1, while the risk among groups 2 and 3 were similar (Fig. 2A and Supplementary Table 5, p<0.001). Differences in the cumulative risk of hospitalization stratified by initial surgery type are shown in Supplementary Fig. 4. Overall, and within each group, the cumulative risk of hospitalization was lower after 2005 than before 2005 (overall: p<0.001, group 1: p<0.001, group 2: p = 0.004, group 3: p<0.001, Fig. 2B-C and Supplementary Fig. 2B).
      Fig 2:
      Fig. 2Cumulative risk of hospitalization among Crohn's disease patients in a Danish nationwide cohort with an initial resection within 29 days (group 1) of their diagnosis, between 30 and 180 days after (group 2) their diagnosis, and more than 180 days after (group 3) their diagnosis.

      3.3 Medications

      IBD medication was administered to 324 (65.7%), 303 (64.2%) and 989 (65.2%, p = 0.880) patients from groups 1, 2 and 3, respectively, within four months of initial surgery (IQR: 1–19, 1–19, 1–17, p = 0.033). The number of patients from groups 2 and 3 who continued their medication beyond surgery was 199 (71.6%) and 945 (70.9%), respectively. In addition, 79 (28.4%) patients from group 2 and 388 (29.1%) patients from group 3 discontinued their IBD medication prior to their initial surgery.
      The overall use and cumulative use of IBD medication after surgery are shown in Supplementary Fig. 5 (p = 0.003 and p<0.001). In all three groups, the IBD medication with the highest overall use was systemic steroids. In groups 1 and 3, immunomodulators were the medication type with the highest cumulative use, while it was systemic steroids in group 2. The use of IBD medication stratified by six months, one year, five years and 10 years after initial surgery is presented in Table 3. The use of 5-ASA and rectal steroids in both groups decreased significantly after 2005 (Supplementary Tables 6 and 7). After 2005 an increase in the use of immunomodulators and biological therapy was found in the whole population and within each group (immunomodulators overall: p<0.001, group 1: p = 0.006, group 2: p = 0.006, group 3: p = 0.180; biological therapy overall: p<0.001, group 1: p = 0.004, group 2: p = 0.055, group 3: p<0.001, Supplementary Figs. 2C-D and 6).
      Patients from group 2 had the lowest risk of needing immunomodulators in the study period overall (p<0.001, Fig. 3A and Supplementary Table 5). Patients from group 3 had a higher cumulative risk of receiving biological therapy, but the difference between the three study groups was statistically insignificant (p = 0.210, Fig. 3B and Supplementary Table 5). Immunomodulators were the preferred treatment for post-operative prophylaxis in all groups (Table 3). Furthermore, group 1 was less frequently treated with corticosteroids following surgery than were groups 2 and 3 (0.379 regimens per patient vs. 0.468 and 0.461 regimens per patient, respectively, p = 0.177, Table 3). However, group 1 had more budesonide regimens than groups 2 and 3 (0.839 vs. 0.538 and 0.596, respectively, p<0.001, Table 3).
      Fig 3:
      Fig. 3Cumulative risk of receiving immunomodulators (A) and biologicals (B) after initial surgery among Crohn's disease patients in a Danish nationwide cohort with an initial resection within 29 days (group 1) of their diagnosis, between 30 and 180 days after (group 2) their diagnosis, and more than 180 days after (group 3) their diagnosis.
      When analysing re-operation rates stratified by biological exposure after surgery, 569 patients out of 2329 non-exposed patients underwent a re-operation (24.4%). Thirty-eight patients of 154 exposed to biologicals later had a re-operation (24.7%, p = 0.650). No statistically significant differences were found in survival analysis when analysing all groups together or separately (group 1: p = 0.592, group 2: p = 0.342, group 3: p = 0.494).
      Of the 2406 patients not exposed to biological therapy, 1364 were hospitalized (56.7%), while of the 77 patients exposed to biologicals, 37 patients were later hospitalized (48.1%, p = 0.062). There was no statistically significant difference in the survival analysis overall, nor in each group separately (group 1: p = 0.016 (exposed group was fewer than 10 patients), group 2: p = 0.513, group 3: p = 0.711).

      3.4 Cox regression analysis

      Table 2 presents the multivariate Cox regression analysis, which showed a 27.6% higher risk of re-operation among patients in group 3 than patients in group 1 (HR: 1.276 [1.027–1.586], p = 0.028). Patients who initially underwent ileocecal resection had a lower risk of re-operation than patients who underwent other types of surgery.
      Table 2Multivariate Cox regression analysis of re-operations, hospitalisations and use of immunomodulators post-operatively among Crohn's disease (CD) patients in a Danish nationwide cohort with an initial resection within 29 days (group 1) of their diagnosis, between 30 and 180 days (group 2) after their diagnosis, and more than 180 days (group 3) after their diagnosis.
      re-operationhospitalizationImmunomodulator
      Resection time
      Group 1referencereferencereference
      Group 21.042 (0.796–1.362)

      p = 0.766
      0.719 (0.610–0.846)

      p<0.001
      0.560 (0.443–0.709)

      p<0.001
      Group 31.276 (1.027–1.586)

      p = 0.028
      0.667 (0.584–0.762)

      p<0.001
      0.825 (0.694–0.981)

      p = 0.030
      Gender
      Malereferencereferencereference
      Female

      1.106 (0.941–1.301)

      p = 0.223
      1.094 (0.984–1.217)

      p = 0.098
      0.841 (0.732–0.967)

      p = 0.015
      Age at diagnosis
      Under 40referencereferencereference
      40 and above0.828 (0.696–0.985)

      p = 0.033
      0.833 (0.743–0.934)

      p = 0.002
      0.643 (0.551–0.750)

      p<0.001
      Initial surgery
      Performed after 2005referencereferencereference
      Performed before 20050.8794 (0.738–1.049)

      p = 0.152
      1.478 (1.326–1.647)

      p<0.001
      0.717 (0.615–0.835)

      p<0.001
      Charlson comorbidity index
      Under 2referencereferencereference
      2 and above0.812 (0.572–1.152)

      p = 0.242
      0.666 (0.520–0.853)

      p = 0.001
      0.544 (0.376–0.787)

      p = 0.001
      Initial surgery type
      Ileocecal resectionreferencereferencereference
      Colon re1section1.826 (1.513–2.205)

      p<0.001
      1.143 (1.008–1.297)

      p = 0.038
      0.857 (0.721–1.019)

      p = 0.080
      Small bowel resection1.798 (1.415–2.284)

      p<0.001
      1.084 (0.920–1.277)

      p = 0.335
      1.237 (1.012–1.512)

      p = 0.038
      Colon and small bowel resection1.586 (1.138–2.210)

      p = 0.007
      1.380 (1.113–1.709)

      p = 0.003
      1.153 (0.864–1.540)

      p = 0.333
      Uncategorized resection1.064 (0.438–2.585)

      p = 0.891
      1.157 (0.667–2.006)

      p = 0.605
      1.455 (0.775–2.731)

      p = 0.243
      Table 3Prevalence of IBD medication received after initial surgery among Crohn's disease (CD) patients in a Danish nationwide cohort with an initial resection within 29 days (group 1) of their diagnosis, between 30 and 180 days (group 2) after their diagnosis, and more than 180 days (group 3) after their diagnosis.
      Group 1Group 2Group 3p
      Any use0.003
      5-ASA and rectal steroids, n (%)
      6 months94 (19.1)59 (12.5)181 (11.9)
      1 year103 (20.9)68 (14.4)202 (13.3)
      5 years124 (25.2)81 (17.2)269 (17.7)
      10 years130 (26.4)84 (17.8)281 (18.5)
      Systemic steroids, n (%)
      6 months64 (13.0)70 (14.8)182 (12.0)
      1 year91 (18.5)98 (20.8)256 (16.9)
      5 years169 (34.3)183 (38.8)512 (33.7)
      10 years203 (41.2)206 (43.6)588 (38.7)
      Immunomodulators, n (%)
      6 months76 (15.4)71 (15.0)309 (20.4)
      1 year93 (18.9)78 (16.5)360 (23.7)
      5 years151 (30.6)102 (21.6)466 (30.7)
      10 years178 (36.1)111 (23.5)492 (32.4)
      Biologic therapy, n(%)
      6 months4 (0.8)6 (1.3)56 (3.7)
      1 year7 (1.4)10 (2.1)73 (4.8)
      5 years24 (4.9)18 (3.8)97 (6.4)
      10 years37 (7.5)25 (5.3)103 (6.8)
      Initial prophylactic therapy following surgery, n(%)0.003
      Immunomodulators

      Before/after 2005
      191 (95.5)

      89 (93.7) /102 (97.1)
      112 (91.1)

      35 (94.6) /77 (89.5)
      478 (85.7)

      152 (98.1) /326 (80.9)
      Biologic therpay

      Before/after 2005
      8 (4.0)

      5 (5.3) /3 (2.8)
      11 (8.9)

      2 (5.4) /9 (10.5)
      77 (13.8)

      2 (1.3) /75 (18.6)
      In combination

      Before/after 2005
      1 (0.5)

      1 (1.1)/ ÷
      ÷3 (0.5)

      1 (0.6)/2 (0.5)
      Corticosteroid regimens following surgery, n regimens (regimens per patient)187 regimens

      (0.379)
      221 regimens

      (0.468)
      699 regimens

      (0.461)
      0.177
      Oral budenosid regimens following surgery, n regimens (regimens per patient)413 regimens

      (0.839)
      254 regimens

      (0.538)
      904 regimens

      (0.596)
      <0.001
      Patients from groups 2 and 3 had a 28.1% and 33.3% lower risk of hospitalization, respectively, than patients in group 1 (HR: 0.719 [0.610–0.846], p<0.001 and HR: 0.667 [0.584–0.762], p<0.001). Additionally, patients whose initial surgery was performed before 2005 had a 47.8% higher risk of hospitalization than patients who underwent surgery after 2005 (HR: 1.478 [1.326–1.647], p<0.001, Table 2).
      Multivariate Cox regression analysis showed that group 1 had the highest risk of needing immunomodulators after surgery (group 2 HR: 0.560 (0.443–0.709), p<0.001 and group 3 HR: 0.825 (0.694–0.981), p = 0.030). The risk of needing immunomodulators increased by 28.3% among patients who underwent surgery after 2005 compared to patients with surgery before 2005 (HR: 0.717 (0.615–0.835), p<0.001).

      4. Discussion

      In this nationwide cohort study, patients undergoing resection within 29 days of their diagnosis had a lower cumulative risk of re-operation. They were also less frequently treated with immunomodulators during the first three years after their operation. The use of immunomodulators and biologicals increased after 2005 in all three study groups; only the hospitalization rates decreased, whereas the rates of re-operation remained stable.
      A previous Danish population-based study found the five-year cumulative risk of re-operation to be 35% for CD patients, while we found risks of only 16.5% and 18.2% in groups 1 and 2, respectively [
      • Vester-Andersen M.K.
      • Vind I.
      • Prosberg M.V
      • et al.
      Hospitalisation, surgical and medical recurrence rates in inflammatory bowel disease 2003–2011—A Danish population-based cohort study.
      ]. The LIR!C trial [
      • Stevens T.W.
      • Haasnoot M.L.
      • D'Haens G.R.
      • et al.
      Laparoscopic ileocaecal resection versus infliximab for terminal ileitis in Crohn's disease: retrospective long-term follow-up of the LIR!C trial.
      ] and a retrospective study from Italy [
      • ARATARI A.
      • PAPI C.
      • LEANDRO G.
      • et al.
      Early versus late surgery for ileo-caecal Crohn's disease.
      ] also reported a better prognosis for early-resected CD patients (with short ileocecal disease) than for medically treated patients or those who underwent surgery late in their disease course. The Italian study found a longer post-operative period without clinical recurrence in patients who underwent an early resection compared to patients with a later resection. The LIR!C trial showed that early ileocecal resected patients switched less frequently to anti-TNF treatment (26%) compared to those initially treated with anti-TNF who later required surgery (48%) [
      • ARATARI A.
      • PAPI C.
      • LEANDRO G.
      • et al.
      Early versus late surgery for ileo-caecal Crohn's disease.
      ,
      • Stevens T.W.
      • Haasnoot M.L.
      • D'Haens G.R.
      • et al.
      Laparoscopic ileocaecal resection versus infliximab for terminal ileitis in Crohn's disease: retrospective long-term follow-up of the LIR!C trial.
      ,
      • Windsor A.C.J.
      Ileal Crohn's disease is best treated by surgery.
      ].
      The lower proportion and cumulative risk of re-operation in patients undergoing surgery within 29 days of diagnosis indicate a more benign post-operative disease course than in late-resected patients. These patients tended to be older at diagnosis than patients resected 180 days or more after their diagnosis, which could indicate that they had a longer period of subclinical inflammation before developing stricture or acute symptoms, while patients who were resected later might have had a chronic disease course.
      The multivariate Cox regression analysis showed that immunomodulators were more commonly used post-operatively among patients resected within 29 days of a diagnosis. We hypothesize that these patients were likely operated on initially due to fibro-stenotic disease and were therefore offered treatment with immunomodulators at an earlier stage. However, the difference in the cumulative use of immunomodulators between patients resected within 29 days of, and 180 days after, their diagnosis was modest. One possible explanation for this is that late-resected patients could have continued the medication they were taking prior to surgery, while patients resected within 29 days would have been naïve to IBD medication. This could also explain the peak in the cumulative prevalence of immunomodulators and biologicals among early-resected patients in group 1 later on in the disease course.
      Nearly 70% of CD patients treated with 5-ASA in our cohort were treated before 2005, and several studies around this time describe some ineffectiveness in the use of prophylactic 5-ASA [
      • Torres J.
      • Bonovas S.
      • Doherty G.
      • et al.
      ECCO guidelines on therapeutics in Crohn's disease: medical treatment.
      ]. These findings of frequent use of 5-ASA despite the lack of clear clinical benefit of 5-ASA in CD is in line with other recent studies [
      • Burisch J.
      • Kiudelis G.
      • Kupcinskas L.
      • et al.
      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.
      ,
      • Burisch J.
      • Bergemalm D.
      • Halfvarson J.
      • et al.
      The use of 5-aminosalicylate for patients with Crohn's disease in a prospective European inception cohort with 5 years follow-up – an Epi-IBD study.
      ,
      • Niewiadomski O.
      • Studd C.
      • Hair C.
      • et al.
      Prospective population-based cohort of inflammatory bowel disease in the biologics era: disease course and predictors of severity.
      ]. The overall use of immunomodulators and biological therapy was low but a population-based study also demonstrates a lower use of biologicals in Denmark around the early 2000s [
      • Lo B.
      • Vind I.
      • Vester-Andersen M.K.
      • et al.
      Direct and indirect costs of inflammatory bowel disease: ten years of follow-up in a danish population-based inception Cohort.
      ]. This illustrates the distance between evidence-based recommendations and real-world clinical practice. Nevertheless, we did observe a trend of increased use of both immunomodulators and biologicals after 2005 in all three groups. Despite this increase, we found a stable rate of re-operation after 2005. Interestingly, the time to re-operation was much shorter after 2005. One explanation for this could be that close monitoring strategies have been implemented in IBD care during recent decades [
      • De Cruz P.
      • Hamilton A.L.
      • Burrell K.J.
      • et al.
      Endoscopic prediction of Crohn's disease postoperative recurrence.
      ], resulting in faster reactions to post-operative disease recurrence.
      Our study found a decrease in the cumulative risk of hospitalization after 2005 in all groups, an observation supported by several other studies [
      • Mao E.J.
      • Hazlewood G.S.
      • Kaplan G.G.
      • et al.
      Systematic review with meta-analysis: comparative efficacy of immunosuppressants and biologics for reducing hospitalisation and surgery in Crohn's disease and ulcerative colitis.
      ,
      • Jeuring S.F.G.
      • van den Heuvel T.R.A.
      • Liu L.Y.L.
      • et al.
      Improvements in the long-term outcome of crohn's disease over the past two decades and the relation to changes in medical management: results from the population-based IBDSL Cohort.
      ]. A greater number of patients resected within 29 days of their diagnosis required hospitalization than late-resected patients. However, groups 2 and 3 had a higher number of hospitalisations per patient. The cumulative risk of hospitalization among patients undergoing surgery within 29 days was especially high during the first year after the initial surgery, as observed in several previous studies [
      • Vester-Andersen M.K.
      • Vind I.
      • Prosberg M.V
      • et al.
      Hospitalisation, surgical and medical recurrence rates in inflammatory bowel disease 2003–2011—A Danish population-based cohort study.
      ,
      • Golovics P.A.
      • Lakatos L.
      • Mandel M.D.
      • et al.
      Prevalence and predictors of hospitalization in Crohn's disease in a prospective population-based inception cohort from 2000 to 2012.
      ]. The fact that newly diagnosed CD patients were hospitalised more frequently could be due to differences in disease phenotype, since we assume these early-resected patients had B2 disease behavior. They probably also experienced their CD for the first time in an acute setting, and several studies have indicated that stressful events and psychological strain increase disease activity [
      • Agostini A.
      • Ballotta D.
      • Righi S.
      • et al.
      Stress and brain functional changes in patients with Crohn's disease: a functional magnetic resonance imaging study.
      ,
      • Agostini A.
      • Filippini N.
      • Benuzzi F.
      • et al.
      Functional magnetic resonance imaging study reveals differences in the habituation to psychological stress in patients with Crohn's disease versus healthy controls.
      ]. A final possibility is that differences in treatment schedules and follow-up among the study groups could explain some of these varying results.
      Few studies have focused on the disease course in early-resected CD patients and, to our knowledge, these have only been among ileocecal-resected patients. As such, our study is unique for investigating CD patients with a major abdominal surgery both early and later on in their disease course. Another major strength of this study is its population-based design, the large number of patients included, and its long follow-up. However, the study has several limitations. It is based on the Danish National Patient Register, which only contains data on diagnoses, surgeries, and prescriptions; as such, we did not have access to data about symptoms or other clinical activity and therefore any diagnostic delay could not be determined. Furthermore, we did not have data about disease location or smoking habits; instead, we tried to compensate by analysing different surgery types, age at the time of diagnosis, and specific outcomes.
      Information on phenotypes in the Danish National Patient Register has previously been shown to have a low validity, therefore we chose not to include these data [
      • Lo B.
      • Vind I.
      • Vester-Andersen M.K.
      • et al.
      Validation of ulcerative colitis and Crohn's disease and their phenotypes in the Danish National Patient Registry using a population-based cohort.
      ]. Another drawback is that coding errors for procedures and diagnoses in the registers could have resulted in inaccurate classifications; however, previous studies have shown a high validity for IBD-specific diagnosis codes [
      • Lo B.
      • Vind I.
      • Vester-Andersen M.K.
      • et al.
      Validation of ulcerative colitis and Crohn's disease and their phenotypes in the Danish National Patient Registry using a population-based cohort.
      ]. To our knowledge, the codes for resections have not yet been formally validated. It is possible that some surgeries were performed for reasons other than IBD. We found nine patients that received a diagnosis of colorectal cancer within 30 days of their IBD surgery; they were not included in the cohort. We attempted to limit the effect of any coding errors by grouping procedure types according to anatomical location. Reassuringly, the number of patients with uncategorised resections in our study was relatively small.
      In this nationwide cohort study, we found that re-operation rates and the cumulative prevalence of immunomodulators after surgery were lower among Crohn's disease patients resected within 29 days of, and between 30 and 180 days after, their diagnosis compared to patients resected after 180 days, suggesting a more benign post-operative disease course in early-resected patients. We observed an overall increase in the use of biological therapy and immunomodulators as post-operative prophylaxis treatment after 2005. We also observed a reduction in hospitalisations after 2005, while the difference in the cumulative risk of undergoing a re-operation was statistically insignificant. Additionally, we found that re-operations were performed earlier on in the disease course from 2005 onwards, suggesting closer monitoring and quicker decisions to perform re-operations. Further studies are needed to elaborate on the prognoses of early- and late-resected CD patients, and especially prospective studies with access to data on disease phenotypes, in order to identify patient groups who will benefit most from early surgery.

      Authors’ contributions

      All authors have critically reviewed the manuscript for content and approved it for publication. MZ. Sarikaya and M. Zhao had full access to the data and take responsibility for its veracity and statistical analysis. All authors were responsible for interpreting the data. F. Bendtsen and J. Burisch supervised the study.

      Funding

      This study received no funding.

      Declaration of competing interest

      MS. Sarikaya, Melek: None. MD. Zhao, Mirabella: Travel fees from Takeda Pharma A/S. MD. Lo, Bobby: None. Prof. Bendtsen, Flemming: Personal fee and grants from Ferring A/S, none of which were related to this study. MD. and PhD Burisch, Johan: Personal fees from AbbVie, Janssen-Cilag, Celgene, MSD, Pfizer; Pharmacomos, Jansen, Tillots Pharma grants and personal fees from Takeda, Tillots Pharma, MSD, Novo Nordisk Foundation, BMS none of which were related to this study.

      Appendix. Supplementary materials

      References

        • Alatab S.
        • Sepanlou S.G.
        • Ikuta K.
        • et al.
        The global, regional, and national burden of inflammatory bowel disease in 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017.
        Lancet Gastroenterol Hepatol. 2020; 5: 17-30
        • Dittrich A.E.
        • Sutton R.T.
        • Haynes K.
        • et al.
        Incidence Rates for Surgery in Crohn's Disease Have Decreased: a Population-based Time-trend Analysis.
        Inflamm Bowel Dis. 2020; 26: 1909-1916
        • Mao E.J.
        • Hazlewood G.S.
        • Kaplan G.G.
        • et al.
        Systematic review with meta-analysis: comparative efficacy of immunosuppressants and biologics for reducing hospitalisation and surgery in Crohn's disease and ulcerative colitis.
        Aliment Pharmacol Ther. 2017; 45: 3-13
        • Burisch J.
        • Kiudelis G.
        • Kupcinskas L.
        • et al.
        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.
        Gut. 2019; 68: 423-433
        • Murthy S.K.
        • Begum J.
        • Benchimol E.I.
        • et al.
        Introduction of anti-TNF therapy has not yielded expected declines in hospitalisation and intestinal resection rates in inflammatory bowel diseases: a population-based interrupted time series study.
        Gut. 2020; 69: 274-282
        • Cosnes J.
        • Gower–Rousseau C.
        • Seksik P.
        • et al.
        Epidemiology and Natural History of Inflammatory Bowel Diseases.
        Gastroenterology. 2011; 140: 1785-1794
        • Kalman T.D.
        • Everhov Å.H.
        • Nordenvall C.
        • et al.
        Decrease in primary but not in secondary abdominal surgery for Crohn's disease: nationwide cohort study, 1990–2014.
        Br J Surg. 2020; 107: 1529-1538
        • Silverstein M.D.
        • Loftus E.V.
        • Sandborn W.J.
        • et al.
        Clinical course and costs of care for Crohn's disease: markov model analysis of a population-based cohort.
        Gastroenterology. 1999; 117: 49-57
        • Ha F.J.
        • Thong L.
        • Khalil H.
        Quality of life after intestinal resection in patients with crohn disease: a systematic review.
        Dig Surg. 2017; 34: 355-363
        • D'Ugo S.
        • Romano F.
        • Sibio S.
        • et al.
        Impact of surgery on quality of life in Crohn's disease: short- and mid-term follow-up.
        Updates Surg. 2020; 72: 773-780
        • Wright E.K.
        • Kamm M.A.
        • De Cruz P.
        • et al.
        Effect of Intestinal Resection on Quality of Life in Crohn's Disease.
        J Crohn's Colitis. 2015; 9: 452-462
        • Lo B.
        • Prosberg M.V .V..
        • Gluud L.L.L.
        • et al.
        Systematic review and meta-analysis: assessment of factors affecting disability in inflammatory bowel disease and the reliability of the inflammatory bowel disease disability index.
        Aliment Pharmacol Ther. 2018; 47: 6-15
        • An V.
        • Cohen L.
        • Lawrence M.
        • et al.
        Early surgery in Crohn's disease a benefit in selected cases.
        World J Gastrointest Surg. 2016; 8: 492
        • ARATARI A.
        • PAPI C.
        • LEANDRO G.
        • et al.
        Early versus late surgery for ileo-caecal Crohn's disease.
        Aliment Pharmacol Ther. 2007; 26: 1303-1312
        • Stevens T.W.
        • Haasnoot M.L.
        • D'Haens G.R.
        • et al.
        Laparoscopic ileocaecal resection versus infliximab for terminal ileitis in Crohn's disease: retrospective long-term follow-up of the LIR!C trial.
        Lancet Gastroenterol Hepatol. 2020; 5: 900-907
        • Windsor A.C.J.
        Ileal Crohn's disease is best treated by surgery.
        Gut. 2002; 51: 11-12
        • Bøcker Pedersen C.
        • Gøtzsche H.
        • Ostrup Møller J.
        • et al.
        The Danish Civil Registration System. A cohort of eight million persons.
        Dan Med Bull. 2006; 53: 441-449
        • Andersen T.F.
        • Madsen M.
        • Jørgensen J.
        • et al.
        The Danish National Hospital Register. A valuable source of data for modern health sciences.
        Dan Med Bull. 1999; 46: 263-268
        • Wallach Kildemoes H.
        • Toft Sørensen H.
        • Hallas J
        The Danish national prescription registry.
        Scand J Public Health. 2011; 39: 38-41
        • Lo B.
        • Vind I.
        • Vester-Andersen M.K.
        • et al.
        Validation of ulcerative colitis and Crohn's disease and their phenotypes in the Danish National Patient Registry using a population-based cohort.
        Scand J Gastroenterol. 2020; 55: 1171-1175
        • Lophaven S.N.
        • Lynge E.
        • Burisch J.
        The incidence of inflammatory bowel disease in Denmark 1980-2013: a nationwide cohort study.
        Aliment Pharmacol Ther. 2017; 45: 961-972
        • Fonager K.
        • Sørensen H.T.
        • Rasmussen S.N.
        • et al.
        Assessment of the diagnoses of Crohn's disease and ulcerative colitis in a Danish hospital information system.
        Scand J Gastroenterol. 1996; 31: 154-159
        • Vester-Andersen M.K.
        • Vind I.
        • Prosberg M.V
        • et al.
        Hospitalisation, surgical and medical recurrence rates in inflammatory bowel disease 2003–2011—A Danish population-based cohort study.
        J Crohn's Colitis. 2014; 8: 1675-1683
        • Torres J.
        • Bonovas S.
        • Doherty G.
        • et al.
        ECCO guidelines on therapeutics in Crohn's disease: medical treatment.
        J Crohn's Colitis. 2020; 14: 4-22
        • Burisch J.
        • Bergemalm D.
        • Halfvarson J.
        • et al.
        The use of 5-aminosalicylate for patients with Crohn's disease in a prospective European inception cohort with 5 years follow-up – an Epi-IBD study.
        United Eur Gastroenterol J. 2020; 8: 949-960
        • Niewiadomski O.
        • Studd C.
        • Hair C.
        • et al.
        Prospective population-based cohort of inflammatory bowel disease in the biologics era: disease course and predictors of severity.
        J Gastroenterol Hepatol. 2015; 30: 1346-1353
        • Lo B.
        • Vind I.
        • Vester-Andersen M.K.
        • et al.
        Direct and indirect costs of inflammatory bowel disease: ten years of follow-up in a danish population-based inception Cohort.
        J Crohn's Colitis. 2020; 14: 53-63
        • De Cruz P.
        • Hamilton A.L.
        • Burrell K.J.
        • et al.
        Endoscopic prediction of Crohn's disease postoperative recurrence.
        Inflamm Bowel Dis. 7 July 2021; (Epub ahead of print)https://doi.org/10.1093/ibd/izab134
        • Jeuring S.F.G.
        • van den Heuvel T.R.A.
        • Liu L.Y.L.
        • et al.
        Improvements in the long-term outcome of crohn's disease over the past two decades and the relation to changes in medical management: results from the population-based IBDSL Cohort.
        Am J Gastroenterol. 2017; 112: 325-336
        • Golovics P.A.
        • Lakatos L.
        • Mandel M.D.
        • et al.
        Prevalence and predictors of hospitalization in Crohn's disease in a prospective population-based inception cohort from 2000 to 2012.
        World J Gastroenterol. 2015; 21: 7272-7280
        • Agostini A.
        • Ballotta D.
        • Righi S.
        • et al.
        Stress and brain functional changes in patients with Crohn's disease: a functional magnetic resonance imaging study.
        Neurogastroenterol Motil. 2017; 29: e13108
        • Agostini A.
        • Filippini N.
        • Benuzzi F.
        • et al.
        Functional magnetic resonance imaging study reveals differences in the habituation to psychological stress in patients with Crohn's disease versus healthy controls.
        J Behav Med. 2013; 36: 477-487