Pre-operative use of anti-TNF-α agents and the risk of post-operative complications in patients with Crohn's disease – a nationwide cohort study

Summary

Background

A possible negative role of pre-operative use of antitumour necrosis factor-alpha (anti-TNF-α) agents on post-operative outcomes in Crohn's disease (CD) patients is still debated.

Aim

To examine the impact of pre-operative anti-TNF-α agents on post-operative outcomes 30 and 60 days after CD surgery in a nationwide Danish cohort. Outcomes were death, reoperation, anastomosis leakage, intra-abdominal abscess and bacteraemia.

Methods

We identified all patients having surgical procedures from 1 January 2000 to 31 December 2010 (n = 2293). Patients were classified according to use of anti-TNF-α agents within 12 weeks before surgery (exposed) or not (unexposed). Outcomes were obtained from nationwide registries and a bacteraemia registry. Sub-analyses were performed for bacteraemia and for impact of pre-operative timing of anti-TNF-α agents.

Results

Among surgical procedures for CD, 214 were exposed and 2079 were not. We found no increased relative risks of death or abscess drainage 30 or 60 days after follow-up. Among exposed, 7.5% had a reoperation within 30 days vs. 8.6% among unexposed, adjusted odds ratio (OR) = 0.92, 95% confidence interval (CI): 0.52–1.63. Among exposed, 3.8% had an anastomosis leakage within 30 days after surgery vs. 2.8% among unexposed, adjusted OR = 1.33, 95% CI: 0.59–3.02. No further cases of anastomosis leakages appeared within 60 days. Sub-analyses indicated no increased risk of bacteraemia after 30 days and no increased risks when anti-TNF-α agents were given ≤14 days before surgery.

Conclusion

We found no significantly increased relative risks of post-operative complications after use of anti-TNF-α agents either 12 weeks or ≤14 days before surgery for Crohn's disease.

Introduction

Previous studies have demonstrated the efficacy of antitumour necrosis factor-alpha (anti-TNF-α) agents (including infliximab and adalimumab) in patients with Crohn's disease (CD)[1-4]; and thus infliximab has shown efficacy in the short-term treatment of refractory luminal and fistulising CD,[4, 5] and also efficacy in a more long-term perspective.[1, 6-8] Since a number of patients will, however, require surgery after having anti-TNF-α agent therapy in the recent past, the surgeons will, therefore, be increasingly faced with abdominal surgery on CD patients with suppressed immune response in the pre-operative setting and consequently concerns about an increased risk of post-operative complications.

The effect of anti-TNF-α agent use on post-operative outcomes in CD patients is still debated because of diverging results[9]; existing studies have used different study approaches and time windows for anti-TNF-α agent exposure, and the numbers of included patients have been modest because anti-TNF-α agents are still relatively new medical therapies.

When examining post-operative complications after use of anti-TNF-α agents, a time window of 12 weeks before surgery is most often applied as the pharmacokinetics of infliximab after intravenous infusion is characterized by a half-life of 10 days, and in week 12 after infusion of infliximab (5 mg/kg) levels are no longer detectable.[7, 10, 11] The two largest studies on post-operative outcomes after anti-TNF-α agent therapy, given within 12 weeks before surgery, include 60 and 65 exposed CD patients.[12, 13] One study indicated an increased risk of post-operative sepsis, abscess and readmissions,[12] whereas the other did not indicate increased rate of complications.[13] Other studies, with fewer exposed CD patients, have not found increased risk of different post-operative outcomes after use of anti-TNF-α agents.[14-16] The association between anti-TNF-α agents and post-operative complications has also been examined in studies not separating the analyses to only CD patients,[17-19] while others included patients treated both pre- and post-operatively.[20-22]

In this nationwide Danish cohort study of patients having intestinal resection for CD, we aimed to examine the risk of 30 and 60 days post-operative complications after use of anti-TNF-α agents within 12 weeks before surgery.

Methods

Study population and setting

In Denmark (population approximately 5.5 million people), all citizens have free access to a tax-supported healthcare system, and its uniform organization allowed us to use a population-based study design. In addition, the availability of nationwide Danish registries made it possible to assess data from The Danish National Patient Registry (NPR) on (i) patients with CD, (ii) treatment with anti-TNF-α agents, (iii) post-operative complications (reoperation, anastomosis leakage, percutaneous abscess drainage and bacteraemia), The Danish Medicine Agency at Statistic Denmark on, (iv) therapeutic drug use before date of surgery for CD, The Central Personal Registration system on, (v) death, and regional bacteraemia registries on, (vi) cases of bacteraemia.

The unique civil registration number (assigned to all Danish citizens at birth from the Central Personal Registration system) is used in all Danish registries and ensures accurate linkage between registries.[23]

Infliximab was licensed by The Danish Medicine Agency in 1999 and was followed by other kinds of anti-TNF-α agents, but the first use of anti-TNF-α agents within 12 weeks before surgery for CD was registered in 2003 (start of study period). To be eligible for the study, patients had to have (i) a valid civil registration number, (ii) a discharge diagnosis of CD during the period from 1 January 1977 through 2010, (iii) an age of ≥15 year at time of surgery, (iv) a surgery for CD within the study period of 1 January 2003–31 December 2010, and (v) be available for follow-up in Denmark 30 and 60 days after surgery.

Identification of surgeries for CD in CD patients

The NPR contains records of all discharges from Danish hospitals since 1977 and all outpatient visits since 1994.[24] Information in the NPR includes patients’ civil registration numbers, hospital, department, dates of admission and discharge, procedures performed and up to 20 discharge diagnoses based on the International Classification of Diseases (ICD-8 before 1994 and ICD-10 from 1994 onward).

All eligible patients had a discharge history of CD from any hospital in Denmark since January 1, 1977 (ICD-8 codes: 56301; ICD-10 codes: DK50.x); and included in the study were patients with a bowel operation for CD at age ≥15 years in the period 1 January 2003–31 December 2010. Operations for CD were (according to the ‘Nordic Classification of Surgical Procedures’) ‘small bowel resections’ KJFB00 and KJFB01; ‘ileocecal resection and right hemicolectomy’ KJFB20,21,30,31; ‘Other resections of small bowel and colon’ KJFB33 and KJFB34; ‘Other colon resections’ KJFB40, 41, 43, 44, 46, 47, 50, 51, 60, 61, 63, 64; ‘Other bowel resections’ KJFB96 and KJFB97; ‘Colectomies’ KJFH00, 01, 10, 11, 20, 96; ‘Resections on rectum’ KJGB00, 01, 10, 11, 30, 31; ‘Intestinal plastic for small bowel stricture’ KJFA60, with a contemporary diagnosis of CD (DK50.x) in the same NPR record.

In a sub-analysis, we studied CD patients with a first time bowel resection in the study period; and thus, patients with a resection for CD before 1 January 2003 were excluded according to versions of the ‘Nordic Classification of Surgical Procedures: 43440, 43460, 43520, 43540, 43680, 43700, 43740, 43760, 43780, 43800, 43820, 43840, 43860, 43880, 44060, 44120, 44150, 44160, 44790, 44900, 44920, 44940, 44960, 44980, 45020, 45060, 45080, 45100, 45120, 45200, 45240, 45320, 45480, 45690, 45840, 45860, 45880, 46290 or one/more of the above mentioned codes within KJFB, KJFH, KJGB, KJFA.

Pre-operative therapeutic drug use

Anti-TNF-α agent therapy

In Denmark, anti-TNF-α agents are administered only in hospitalized patients or in hospital-based outpatient settings in accordance with the National Board of Health. Data on anti-TNF-α agent therapy were obtained from the NPR, where date of procedures for anti-TNF-α agent therapy is coded by BOHJ18 [or specific codes for infliximab (BOHJ18A1), adalimumab (BOHJ18A3) or certolizumab pegol (BOHJ18A5), or procedures coded with additional relevant Anatomical Therapeutical Chemical classification (ATC) codes (infliximab L04AA12 or L04AB02, adalimumab L04AA17 or L04AB04, or certolizumab pegrol L04AB05)].

Other therapeutic drug treatment

Data on other therapeutic drug treatment, besides anti-TNF-α agents, were obtained by outpatient drug prescriptions from the nationwide prescription database maintained by the Danish Medicine Agency. Since 1 January 1995, data on prescriptions have been available from all pharmacies in Denmark due to computerized accounting systems, which send key data on outpatient drug prescriptions directly to the database. Thus, we were able to obtain the prescription history of each CD patient, and could classify the time of drug prescriptions before operation in the study period. We identified all prescriptions (including prescription date) with the following ATC-codes: L04A X01 (azathioprine); L01B B02 (mercaptopurine); L01BA01 (methotrexate); L04AD01 (ciclosporin); H02A B06 (prednisolone); H02A B07 (prednisone); A07E A01 (local prednisolone); A07E A02 (local hydrocortisone); and A07E A06 (local budesonide).

Outcome data

Data on reoperations were obtained from the NPR; we identified (i) all kinds of operations according to ‘small bowel resections’ KJFB00 and KJFB01; ‘ileocecal resection and right hemicolectomy’ KJFB20, 21, 30, 31; ‘Other resections of small bowel and colon’ KJFB33 and KJFB34; ‘Other colon resections’ KJFB40, 41, 43, 44, 46, 47, 50, 51, 60, 61, 63, 64; ‘Other bowel resections’ KJFB96 and KJFB97; ‘Colectomies’ KJFH00, 01, 10, 11, 20, 96; ‘Resections on rectum’ KJGB00, 01, 10, 11, 30, 31; ‘Reoperations after gastrointestinal operation’ KJWA, KJWB, KJWC, KJWD, KJWE, KJWF, KJWW; ‘suture of small bowel’ KJFA70, ‘suture of colon’ KJFA80 (ii) specifically the subgroup of anastomosis leakage, procedure code KJWF, and (iii) percutaneous abscess drainage, UXUD (ultrasound examination of the abdomen) together with KTJA40 (percutaneous needle biopsy of the intestine). Data on death were obtained from the Central Personal Registration system.

We linked our data to a bacteraemia registry retrieved from three departments of clinical microbiology at Danish hospitals (data from 14 hospitals). This registry captures well-defined geographical areas (The North Region Denmark, Hvidovre and Herlev Hospitals in the Capital Region of Denmark) with a total background population of approximately 1.75 million, covering the period 1 January 2000–31 December 2011.[25, 26] We omitted positive blood cultures considered to be contamination and defined bacteraemia if the isolated pathogen was given aetiological significance based on clinical and microbiological criteria.[27, 28] We evaluated all outcomes occurring within 30 and 60 days after surgery for CD.

Data on covariates

From the NPR, we obtained data on patient gender, age at time of surgery for CD in the study period, duration of CD at the time of surgery (calculated from the date of the first time the CD diagnosis appears in the NPR until the date of surgery), calendar year of surgery for CD, number of inpatient days at hospital before surgery and comorbid diseases. Comorbidity index score, the Charlson Index,[29] was assessed for each patient and covers 19 major disease categories weighted according to their prognostic impact on patient survival. The Charlson Index has been adapted for use with hospital discharge registry data, and we computed the index based on diagnoses recorded during all previous hospitalizations since 1977. Three index levels were defined to capture increasing degrees of comorbidity: no comorbidity (Charlson Index 0), intermediate comorbidity level (Charlson Index 1–2), and high comorbidity level (Charlson Index >2). Data on prescriptions for steroids (prednisolone, prednisone, local prednisolone, local hydrocortisone, or local budesonide) before surgery for CD were obtained from the nationwide prescription database.

Statistical analysis

Classification of exposed and unexposed

CD patients might have more than one surgical procedure for CD in the study period, and each new index CD surgery was included as an independent event in our main analyses. For a surgery to count as a new index surgery, the patient was not allowed to have any CD-related surgery within a time interval of 1 year, prior to the index surgery. Thus, operations occurring within the first year from index surgical procedure were not included as an index procedure, but may be included as a reoperation.

The associations between pre-operative anti-TNF-α agent therapy and adverse post-operative outcomes (reoperations, anastomosis leakage, intraabdominal abscess, bacteraemia, and death) were studied in sub-cohorts of patients undergoing operation for CD in the study period, classified according to therapeutic drug treatment pre-operatively:

1. Exposed cohort constituted all CD operations where the patients had been treated at least once with anti-TNF-α agents within 12 weeks before surgery. Patients were allowed to be treated with other kinds of drugs for CD (for instance, steroids, and/or azathioprine/mercaptopurine).
2. Unexposed cohort 1 constituted all CD operations where the patients were not treated with anti-TNF-α agents within 12 weeks before surgery, but allowing all kinds of other treatments (for instance steroids, and/or azathioprine/mercaptopurine).
3. Unexposed cohort 2 constituted all CD operations where the patients were not treated with anti-TNF-α agents within 12 weeks before surgery, but as recorded by the national prescription database, the patients had at least one filled prescription for steroids, azathioprine/mercaptopurine, methotrexate, or ciclosporin within 12 weeks before surgery. Thus, this group constitutes a subgroup of the unexposed cohort 1. The purpose of designing unexposed cohort 2 was to give the possibility to counter in a potential impact of disease activity.

Adverse post-operative complications

Adverse post-operative complications (death, reoperation, anastomosis leakage, and intra-abdominal abscess) in the exposed cohort were compared with outcomes in unexposed cohort 1 and unexposed cohort 2.

Crude and adjusted analyses

We constructed contingency tables for the main study variables, and computed the relative risk estimates [prevalence odds ratio, (OR)], with 95% confidence intervals (95% CI). We used logistic regression analyses to compute relative risk estimates for adverse post-operative outcomes (reoperations, anastomosis leakage, intra-abdominal abscess, bacteraemia, and death within 30 days and 60 days after CD operation) associated with anti-TNF-α agent therapy pre-operatively, adjusted for potential confounders. Adjustment was made for age (years: 15–25 as reference, 26–55, >55), gender (females as reference), comorbidity (Charlson Index 0 as reference, Charlson Index 1–2, Charlson Index >2), calendar period (2003–2004 as reference, 2005–2006, 2007–2008, 2009–2010), and duration of CD (<5 years as reference, ≥5 years), use of steroids within 4 weeks before surgery (no as reference, yes), number of inpatient days at hospital within 4 weeks before surgery (<2 weeks as reference, ≥2 weeks) in a logistic regression model. Confounders were included in the models based on the ‘change-in-estimate’ method. When estimating the risk of reoperations, anastomosis leakage, intra-abdominal abscess, and bacteraemia, only patients available for full follow-up were included.

Sub-analyses

In sub-analysis 1, we restricted the analyses to first time operation for CD in the study period for each individual; and the risk of death, reoperation, anastomosis leakage, and intra-abdominal abscess was estimated among exposed patients relative to outcomes in unexposed cohort 1.

In sub-analysis 2, we changed the exposure for anti-TNF-α agents into three time-intervals according to the time span from last use of anti-TNF-α agent until time of surgery (≤14 days, 15–30 days, and 31–84 days). In this analysis the risk of death, reoperation, anastomosis leakage, and intra-abdominal abscess was estimated among exposed operations relative to outcomes in unexposed cohort 1.

In sub-analysis 3, we estimated the risk of bacteraemia in a sub-cohort of the study population with an operation for CD in The North Region Denmark, Hvidovre and Herlev Hospitals in the Capital Region of Denmark, relative to the outcomes in unexposed cohort 1. All analyses were performed using Stata 12 software (StataCorp LP, College Station, TX, USA). The study was approved by the Danish Data Protection Agency (j.nr. 2010-41-5660).

Results

During the study period, 2293 surgical procedures for CD were performed. The exposed cohort comprised 214 (9.3%) operations where patients were treated with anti-TNF-α agents within 12 weeks before surgery, and in the remaining operations, no anti-TNF-α agent was given within 12 weeks before surgery [unexposed cohort 1 = 2079 (90.7%)]. Among those treated with anti-TNF-α agents within 12 weeks before surgery, 53.3% (114/214) were treated in 2009 and 2010; 24.3% (52/214) in 2007 and 2008; and 22.4% (48/214) from 2003 through 2006. During the study period, infliximab counted for most of the anti-TNF-α agents given before surgery with 57.0% (122/214), adalimumab with 27.1% (58/214) and other anti-TNF-α agents with 15.9% (34/214).

Characteristics of study cohorts

The characteristics of study cohorts are given in Table 1. The most commonly used procedure among exposed, unexposed cohort 1 and unexposed cohort 2 was ‘Ileocecal resection and right hemicolectomy’; 37.9%, 46.0% and 49.5% respectively. The median age among exposed, unexposed cohort 1 and unexposed cohort 2 was 33, 41 and 38 years, respectively. In the exposed cohort, the duration of CD was less than 5 years for the majority of patients, while most patients in unexposed cohorts had disease duration of ≥5 years. For all cohorts, the vast majority of patients had (i) no comorbid diseases, (ii) inpatient days at hospital of <14 days within 4 weeks before surgery, (iii) and no steroid prescriptions within 4 weeks before surgery.

Table 1. Characteristics of study cohorts of Crohn's disease (CD) operations from 1 January 2003 through 2010

	Exposed cohort, N = 214b (receiving anti-TNF-α agents within 12 weeks before surgery)	Unexposed cohort 1, N = 2079c (no anti-TNF-α agents within 12 weeks before surgery)	P-valuea	Unexposed cohort 2, N = 887d (no anti-TNF-α agents within 12 weeks before surgery but other kind of immunosuppressant's)	P-valuea
^a P-values: by Chi-square testing exposed cohort vs. unexposed cohort 1 and exposed cohort vs. unexposed cohort 2. ^b Representing 212 CD patients. ^c Representing 1951 CD patients. ^d Unexposed cohort 2 is a sub-group of unexposed cohort 1 and representing 855 CD patients. ^e First (main) registered procedure code for CD surgery. ^f No comorbidity (Charlson Index 0), intermediate comorbidity level (Charlson Index 1–2), high comorbidity level (Charlson Index >2).
Type of CD surgery,e number (%)
Ileocecal resection and right hemicolectomy KJFB20, 21, 30, 31	81 (37.9%)	956 (46.0%)	<0.01	439 (49.5%)	<0.01
Colectomies KJFH00, 01, 10, 11, 20, 96	51 (23.8%)	228 (11.0%)		116 (13.1%)
Small bowel resections KJFB00, 01	32 (15.0%)	427 (20.5%)		157 (17.7%)
Other resections of small bowel and colon  KJFB33, 34	19 (8.9%)	182 (8.8%)		69 (7.8%)
Other colon resections  KJFB40, 41, 43, 44, 46, 47, 50, 51, 60, 61, 63, 64	18 (8.4%)	184 (8.9%)		79 (8.9%)
Resections on rectum  KJGB00, 01, 10, 11, 30, 31	13 (6.1%)	102 (4.9%)		27 (3.0%)
Gender
Females, number (%)	117 (54.7%)	1227 (59.0%)	0.22	518 (58.4%)	0.32
Males, number (%)	97 (45.3%)	852 (41.0%)		369 (41.6%)
Age at time CD surgery
Median (Min–max), in years	33 (16–86)	41 (15–90)	<0.01	38 (15–85)	<0.01
Age 15–25, number (%)	49 (22.9%)	327 (15.7%)		197 (22.2%)
Age 26–55, number (%)	142 (66.4%)	1208 (58.1%)		512 (57.7%)
Age >55, number (%)	23 (10.7%)	544 (26.2%)		178 (20.1%)
Duration of CD at time of surgery
<5 years with CD, number (%)	131 (61.2%)	975 (46.9%)	<0.01	419 (47.2%)	<0.01
≥5 years with CD, number (%)	83 (38.8%)	1,104 (53.1%)		468 (52.8%)
Comorbidity level at time of surgeryf
No comorbidity, number (%)	172 (80.4%)	1,537 (73.9%)	0.01	687 (77.5%)	0.09
Intermediate comorbidity level, number (%)	40 (18.7%)	434 (20.9%)		165 (18.6%)
High comorbidity level, number (%)	2 (0.9%)	108 (5.2%)		35 (3.9%)
Inpatient days at hospital within 4 weeks before surgery
<14 days, number (%)	195 (91.1%)	1853 (89.1%)	0.37	796 (89.7%)	0.61
≥14 days, number (%)	19 (8.9%)	226 (10.9%)		91 (10.3%)
Steroid prescriptions within 4 weeks before surgery
No, number (%)	195 (91.1%)	1785 (85.9%)	0.03	593 (66.9%)	<0.01
Yes, number (%)	19 (8.9%)	294 (14.1%)		294 (33.1%)

Post-operative outcomes

Table 2 gives the numbers of post-operative complications, and crude and adjusted risk estimates with unexposed cohort 1 as the reference. Among the 2293 surgical procedures for CD performed, there were 55 deaths within 30 days after surgery (2.4%). There was one death among exposed and 54 among unexposed during 30 days of follow up, adjusted OR = 0.38, 95% CI: 0.05–2.91. No further cases of death were found among exposed up to 60 days after surgery. In the exposed cohort, 7.5% had a reoperation within 30 days vs. 8.6% among unexposed, adjusted OR = 0.92, 95% CI: 0.52–1.63. Also, no increased risk of reoperation was found within 60 days after surgery. In the exposed cohort, 3.8% had an anastomosis leakage within 30 days after surgery vs. 2.8% among unexposed, adjusted OR = 1.33, 95% CI: 0.59–3.02. No further cases of anastomosis leakages were found among exposed up to 60 days after surgery. No cases of abscess drainage were found among exposed during follow-up.

Table 2. Post-operative outcomes in study cohorts of patients with Crohn's disease (CD) with a surgery from 1 January 2003 through 2010

	Exposed cohort, N = 214a (anti-TNF-α agents within 12 weeks before operation)	Unexposed cohort 1, N = 2079b (no anti-TNF-α agents within 12 weeks before operation)	Crude OR (95% confidence interval)	ORc (95% confidence interval)
^a Representing 212 CD patients. The one patient who died within the 30 (and 60)-day time window was excluded in the analyses of reoperation, anastomosis leakage and abscess drainage within 30 (and 60) days. ^b Representing 1951 CD patients. Fifty four patients who died within the 30-day time window were excluded in the analyses of reoperation, anastomosis leakage and abscess drainage within 30 days, and 62 patients who died within the 60-day time window were excluded in the analyses of reoperation, anastomosis leakage and abscess drainage within 60 days. ^c Adjusted for age (years: 15–25, 26–55, >55), gender, comorbidity (no comorbidity, comorbidity level 1, comorbidity level 2), calendar period (2003–2004, 2005–2006, 2007–2008, 2009–2010), duration of CD (<5 years, ≥5 years), steroid prescriptions (no/yes) within 4 weeks before surgery, and number of inpatient days at hospital within 4 weeks before surgery (<2 weeks, ≥2 weeks) in a logistic regression model.
Post-operative outcomes
Death within 30 days after CD surgery
Yes, number (%)	1 (0.5%)	54 (2.6%)	0.18 (0.02–1.28)	0.38 (0.05–2.91)
No, number (%)	213 (99.5%)	2025 (97.4%)
Death within 60 days after CD surgery
Yes, number (%)	1 (0.5%)	62 (3.0%)	0.15 (0.02–1.11)	0.34 (0.04–2.62)
No, number (%)	213 (99.5%)	2017 (97.0%)
Reoperation within 30 days after CD surgery
Yes, number (%)	16 (7.5%)	175 (8.6%)	0.86 (0.50–1.46)	0.92 (0.52–1.63)
No, number (%)	197 (92.5%)	1850 (91.4%)
Reoperation within 60 days after CD surgery
Yes, number (%)	18 (8.5%)	181 (9.0%)	0.94 (0.56–1.56)	1.03 (0.60–1.77)
No, number (%)	195 (91.5%)	1836 (91.0%)
Anastomosis leakage within 30 days after CD surgery
Yes, number (%)	8 (3.8%)	56 (2.8%)	1.37 (0.64–2.92)	1.33 (0.59–3.02)
No, number (%)	205 (96.2%)	1969 (97.2%)
Anastomosis leakage within 60 days after CD surgery
Yes, number (%)	8 (3.8%)	57 (2.8%)	1.34 (0.63–2.85)	1.27 (0.57–2.85)
No, number (%)	205 (96.2%)	1960 (97.2%)
Percutaneous abscess drainage within 30 days after CD surgery
Yes, number (%)	0 (0.0%)	3 (0.1%)	–	–
No, number (%)	213 (100.0%)	2022 (99.9%)
Percutaneous abscess drainage within 60 days after CD surgery
Yes, number (%)	0 (0.0%)	3 (0.1%)	–	–
No, number (%)	213 (100.0%)	2014 (99.9%)

Results of post-operative outcomes, with unexposed cohort 2 as reference, were virtually unchanged; i.e. no increased adjusted risks of death, reoperation, or abscess drainage (data not shown); and the adjusted OR for anastomosis leakages within 30 days after surgery was 1.40, 95% CI: 0.56–3.49.

Length of hospitalization after surgery (≥8 days vs. <8 days) was also examined. In the exposed cohort, 52.6% (112/213) had inpatient days of ≥8 days after surgery with similar results among unexposed cohort 1, 50.8% (1029/2025).

Post-operative outcomes after first-time surgery

The results of sub-analysis 1 (restricted to first time surgery for CD for each individual) are detailed in Table S1. Among those treated with anti-TNF-α agents within 12 weeks before surgery, we found no cases of death or abscess drainage during 30 days or 60 of follow-up. The adjusted ORs for reoperation were close to unity, and the adjusted OR for anastomosis leakage 30 days after surgery was 1.21, 95% CI: 0.46–3.16.

Post-operative outcomes and pre-operative timing of anti-TNF-α agents

In sub-analysis 2, stratified by three time periods for last anti-TNF-α agents prior to surgery, we found no increased risk of post-operative outcomes among those treated with anti-TNF-α agents ≤14 days before surgery compared with those without TNF-α agents (Table 3). For those treated within 31–84 days before surgery, the relative risk estimate for anastomosis leakage within 30 days after surgery was OR = 1.55, 95% CI: 0.59–4.07. The relative risks of abscess drainage were not estimated (no cases among exposed).

Table 3. Relative risk of post-operative outcomes, with 95% confidence interval (CI), stratified by time periods for last anti-TNF-α agents prior to surgery

	Exposed cohort, N = 43(anti-TNF-α agents within ≤14 days before surgery)	Exposed cohort, N = 55 (anti-TNF-α agents 15–30 daysbefore surgery)	Exposed cohort, N = 116 (anti-TNF-α agents 31–84 days before surgery)	Unexposed cohort1, N = 2079 (no anti-TNF-α agents within 12 weeks before surgery)
^a Adjusted relative risk of post-operative outcomes, relative to unexposed cohort 1, adjusted for age (years: 15–25, 26–55, >55), gender, comorbidity (no comorbidity, comorbidity level 1, comorbidity level 2), calendar period (2003–2004, 2005–2006, 2007–2008, 2009–2010), duration of CD (<5 years, ≥5 years), steroid prescriptions (no/yes) within 4 weeks before surgery, and number of inpatient days at hospital within 4 weeks before surgery (<2 weeks, ≥2 weeks) in a logistic regression model.
Post-operative outcomes
Death within 30 days after CD surgery
Yes, number (%)	0 (0.0%)	0 (0.0%)	1 (0.9%)	54 (2.6%)
No, number (%)	43 (100.0%)	55 (100.0%)	115 (99.1%)	2025 (97.4%)
Crude OR (95% CI)			0.33 (0.04–2.38)	Reference
ORa (95% CI)			0.64 (0.08–5.43)
Death within 60 days after CD surgery
Yes, number (%)	0 (0.0%)	0 (0.0%)	1 (0.9%)	62 (3.0%)
No, number (%)	43 (100.0%)	55 (100.0%)	115 (99.1%)	2017 (97.0%)
Crude OR (95% CI)			0.28 (0.04–2.06)	Reference
ORa(95% CI)			0.56 (0.07–4.81)
Reoperation within 30 days after CD surgery
Yes, number (%)	1 (2.3%)	6 (10.9%)	9 (7.8%)	175 (8.6%)
No, number (%)	42 (97.7%)	49 (89.1%)	106 (92.2%)	1850 (91.4%)
Crude OR (95% CI)	0.25 (0.03–1.84)	1.29 (0.55–3.07)	0.90 (0.45–1.80)	Reference
ORa (95% CI)	0.26 (0.04–1.97)	1.42 (0.59–3.44)	0.95 (0.46–1.97)
Reoperation within 60 days after CD surgery
Yes, number (%)	1 (2.3%)	6 (10.9%)	11 (9.6%)	181 (9.0%)
No, number (%)	42 (97.7%)	49 (89.1%)	104 (90.4%)	1836 (91.0%)
Crude OR (95% CI)	0.24 (0.03–1.77)	1.24 (0.52–2.94)	1.07 (0.57–2.04)	Reference
ORa (95% CI)	0.26 (0.04–1.97)	1.37 (0.57–3.27)	1.17 (0.60–2.28)
Anastomosis leakage within 30 days after CD surgery
Yes, number (%)	1 (2.3%)	2 (3.6%)	5 (4.3%)	56 (2.8%)
No, number (%)	42 (97.7%)	53 (96.4%)	110 (95.7%)	1969 (97.2%)
Crude OR (95% CI)	0.84 (0.11–6.20)	1.33 (0.32–5.59)	1.60 (0.63–4.07)	Reference
ORa (95% CI)	0.80 (0.10–6.23)	1.31 (0.30–5.69)	1.55 (0.59–4.07)
Anastomosis leakage within 60 days after CD surgery
Yes, number (%)	1 (2.3%)	2 (3.6%)	5 (4.3%)	57 (2.8%)
No, number (%)	42 (97.7%)	53 (96.4%)	110 (95.7%)	1960 (97.2%)
Crude OR (95% CI)	0.82 (0.11–6.06)	1.30 (0.31–5.46)	1.56 (0.61–3.98)	Reference
ORa (95% CI)	0.78 (0.10–6.03)	1.23 (0.28–5.32)	1.48 (0.57–3.85)

Analyses of bacteraemia

In sub-analysis 3, we estimated the relative risk of bacteraemia within 30 days after surgery for CD; among those treated with anti-TNF-α agents within 12 weeks before surgery, 1/68 (1.5%) had a bacteraemia, compared to 11/591 (1.9%) among unexposed, yielding an OR = 0.79, 95% CI: 0.10–6.22. We found three cases of bacteremia within 60 days after surgery among exposed (4.4%), compared to 1.9% among unexposed, yielding an OR = 2.43, 95% CI: 0.66–8.94.

Discussion

This nationwide study showed no increased relative risks of death, reoperation, or abscess drainage 30 or 60 days after surgery among CD patients treated with anti-TNF-α agents within 12 weeks before surgery. The relative risk of anastomosis leakage was slightly, but not significantly, increased. Sub-analyses indicated no increased relative risk of complications when anti-TNF-α agents were given ≤14 days before the CD surgery, and no increased risk of bacteraemia within 30 days after surgery for CD.

This study has several strengths: (i) it is population-based due to a unique availability of nationwide registries in Denmark. The NPR records more than 99% of all hospital discharges for somatic diseases.[24] and thus we had access to the obligatory registration from hospitals since 1977 and all outpatient visits since 1994. Hence, we used the NPR to identify CD patients based on hospital discharge diagnoses and specific codes for surgical procedures for CD, (ii) the completeness and the validity of CD diagnoses in the NPR are of high quality. The completeness of diagnoses of CD has been examined in a Danish study using the pathology system as a reference standard[30] – showing that among all patients, with a confirmed diagnosis of CD, 94% were included in the NPR.[30] The overall validity of diagnosis of CD in the NPR was 97%,[30] (iii) there is no major misclassification between CD and ulcerative colitis patients. Thus, patients were identified with a discharge history of CD in the NPR since 1977, were included only if coded as having specific procedure codes for CD surgery, and patients had to have a discharge diagnosis of CD in the same record as the code for CD surgery occurred, (iv) data on treatment with anti-TNF-α agents were based on procedure codes in the NPR, which are usually of high quality.[31-34] Some misclassification of anti-TNF-α agents (the exposure) cannot be ruled out if the coding in the NPR has not been entirely accurate. A possible misclassification of anti-TNF-α agent therapy is unrelated to the outcomes examined, and will bias our risk estimates towards the null hypothesis, (v) the access to high quality data on outpatient drug prescriptions from the nationwide prescription database making it possible to identify unexposed cohort 2. It is a strength that this group is based on prescriptions and not on recall, as drug exposure based on self-reported use may lead to recall bias or under-ascertainment; and data from the prescription database are of high quality as a result of direct computerized transfer of information when a prescribed drug is dispensed at a pharmacy, (vi) outcome data were obtained independently of the hypotheses investigated, and independently of exposure measurement preventing differential misclassification of the outcome measurements; and procedure codes in the NPR have positive predictive values of 94–100%,[31-34] and (vii) available information on several confounders.

In an observational, register-based, study like this, it impossible to take into accounts all potential confounders. Thus, the study also had limitations as we had no opportunity to assess possible impacts of clinical patient details, severity of disease at the time of surgery, factors related to surgery, timing of operation, acute or elective surgery, in-hospital drug use, and factors related to post-operative care. However, an impact of such possible confounders is unlikely as we have no reason to believe that factors such as experience of the surgeons and factors related to post-operative care differ between exposed and unexposed. Moreover, this study did not examine a possible cumulative effect of anti-TNF-α agent therapy given prior to the 12-week period before surgery. It is a special challenge to handle a possible impact of disease activity; the main approach was to design the unexposed cohort 2 to counter possible arguments that patients in the exposed cohort had higher degree of disease activity than those in the unexposed cohort 1; i.e. if we had found increased risks of post-operative outcomes, one could have argued that disease activity, and not anti-TNF-α agent therapy, was responsible for adverse outcomes. With our results of no increased risks, we did not, however, expect that the analyses using unexposed cohort 2 changed our results – and they did not. Thus, our results indicate that disease activity had no major impact on our post-operative outcomes or perhaps even more likely, disease severity was equally distributed between exposed and unexposed as all patients were so severely diseased that they underwent surgery. Another approach to handle the disease activity issue was to include surrogate measure of disease severity into the logistic regression models (use of steroids and number of inpatient days at hospital before surgery).

This study is the largest to date on the risk of post-operative outcomes after pre-operative exposure to anti-TNF-α agents in patients only with CD. In our view, it is important to examine the association between pre-operative exposure to anti-TNF-α agents and the risk of post-operative outcomes according to each specific disease (CD or ulcerative colitis) – in respect for differences according to pathophysiology, clinical characteristics, cause of disease and treatment regimes. We have earlier examined the risk of post-operative outcomes in patients with ulcerative colitis in a similar setting[35]; and found differences according to specific types of surgical procedures performed for CD and ulcerative colitis, but another apparent difference is that the proportion of patients having a reoperation within 30 days after surgery was considerably higher among ulcerative colitis patients (23%)[35] than among CD patients in this study (8%).

Existing data on the safety of pre-operative anti-TNF-α agents according to post-operative outcomes in CD patients are still limited and the results are diverging. Until now, the two largest studies that included solely CD patients are from Appau et al. examining 60 patients exposed to infliximab within 3 months before surgery[12] and from Canedo et al. examining 65 patients.[13]Appau et al. concluded that infliximab used pre-operatively was associated with increased 30-day post-operative sepsis (adjusted OR = 2.62, 95% CI: 1.12–6.13), abscess (OR = 5.78, 95% CI: 1.69–19.7), and readmissions (adjusted OR = 2.33, 95% CI: 1.02–5.33).[12] Regarding the increased risk of abscess and sepsis, our data could not confirm these results – on the contrary – among our 214 operations for CD, we found no treatments of abscesses during 30 or 60 days of follow-up, and in our sub-analysis of bacteraemia, we found no increased risk after 30 days. We did find an OR = 2.43, 95% CI: 0.66–8.94, of bacteraemia after 60 days of follow-up, but the estimate lacks statistical precision. The differences between the findings in the study by Appau et al. and our study may be explained by the definition of outcome variables. Canedo et al. concluded that there were no differences in the rate of post-operative complications among the groups of patients undergoing surgery for CD pre-treated with infliximab or other immunosuppressive drugs, but as it is difficult to determine the time window of follow-up period and the study does not provide risk estimates, the results are difficult to compare with our results.[13] Other studies on CD patients, including fewer exposed patients than the studies by Appau et al. and Canedo et al., mainly concluded that pre-operative infliximab was not associated with post-operative complications.[14-16] When it comes to impact of the timing of treatment with the anti-TNF-α agents before surgery, there have been no studies solely on CD patients. One large study by Waterman et al. included 195 patients with exposure to biologics, but did not separate analyses on UC and CD patients.[19] That study examined for the first time possible complications of biologics according to time intervals between the last dose of biologics and surgery, and found that exposure within 14 days before surgery was not associated with increased rate of complications compared to exposure 15–30 days and 31–180 days before surgery.[19] We estimated the risk of complications relative to those not treated with anti-TNF-α agents and also found no increased risk of post-operative outcomes among those treated with anti-TNF-α agents ≤14 days before surgery. Moreover, in our study, we found no significantly increased relative risks of post-operative complications within the other time intervals.

This is the first nationwide study on short-term post-operative adverse outcomes in CD patients after pre-operative use of anti-TNF-α agents, and the study provides reassuring results. Furthermore, sub-analyses indicate no increased risk of adverse post-operative adverse outcomes when anti-TNF-α agents are given close to the time of surgery (within 14 days before surgery). As anti-TNF-α agent therapy has been introduced in recent years, it is of course especially important to monitor a possible negative effect on post-operative outcomes; thus, these first reassuring results from nationwide data should be confirmed in other settings.

Authorship

Guarantor of the article: Bente Mertz Nørgård is the guarantor of the study and accepts full responsibility for the work, the conduct of the study, the access to the data, and controlled the decision to publish.

Author contributions: Bente Mertz Nørgård has contributed to the conception and design, the analysis and interpretation of data, and drafting the article. Jan Nielsen has contributed to the design, the analysis, the interpretation of data, and revising the article critically for important intellectual content. Kim Oren Gradel has contributed to the analysis and interpretation of data, and revising the article critically for important intellectual content. Niels Qvist, Ove Bernhard Schaffalitzky de Muckadell and Jens Kjeldsen have contributed to the conception and design, the interpretation of data, and revising the article critically for important intellectual content. All authors have approved the final version of the manuscript.

Acknowledgement

Declaration of personal interests: None.

Declaration of funding interests: The study was funded in part by the Danish Colitis-Crohn Society (Colitis-Crohn Foreningen).