Article Text
PDF
Abstract
Objective The aim of the study was to assess whether azathioprine and antitumour necrosis factor (TNF) treatment decrease the long-term need for surgery in patients with Crohn's disease.
Methods This was an observational study of a referral centre cohort. The cumulative incidence of the first Crohn's disease-related major abdominal surgery was estimated using the Kaplan–Meier method, and independent predictors of surgery were identified using Cox proportional hazards regression with propensity scores adjustment. Receiver operating characteristic (ROC) analysis was used to identify optimal cut-offs for duration of maintenance treatments. The electronic charts of 296 incident cases of Crohn's disease from Nancy University Hospital, France, diagnosed between 2000 and 2008, were reviewed through January 2010.
Results The median follow-up time per patient was 57 months. Seventy-six patients (26%) underwent at least one major abdominal surgical procedure. The cumulative probabilities of the first Crohn's disease-related major abdominal surgery were 6.5, 25.9 and 44.3 at 1, 5 and 9 years, respectively. In the ROC analysis, the duration of anti-TNF and azathioprine treatment had significant cut-off values (≤475 days ∼16 months and ≤45 days ∼1.5 months, respectively) with positive likelihood ratios (PLRs) of 1.52 (p<0.0001) and 1.51 (p=0.003) for the first Crohn's disease-related major abdominal surgery. Using multivariate Cox proportional hazards regression analysis (after propensity score adjustment), independent positive predictors of major abdominal surgery were stricturing (HR=12.01; 95% CI 5.97 to 24.17) or penetrating (HR=10.77; 95% CI 4.87 to 23.80) disease behaviour at diagnosis, duration of anti-TNF treatment of <16 months (HR=3.86; 95% CI 1.77 to 8.45) and duration of azathioprine treatment of <1.5 months (HR=2.00; 95% CI 1.20 to 3.34).
Conclusions Non-complicated inflammatory disease behaviour and long-term anti-TNF treatment are associated with a lower risk for surgery whereas azathioprine only modestly lowers this risk.
- Crohn's disease-related surgery
- tumour necrosis factor antagonists
- azathioprine
- IBD
- IBD clinical
- infliximab
- surgery for IBD
Statistics from Altmetric.com
- Crohn's disease-related surgery
- tumour necrosis factor antagonists
- azathioprine
- IBD
- IBD clinical
- infliximab
- surgery for IBD
Significance of this study
What is already known about this subject?
In paediatric Crohn's disease, azathioprine treatment was associated with a decreased requirement for surgery.
In adult Crohn's disease, an increased and earlier thiopurine use was temporally associated with a reduction in the need for surgery.
The effect of antitumour necrosis factor (TNF) treatment on the long-term risk for surgery in Crohn's disease is unknown.
What are the new findings?
This is the first study that investigates the association between the duration of azathioprine and anti-TNF treatments and the requirement for surgery in patients newly diagnosed with Crohn's disease.
Anti-TNF treatment was associated with a reduction in the need for surgery in patients newly diagnosed with Crohn's disease.
Compared with anti-TNF agents, azathioprine had modest efficacy in reducing the risk for long-term surgery in patients newly diagnosed with Crohn's disease.
How might it impact on clinical practice in the foreseeable future?
Both azathioprine and TNF antagonists may be disease-modifying agents, with azathioprine only modestly reducing the risk of surgery in the long term.
Introduction
Crohn's disease is a condition that causes disability over time. Many patients with Crohn's disease will require surgical intervention during the course of the disorder.1 Historic referral centre-based cohort studies have shown that three-quarters of patients with Crohn's disease require surgical treatment during follow-up.2 3 More recently, a French referral centre reported that approximately one-third of its patients with Crohn's disease underwent surgery during follow-up.2 Furthermore, population-based cohort studies indicate that the cumulative risk for surgery 10 years after diagnosis ranges from 40% to 71%.3–8
Despite the current widespread use of biologics and immunomodulators in Crohn's disease, long-term data on the impact of potentially disease-modifying agents on the need for surgery are scarce. In Paris, France, the increasing use of immunosuppressants such as azathioprine has not been accompanied by a commensurate decrease in the need for surgery in Crohn's disease.2 It is important to note, however, that <10% of patients with Crohn's disease received azathioprine before surgery.2 Recently, in Cardiff, Wales, a well-defined, population-based cohort study found that a reduction in the need for surgery was temporally associated with increased and earlier thiopurine use.8 However, the authors did not assess the impact of antitumour necrosis factor (TNF) treatment on the risk of surgery.
The need for surgery in patients with Crohn's disease treated with infliximab was assessed in two randomised, placebo-controlled trials.9 10 In both trials, treatment of the patients with anti-TNF therapy was associated with a lower risk for surgery.9 10 Whether anti-TNF treatment is associated with a decreased need for long-term surgery is unknown, and the impact of thiopurines on such a requirement remains poorly understood, despite the increasing use of immunosuppressants.
The aims of this study were therefore (1) to evaluate the cumulative probability of surgery in Crohn's disease in the era of biologics; (2) to assess the cumulative probability of receiving Crohn's disease medications (5-aminosalicylates, steroids, thiopurines, methotrexate and anti-TNF agents); and (3) to assess the association of baseline factors and Crohn's disease treatment duration with need for surgery. For that purpose, we used a cohort of patients with newly diagnosed Crohn's disease who were followed in a French referral centre according to a standardised protocol.
Methods
Study population
The Nancy inflammatory bowel disease (IBD) cohort is a referral centre-based cohort including 296 patients newly diagnosed with Crohn's disease. The patients were first diagnosed with Crohn's disease between 15 January 2000 and 24 September 2008, using the Lennard-Jones criteria.11 The duration of follow-up for each patient was based on the date of diagnosis and the date of the last follow-up, up to January 2010.
All demographic information—including date of birth, date of Crohn's disease diagnosis, disease type and extent, and dates and type of surgeries—were prospectively recorded. In addition, we also collected information regarding treatment with oral 5-aminosalicylates, oral and/or intravenous steroids, thiopurines, methotrexate and anti-TNF agents. The start and end dates for each maintenance treatment (mesalamine, thiopurines and anti-TNF agents) were recorded, and only the patients' electronic charts were reviewed to conduct this study.
Information about the Nancy IBD cohort is reported to the Commission Nationale de l'Informatique et des Libertés (no. 1404720), which supervises the implementation of the act regarding data processing, data files and individual liberties that came into effect on 6 January 1978, and was amended on 6 August 2004, to protect the personal data of individuals.
Statistical analysis
All quantitative variables are described as medians and percentiles (IQR, 25–75th percentile). All proportions are expressed as percentages with 95% CIs. The ‘event’ of first Crohn's disease-related major abdominal surgery was evaluated using survival analysis. The cumulative probabilities of event-free survival were estimated using the Kaplan–Meier method. The time to first surgery was considered to begin at the date of Crohn's disease diagnosis and end at the date of occurrence of the ‘event’ or last known follow-up. Patients with follow-up times of <12 months and patients in whom surgery was performed at diagnosis were excluded from the analysis. Major abdominal surgery was defined as any surgery, with the exception of perianal surgery or endoscopic dilation. The treatment of each patient with Crohn's disease with different medications and follow-up times was calculated and expressed in days.
To identify factors predictive of Crohn's disease-related major abdominal surgery among the patients' baseline characteristics, we performed a univariate analysis using the log-rank test. When considering the continuous variables for dichotomous analysis, cut-off values were determined using receiver operating characteristic (ROC) analysis, as described by DeLong et al,12 using surgery outcome as a classification variable. To identify independent predictors of surgery using a multivariate analysis, all significant variables evaluated in the log-rank test were integrated into a Cox proportional hazards regression using a stepwise selection method. All variables with p values of <0.1 were initially included in the model, and variables with p values of <0.05 were retained in the model. The results are shown as HRs with 95% CIs.
Given the observational nature of this study, we attempted to adjust HRs for the propensity of patients with Crohn's disease to be treated with disease-specific medications. We constructed a logistic regression model to generate a propensity score that expressed the likelihood of being treated with a Crohn's disease-specific medication. The outcome variable in the propensity score model was defined by the prescription of a Crohn's disease medication (azathioprine, infliximab, adalimumab or certolizumab pegol), on the basis of the demographic and clinical characteristics of each patient.13 Stepwise logistic regression analysis was used to perform this calculation. The covariates that constituted the final models were as follows: the Montreal classification items (A, B, L and P)14 and sex for all the propensity scores; receiving azathioprine treatment before infliximab for the ‘infliximab propensity score’; and receiving treatment with azathioprine, infliximab and adalimumab before the initiation of certolizumab pegol therapy for the ‘certolizumab pegol propensity score’. No covariate was retained to determine the adalimumab propensity score (data not shown). We adjusted for the propensity score in the final Cox proportional hazards regression model by using the surgery outcome as the dependent variable and the propensity scores as covariates.15
All the reported p values were two-sided, and p values of <0.05 were considered statistically significant. Statistical analyses were performed using MedCalc software, version 11.3.3.0.
Results
The baseline characteristics of the 296 patients with Crohn's disease that we evaluated are shown in table 1. These patients were followed for a total of 1374 person-years, with a median follow-up duration of 57 months (∼5 years; IQR 25–75th, 31–76 months). Of these patients, 115 (39%) were male (table 1). The median age at Crohn's disease diagnosis was 24 years (IQR 25–75th, 20–36). Perianal disease was identified at diagnosis in 105 (36%) patients. When the Montreal classification14 was applied to the baseline characteristics, 190 patients (65%) had no intestinal complications, whereas 75 (25%) patients presented intestinal stricture and 30 (10%) developed a penetrating complication. Only 7% of the patients were under 16 years of age (A1); 74% of the patients were between 17 and 40 years old (A2) and 19% of the patients were >40 years old (A3) at the time of diagnosis. Twenty-one per cent of the patients had ileitis, 30% of the patients had colitis and 49% of the patients had ileocolitis at the time of diagnosis. A familial history of IBD was reported by 9% of the patients (table 1).
Characteristics of the 296 patients with Crohn's disease diagnosed between 2000 and 2008
Incidence of major abdominal surgery
Among the 296 patients in the cohort, 76 (26%; 95% CI 21% to 31%) underwent at least one major abdominal surgical procedure after being diagnosed with Crohn's disease. The cumulative probabilities of the first major abdominal surgery at 1, 5 and 9 years from the time of diagnosis were 6.5% (±1.4%), 25.9% (±3.0%) and 44.3% (±6.2%), respectively (figure 1). Among these 76 patients, only 4 (5%) underwent a second major abdominal surgical procedure. All four patients have received treatment with adalimumab after the second surgery.
Cumulative probability of the first major abdominal surgery after Crohn's disease diagnosis.
Type of first major abdominal surgery
A total of 81 surgical procedures were performed in the 76 patients who underwent a first major abdominal surgical procedure. These surgical procedures included an ileal or ileocaecal resection in 47 out of 76 (62%) patients. A total proctocolectomy with loop ileostomy and a small bowel resection were performed in 15% (11/76) and 9% (7/76) of patients, respectively. The other types of procedures are shown in table 2.
Type of first major abdominal surgery performed in 76 patients with Crohn's disease who underwent at least one major abdominal surgery (per procedure analysis)
Use of azathioprine before the first major abdominal surgery
During the time between Crohn's disease diagnosis and the first major abdominal surgery or last follow-up in patients who did not undergo surgery, 65% of patients (95% CI 59% to 71%) were treated with azathioprine. None of the patients was treated with 6-mercaptopurine. The probabilities of being treated with azathioprine before the first major abdominal surgery were 41.3% (±3.0%), 71.7% (±3.2%) and 80.8% (±4.0%) at 1, 5 and 9 years, respectively, from the time of diagnosis (figure 2). Half of the patients had received treatment with azathioprine in the preoperative period after 19 months of follow-up.
Cumulative probability of receiving azathioprine before the first major abdominal surgery (patients were censored at first major abdominal surgery).
Use of anti-TNF treatment before the first major abdominal surgery
During the time between Crohn's disease diagnosis and the first major abdominal surgery or last follow-up in patients who did not undergo surgery, 60% of patients (95% CI 54% to 65%) received at least one anti-TNF agent (table 1). Among patients who were treated with TNF antagonists before surgery (n=176), 3 (2%) received an ‘infliximab–adalimumab–certolizumab pegol’ sequence, 46 (26%) received an ‘infliximab–adalimumab’ sequence, 2 (1%) received an ‘adalimumab–infliximab’ sequence, 115 (65%) received only infliximab and 10 (6%) received only adalimumab (Supplementary table 1 online).
The probabilities of receiving a first-line TNF antagonist at 1, 5 and 9 years from the time of diagnosis were 23.2% (±2.5%), 64.7% (±3.3%) and 85.0% (±5.5%), respectively (Supplementary figure 1 online). Half of the patients had received anti-TNF treatment in the preoperative period after 33 months of follow-up. The probabilities of receiving at least one infliximab infusion before the first major abdominal surgery at 1, 5 and 9 years from the time of diagnosis were 23.1% (±2.5%), 65.0% (±3.4%) and 83.4% (±5.9%), respectively (figure 3). The corresponding values for adalimumab were 1.4% (±0.7%), 23.4% (±3.2%) and 59.4% (±10.9%) at 1, 5 and 9 years, respectively (figure 3).
Cumulative probability of receiving infliximab, adalimumab or certolizumab pegol before the first major abdominal surgery (patients were censored at the first major abdominal surgery).
Use of Crohn's disease-specific medications during follow-up
For this analysis, we included medications that were administered to the patients before and after surgery, without censoring patients at the time of surgery. During follow-up, 39, 71, 70, 11 and 65% of patients received oral mesalamine, oral and/or intravenous corticosteroids, azathioprine, methotrexate or a first-line anti-TNF agent, respectively (table 1). The probabilities of receiving oral 5-aminosalicylates at 1, 5 and 9 years from the time of diagnosis were 28.8% (±2.7%), 36.5% (±3.1%) and 42.5% (±3.9%), respectively. The corresponding values at 1, 5 and 9 years from the time of diagnosis were 53.7% (±3.0%), 68.2% (±3.0%) and 78.3% (±4.0%), respectively, for corticosteroids (oral and/or intravenous) and 1.7% (±0.8%), 10.5% (±2.1%) and 15.5% (±3.5%), respectively, for methotrexate (figure 4).
Cumulative probability of receiving Crohn's disease medications (oral 5-aminosalicylates, oral and/or intravenous steroids, methotrexate, azathioprine or tumour necrosis factor (TNF) antagonists) from time of diagnosis to last follow-up.
The cumulative probabilities of receiving azathioprine at 1, 5 and 9 years from the time of diagnosis were 40.4% (±2.9%), 71.2% (±3.0%) and 83.4% (±4.1%), while the values corresponding to first-line anti-TNF agent were 23.2% (±2.5%), 67.0% (±3.2%) and 89.1% (±4.4%), respectively (figure 4). The corresponding values at 1, 5 and 9 years from the time of diagnosis were 22.8% (±2.5%), 64.6% (±3.3%) and 85.9% (±5.2%), respectively, for infliximab and 1.7% (±0.8%), 23.3% (±3.0%) and 59.1% (±9.6%), respectively, for adalimumab. During follow-up, half of the patients had received treatment with azathioprine or anti-TNF therapy after 20 months and 32 months, respectively.
Factors associated with major abdominal surgery
Identification of cut-off values for treatment duration using ROC analysis
To reduce the risk of bias related to arbitrarily defined treatment durations and to identify the optimal duration of exposure to maintenance treatment of Crohn's disease, we performed ROC analysis on maintenance treatment durations using the surgery outcome as a classification variable. In the ROC analysis, the duration of anti-TNF treatment had a significant cut-off value (≤475 days ∼16 months) with a positive likelihood ratio (PLR) of 1.52 (p<0.0001) for the first Crohn's disease-related major abdominal surgery. Similarly, analysis of the duration of azathioprine treatment revealed a significant threshold (≤45 days ∼1.5 months) with a PLR of 1.51 (p=0.003). ROC analysis did not identify a significant threshold for the duration of mesalamine treatment with respect to major abdominal surgery (Supplementary table 2 online). The duration of methotrexate treatment was not analysed because of the low number of patients treated during the preoperative period (n=4). In ROC analysis, the time delay between diagnosis and starting treatment with azathioprine or anti-TNF therapy was not predictive of the first Crohn's disease-related major abdominal surgery (data not shown).
Univariate analysis using the log-rank test
The factors that were evaluated in the univariate analysis using the log-rank test are shown in table 3. Age at diagnosis, gender, familial history of IBD, perianal disease, isolated upper gastrointestinal disease and ileocolonic disease were not associated with the risk of major abdominal surgery (table 3). Stricturing (B2) or penetrating (B3) disease, treatment with an anti-TNF agent for <475 days (∼16 months), as defined by ROC analysis, and treatment with azathioprine for <45 days (∼1.5 months), as defined by ROC analysis, were positively associated with major abdominal surgery (table 3). Conversely, isolated colonic disease (L2) was negatively associated with major abdominal surgery (table 3).
Predictors of first Crohn's disease-related major abdominal surgery, as determined by univariate analysis
Multivariate analysis
Compared with patients with non-stricturing, non-penetrating disease (B1), patients with stricturing or penetrating disease were found to be at increased risk for major abdominal surgery, with unadjusted HRs of 12.62 and 11.01, respectively (Supplementary table 3 online). Treatment with an anti-TNF agent for <475 days (∼16 months) was associated with a fourfold increased risk for major abdominal surgery (unadjusted HR=3.7), and azathioprine treatment for ≤45 days (∼1.5 months) was associated with a twofold increased risk for surgery (unadjusted HR=2.02) (Supplementary table 3 online). After adjusting for propensity scores, the adjusted HRs (aHRs) for major abdominal surgery were as follows: B2 (aHR=12.01; 95% CI 5.97 to 24.17); B3 (aHR=10.77; 95% CI 4.87 to 23.80); anti-TNF treatment duration ≤16 months (3.86; 95% CI 1.77 to 8.45); and azathioprine treatment duration ≤1.5 months (aHR=2.00; 95% CI 1.20 to 3.34) (table 4).
Predictors of first Crohn's disease-related major abdominal surgery, as determined by multivariate analysis (Cox proportional hazards regression) after propensity score adjustment
Discussion
This is the first study that investigates the association between the duration of azathioprine and anti-TNF treatments and the requirement for surgery in patients with newly diagnosed Crohn's disease. In the prebiological era, the risk for surgery in patients with Crohn's disease was assessed in several referral centre-based cohort studies. Among 592 patients diagnosed at the Cleveland Clinic Foundation between 1966 and 1969, 438 (74%) underwent surgery after a mean follow-up time of 13 years.16 Similarly, between 1938 and 1970, of 189 patients with Crohn's disease followed up at the Radcliffe Infirmary, Oxford, 78% required surgical treatment.17 More recently, among 565 patients with Crohn's disease seen within the first 3 months after diagnosis at a French referral centre, 190 (34%) underwent at least one surgical procedure during follow-up, and their cumulative risk of intestinal resection at 5 years ranged from 34% to 35%.2
In the era of biologicals and the increasing use of azathioprine, we found that the cumulative probability of major abdominal surgery in a cohort of patients with newly diagnosed Crohn's disease was 26% at 5 years. Interestingly, this result is in agreement with that reported in a population-based cohort of patients diagnosed in Cardiff between 1986 and 2003 in which the resection rate for patients with Crohn's disease at 5 years decreased significantly from 59% (1986–91) to 25% (1998–2003).18 Previous experience in a French referral centre suggested that despite the increased use of azathioprine treatment in patients with Crohn's disease over time, azathioprine did not appear to affect the requirement for surgery.2 However, these results should be interpreted with caution because <10% of patients were treated with azathioprine before surgery.2
In a Belgian referral centre study, approximately one-fifth of patients with Crohn's disease who were treated with infliximab underwent major abdominal surgery after a median follow-up time of 55 months.19 However, the rate of surgery in patients who were not treated with infliximab was not assessed.
The need for surgery in patients with Crohn's disease treated with anti-TNF agents remains poorly investigated by population-based studies. In Stockholm, Sweden, 33 out of 191 (17%) patients with Crohn's disease treated with infliximab underwent major abdominal surgery after a mean of 2.6 infusions.20 Two randomised, placebo-controlled trials evaluated the risk for surgery as a secondary objective in a study of patients with Crohn's disease treated with anti-TNF agents.21 22 In the ACCENT 1 trial, 11 out of 385 patients who received combined, scheduled infliximab treatment (5 or 10 mg/kg) underwent significantly fewer Crohn's disease-related surgical procedures throughout the 54-week trial period when compared with 14/188 in episodic strategy patients (p=0.01).21 In the CHARM trial, fewer major Crohn's disease-related surgical procedures occurred in patients treated with combined adalimumab (3/517, 0.6%) compared with those treated with placebo (10/261, 3.8%; p=0.005) at 1 year.22 A recent study, which used the US Nationwide Inpatient Sample to identify all hospital admissions for Crohn's disease per year between 1993 and 2004,18 reported that the rates of total colorectal and small bowel resections has not declined.18
Whether azathioprine is a disease-modifying agent in Crohn's disease remains a matter of active debate. In addition, the influence of anti-TNF treatment on the requirement for long-term surgery is unknown. In our referral centre-based cohort study, approximately half of the patients with Crohn's disease were treated with anti-TNF therapy before surgery. Our results suggest that long-term anti-TNF treatment may be associated with a lower risk for surgery. To our knowledge, the cumulative probability of receiving different types of Crohn's disease medications is unknown. We found that in patients with newly diagnosed Crohn's disease, the cumulative probabilities of receiving mesalamine, corticosteroids, azathioprine, methotrexate or first-line anti-TNF agent at 5 years were 36.5, 68.2, 71.2, 10.5 and 67.0%, respectively. In addition, we found no association between steroid treatment and the need for surgery. The lack of an association between the duration of treatment with 5-aminosalicylates and the risk for surgery is in agreement with previous reports.23 24
We also demonstrated that azathioprine treatment is associated with a reduced requirement for surgery in patients newly diagnosed with Crohn's disease. This finding is in agreement with the results from two recent population-based studies in adult and paediatric patients with Crohn's disease.8 25 However, in the study conducted in Cardiff, Wales, whether the duration of thiopurine treatment was associated with a reduced need for surgery was not investigated. In a French paediatric population-based cohort study, treatment with azathioprine for >3 months was associated with a decreased risk for surgery25; however, this study used an empirical definition of azathioprine treatment. By using ROC analysis, we confirmed this result in an adult population of patients with Crohn's disease, and we identified a significant threshold of treatment time below which the effectiveness of azathioprine was reduced and was associated with a higher risk for surgery.
The risk factors for undergoing surgery were previously investigated in three population-based cohort studies.4 6 8 In patients with Crohn's disease, the following conditions were identified as independent risk factors for surgery: having ileal, ileocolonic or orojejunal disease; treatment with oral corticosteroids within 3 months of diagnosis; early use of thiopurines within the first year of diagnosis; being <40 years old or >30; and stricturing or penetrating disease behaviour.4 6 8 In our cohort from Nancy, France, multivariate analysis confirmed that compared with patients with non-complicated disease behaviour (B1), those with a complicated disease behaviour (B2 and B3) at diagnosis have an 11- to 12-fold increased risk of undergoing major abdominal surgery.
The type of surgery performed on patients with Crohn's disease was assessed in one population-based cohort. In agreement with our findings, a study in Cardiff, Wales (1986–2003) found that ileocaecal resection was the main indication for surgery, and it was performed in 67% of patients with Crohn's disease.8 Total proctocolectomy with loop ileostomy was performed in 15% of patients from Nancy who were diagnosed between 2000 and 2008. This rate is much higher than that described by the Cardiff researchers, where only 5% of patients underwent this surgical procedure,8 indicating that this procedure is still required in a significant number of patients in our referral centre. However, the discrepancy between patients from Nancy and Cardiff is probably due to the fact that our cohort is referral based, thus including subjects with more severe disease. At Nancy, small bowel resection, subtotal colectomy and left hemicolectomy were performed in 9, 3 and 3% of patients, respectively. These figures are broadly similar to those observed in the Cardiff cohort.8
The strengths of our study reside in the treatment regimen of the patients because two-thirds of them received azathioprine or anti-TNF treatment during follow-up. In addition, information on the duration of maintenance treatment (mesalamine, azathioprine and anti-TNF agents) was available for each patient in this referral centre cohort and only the patients' electronic charts were reviewed to conduct this study. As such, we were able to assess the impact of the duration of azathioprine and anti-TNF treatment on the requirement for surgery. It should be noted that some patients received a combined treatment with azathioprine and an anti-TNF agent. It is important to note that there are several limitations to our study. First, information on disease activity and mucosal healing was not available. In addition, other factors that may influence the course of the disease, such as smoking habits, were not systematically assessed in this cohort. The therapeutic management of Crohn's disease has evolved over the last 10 years, raising the possibility of a time-trend bias in the reported cohort. However, the propensity score adjustment and analysis of exposure times to treatments as well as time delay between Crohn's disease diagnosis and starting treatment with azathioprine or anti-TNF therapy aimed to reduce this bias. Finally, compliance to azathioprine treatment was not assessed.
In conclusion, we found that the cumulative risk of major abdominal surgery after 5 years of disease in patients with Crohn's disease was 26%. Non-complicated inflammatory disease behaviour and long-term treatment with anti-TNF and azathioprine were the only factors independently associated with a reduced risk for surgery. Our results indicate that long-term anti-TNF therapy is associated with a lower risk for surgery whereas azathioprine only modestly lowers this risk. A better understanding of the natural history of Crohn's disease in the era of biologicals together with the identification of risk factors for surgery may be used in future disease modification trials. In addition, the association between azathioprine and anti-TNF use and the requirement for intestinal surgery warrants further investigation.
References
Supplementary materials
Online only appendix
Files in this Data Supplement:
Footnotes
Funding LPB has received consulting fees from Abbott Laboratories and UCB Pharma; lecture fees from speaking at continuing medical education events from Centocor; and grant support from UCB Pharma.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.