OUP user menu

Journal of Crohn's and Colitis: 10 (10)


Laurence J. Egan, Ireland

Associate Editors

Shomron Ben-Horin, IsraelSilvio Danese, ItalyPeter Lakatos, HungaryMiles Parkes, UKJesús Rivera-Nieves, USABritta Siegmund, GermanyGijs van den Brink, NLSéverine Vermeire, Belgium


Published on behalf of

Infliximab and complications after colectomy in patients with ulcerative colitis

David Bregnbak, Christian Mortensen, Flemming Bendtsen
DOI: http://dx.doi.org/10.1016/j.crohns.2011.08.014 281-286 First published online: 1 April 2012


Background: Infliximab treatment may increase the risk of subsequent postoperative complications in patients with ulcerative colitis. The main purpose of the present study therefore was to assess postoperative complications in patients who have undergone colectomy for ulcerative colitis with and without previous infliximab treatment.

Methods: Through a database search within a five-year period ulcerative colitis patients in a single highly specialized department, who had undergone colectomy, were identified. In total 71 ulcerative colitis patients were identified and analyzed according to pretreatment with infliximab or not. Twenty patients who had received infliximab within 12 weeks prior to colectomy were compared to 51 patients on standard treatment. Data on patient background, concomitant medication, endoscopic and the laboratory results, clinical activity, and complications within 30 days after colectomy were recorded.

Results: At primary surgery, patient groups were similar with respect to distribution on gender, age, smoking behavior and concomitant medication. There were significant differences in partial Mayo-scores (7,95 (IFX) vs. 7,64, P = 0.032); preoperative CRP-levels (42,72 (IFX) vs. 63,2, P = 0.05); postoperative hospitalization time (10,9 (IFX) vs. 11,3 days, P = 0.039); and in number of patients who underwent elective surgery (10% vs. 37,3%, P = 0.015). There was no short-term mortality in either group and no significant difference in terms of postoperative complications between patients treated with IFX or not. However, the number of postoperative infectious complications was increased in corticosteroid-treated patients irrespective of IFX or not (45,8% in CS group vs. 13,0%, P = 0.028).

Conclusions: The use of infliximab does not seem to associate with an increased risk of short-term postoperative complications in ulcerative colitis.

  • Corticosteroids
  • Inflammatory bowel disease
  • Surgery
  • Biological therapy

1 Introduction

Ulcerative colitis (UC) is an inflammatory bowel disease (IBD) of unknown etiology that causes chronic inflammation of the digestive tract. The disease is characterized by being erosive or ulcerative with continuous inflammation of the colon mucosa. There is no known cure for UC, but treatment of UC is to a large extent pharmacological, using agents such as 5-aminosalicylic acid (5-ASA), corticosteroids (CS) and immunomodulatory drugs as well as biological therapy. Biological therapy was introduced approximately 10 years ago as an alternative in otherwise pharmacologically resistant patients with the aim of preventing surgery 1.

Traditionally, acute attacks of UC have been treated with high doses of CS. Despite a relatively high response rates to pharmacological treatment, the condition is still associated with a rather high risk of surgery, due to lack of response in up to 30% of patients 2. The surgical procedure is lifesaving and curative, but introduces a significant risk of reduced quality of life; physical; and psychological morbidity.

Today infliximab (IFX) has become an established option for the medical management of UC in both induction 35 and maintenance therapy 1 for moderate-to-severe ulcerative colitis. IFX is a monoclonal antibody that recognizes, binds to and blocks the action of tumor necrosis factor-alpha (TNF-α). TNF-α is a potent pro-inflammatory cytokine found in elevated concentrations in inflammatory tissues. By blocking this cytokine the inflammatory activity is expected to be attenuated 6.

The earliest studies which analyzed postoperative infectious complications in inflammatory bowel disease patients treated with IFX, had their focus on patients with Crohn's Disease (CD). They did not show any increased risk of postoperative complications 79. Even though UC has certain similarities with CD, published studies including UC-patients have been less conclusive 1016.

The aim of this study was to analyze a Danish cohort of patients from a single UC specialized unit undergoing surgery in order to assess whether IFX treatment in the preoperative period results in an increased rate of short-term postoperative complications.

2 Material and methods

2.1 Recruitment

Retrospective data was gathered from a study population treated at Hvidovre Hospital (HH) in the period of January 1, 2005 to September 23, 2010. The cohort of patients studied was initially found through a database query with the Danish National Health Register. This database contains all contacts and hospitalizations within the Danish healthcare system, including primary diagnosis and surgical procedures. A search with the following criteria was pebbormed:

  • Patients with a diagnosis of certain or probable UC based on clinical, endoscopic and histological criteria.

  • and who underwent colectomy during the study period.

The search ended up with a total of 87 candidates. These 87 candidate patient records were systematically scrutinized. A total of 16 patients were excluded. Two of these patients underwent colectomy before the time period of our study; three patient records could not be retrieved; five patients were registered under a diagnosis of ulcerative colitis in the national database, and were during file review demonstrated to fulfill criteria of CD, and were excluded accordingly; and six patients who underwent colectomy for malignancy or dysplasia were excluded because surgery for malignancy may have a different profile risk of postoperative complications, and patients are less likely to receive infliximab in the setting of malignancy, possibly confounding the result. This left a total consecutive cohort of 71 patients. The patients were grouped into 2 separate cohorts based on use or no use of IFX within 90 days prior to primary surgery. Among the 71 patients analyzed, 20 patients received IFX prior to surgery. We chose a cut-off period of 90 days, as it is believed to eliminate the influence of IFX on postoperative outcome in the control group 17. Records were abstracted for clinical characteristics including gender; age at diagnosis; age at primary surgery; duration of disease before surgery; smoking behavior; medical therapy within the last 90 days prior to surgery; biochemical markers (hemoglobin (Hgb); CRP; leukocytes; albumin; creatinine); date of admission; date of surgery; date of discharge; postoperative complications; and co-morbidities. Activity was assessed by collecting information on stool frequency, rectal bleeding and physician's global assessment, which was used to calculate a partial Mayo Score 18,19. The primary outcome measure was complications within 30 days after primary surgery (Fig. 1).

Figure 1

Flowchart illustrating the cohort selection of the study.

2.2 Statistical considerations

Student's t-test and Wilcoxon non-parametric test were used as appropriate. The threshold for statistical significance was predefined as a p-value below 0.05.

2.3 Ethical statement

The study was approved by the Danish Data Protection Agency, ref. #HVH-2010-05.

3 Results

Population demographics; preoperative medications; biochemical markers and disease activity parameters are summarized in Table 1.

View this table:
Table 1

Patient characteristics.

IFX group (n = 20)Control group (n = 51)P-value
Female (%)11/20 (55)22/51 (49)0.455
Smoker (%)1/20 (5)2/51 (3,9)0.255
Median (IQR) age at diagnosis (years)36,1(24,8–40,7)35,3 (25,3–46,2)0.300
Median (IQR) disease duration prior to surgery (years)3,7 (1,1–4,8)5,3(0,3–7,7)0.051
Median (IQR) age at surgery (years)39,8 (29,1–49,5)40,6 (28,7–51,4)0.411
Total days17,7 (10,5–22,8)16,7 (9,0–20,05)0.129
Before surgery7,1 (2,0–9,5)5,3 (1,0–8,0)0.087
After surgery10,9 (7,0–15,0)11,3 (6,0–13,0)0.039
Medical therapy prior to surgery
Corticosteroids (%)15/20 (75)33/51(64,7)0.356
Corticosteroid dose (mg prednisolon/day)51 (50,0–55,0)52,7 (25,0–75,0)0.192
Corticosteroid duration (days)51,8 (11,0–58,0)62,1 (10,0–91,0)0.429
Immunomodulators5/20 (25)16/51 (31,4)0.414
Medical therapy prior to surgery
5-ASA oral (%)10/20 (50)24/51 (47)0.441
5-ASA oral dose (mg/day)2200 (1800–2400)2604 (2400–3200)0.334
5-ASA topical (%)4/20 (20)8/51 (15,7)0.260
Biochemical markers
Hgb [20/49]a (mmol/l, 8.3–10.5)6,97 (6,3–7,4)7,45 (6,4–8,5)0.148
CRP [18/41]a (mg/l, < 10)42,72 (27,8–53,8)63,20 (19,0–88,0)0.050
Leukocytes [18/39]a (billion/l, 3,5–8,8)16,04 (10,2–15,4)12,81 (8,1–15,5)0.426
Albumin [19/45]a (g/l, 36–45)32,50 (27,8–37,1)30,31 (23,3–37,4)0.348
Creatinine [20/47]a (μmol/l 60–105)73,90 (64,8–81,8)72,35 (62,0–81,0)0.734
Partial Mayo-score [20/31]7,957,640.032

n = [IFX group/control group].

  • a Data for patient cohort only partially available.

No significant differences were observed between the IFX group and the control group regarding gender; smoking behavior; age at diagnosis; age at surgery; preoperative hospitalization or total hospitalization days. Patients in the control group were significantly longer hospitalized after surgery (11,3 days [6,0–13] vs. 10,9 days [7,0–15,0], P = 0.039). Indications for surgery were pharmacologically refractory disease, either chronic active disease, or acute and severe disease or steroid dependency. More patients in the control group had planned surgery (37,3% vs. 10%, P = 0.015) as a result of unacceptable quality of life caused by chronic active disease.

Medical therapy registered within 90 days prior to surgery was uniformly distributed in the two groups with regard to corticosteroids, immunomodulatory agents and maintenance treatment with 5-ASA.

In the IFX group a total of 34 infusions were administered. This resulted in a median (IQR) of infusions of 1,7 (1,0–2,0). The median duration between last infusion and primary surgery was 29 days (7,5–59,3 (IQR)).

No statistically significant differences between patients were observed regarding Hgb; leukocytes; albumin; and creatinine. The control group did however have a significantly higher CRP status preoperatively (63,2 [19,0–88,0] vs. 42,72 [27,8–53,8], P = 0.050). The number of patients with CRP-levels above the normal range (0–10 mg/l) was 16/18 (88,9%) in the IFX-treated group, and 36/41 (87,8%) in the control-group, the difference not reaching statistical significance (P = 0.476).

The control group had a significantly lower partial mayo-score (7,95 vs. 7,64, P = 0.032) compared to patients in the IFX group. The individual score of the patients had a low numeric variance.

3.1 Complications

Complications were registered within 30 days of surgery. There were no deaths during follow up in either group. Since dehydration can be the object of subjective appraisal, and anemia is dependent on degree of volume substitution and supply of blood products, these events were not recorded as complications.

A total of 10 patients (50%) in the IFX group and 25 patients (49%) of the control group had at least one postoperative complication (non significant, P = 0.444) (Fig. 2).

Figure 2

Early complications (within 30 days) after colectomy among patients treated versus not treated with infliximab.

The complications were divided into groups of infectious and non-infectious complications. Regardless of whether patients had multiple complications, they were only registered as either infectious, non-infectious or both. Infectious complications were observed in 4 patients (20%) in the IFX group and 21 patients (41%) in the control group, a difference which does not reach statistical significance (P = 0.181). Of the non-infectious complications observed, the IFX Group had a total of 6 (30%) as of 10 (18%) (P = 0.130) in controls (Table 2).

View this table:
Table 2

Short-term postoperative.

IFX groupControl group
Fever (%)4 (20)9 (18)
Wound infection (%)1 (5)7 (14)
Abscess (%)1 (5)2 (4)
Pneumonia (%)1 (5)1 (2)
Gastroenteritis (%)0 (0)2 (4)
Candida infection (%)0 (0)1 (2)
Epididymitis (%)1 (5)0 (0)
Urinary tract infection (%)0 (0)1 (2)
Sepsis (%)0 (0)1 (2)
Infection of undetermined cause (%)0 (0)1 (2)
Ileus (%)6 (30)7 (14)
Wound rupture (%)1 (5)13 (25)
Drug-induced eczema (%)0 (0)2 (4)
Postspinal headache (%)0 (0)2 (4)
Deep vein thrombosis (%)0 (0)1 (2)
Delirium (%)0 (0)1 (2)
Emphysema (%)0 (0)1 (2)

Evaluating the risk of other variables such as CS vs. no CS usage and a dose higher than median of CS (> 50 mg day) turned out as follows: infectious complications were observed in 22/48 (45,8%) patients receiving CS treatment compared to 3/23 (13,0%) of patients not being treated with CS, this difference being significant (P = 0.028). Non infectious complications were not significantly different (P = 0.099) with 8/43 (16,7%) in the CS group compared to 7/23 (30,4%) in the non-CS group. Also total complication rate turned out non significant (P = 0.498) with the CS group accumulating 26/43 (54,2%) risk of complications against 9/43 (39,1%) in the non-CS group. When looking at difference in dose of CS, either above/or below median, neither infectious (below-group 9/27 (33,3%) vs. above-group 13/21 (61,9%), P = 0.427); non-infectious (6/27 (22,2%) vs. 2/21 (9,5%), P = 0.0599); or total complications (13/27 (48,1%) vs. 13/21 (61,9%), P = 0.464).

4 Discussion

The aim of this study was to analyze retrospectively whether UC patients treated with IFX infusions had increased rates of short-term postoperative complications after colectomy.

Since treatment with IFX affects the immune system, it has been anticipated that this treatment is associated with both infectious and other complications 7. We found no association between treatment with IFX 12 weeks prior to colectomy in UC patients and the risk of 30-days postoperative complications. This conclusion is being shared by earlier studies.

Schluender et al. 14 compared an IFX group of 17 patients to a control group of 134 patients. Both patient groups were undergoing surgery in the period of primo 2000 to October 2005. The study showed no statistically significant difference regarding postoperative complications nor in comparing infectious complications alone. A weakness of their study, though, lies within its lack of assessment of disease severity before primary surgery, as well as the small cohort size in the IFX group precludes the possibility of a multivariate regression.

Another study concluding no difference in postoperative complications was published by Ferrante et al. 11, who compared an IFX-group of 22 patients to a control group of 119. This study found no difference in frequency of CS treatment between the groups (64% of IFX Group vs. 69% in the control group (P = 0.626)). Their study period spanned over a 10 year period from 1998 to 2008. The authors concluded that CS but not IFX increases the risk of short-term postoperative infectious complications in patients with UC. Also this study is also missing a measurement of the disease severity before primary surgery.

Couquet-Reinier et al.10 found no increase in complication rates in a small case-matched study with two equal groups of 13 patients undergoing ileal pouch-anal anastomosis.

Recently Gainsbury et al.12 published a study on postoperative outcomes in 81 UC patients from a tertiary academic center. They found no increased risk of short-term postoperative complications in 29 IFX -treated patients when compared to 52 IFX-naïve patients.

Contrary to this a study from the Mayo Clinic by Selvaskar et al. 15 came up with an opposite conclusion published in 2007. This study included 47 IFX patients and 254 control patients all treated between 2002 and 2005. Anastomotic leaks; pouch-specific and infectious complications occurred with a significant higher frequency in the IFX-treated group. IFX remained an independent predictor of postoperative complications, despite the fact that the IFX-group had a higher consumption of high dose CS, AZA, 5-ASA and had an endoscopically verified severe activity in disease. Differences disease severity in the examined groups and the fact that the IFX-group was selected with no regard to cut-off period of last IFX infusion make the study difficult to interpret.

In addition to this data reported by Mor et al. 13 included an IFX group of 85 patients and 438 control patients concluded that IFX-use was associated with an increased risk of early postoperative complications. But the study has some of the same weaknesses as the Mayo study, with an average median of the last IFX infusion being 13.5 weeks. A sustained biological effect of IFX beyond 12 weeks seems unlikely 17. In this study other potentially confounding variables were present, such as the possibility of IFX-treated patients having more severe disease than the control group.

A meta-analysis by Yang et al. 16 pooled data from previously published studies11,1315,20, and reported an association between IFX-use and overall postoperative complications, but found no associations between IFX and infectious or non-infectious complications when analyzed separately.

Our study design had several similarities with the studies listed above, and like these studies our findings also need to be interpreted with caution. The present study was like the previous studies not powered to identify significant differences in risk of infectious complications under IFX therapy. We did not use complete Mayo score, because several of the patients did not have an endoscopy close enough to the colectomy date and in app. 1/3 of the patients in the control group a partial score is lacking, mainly in patients operated electively. However, according to a previous study by Higgins et al. 19, endoscopy contributes little additional information to indices of disease activity in UC. Looking at potential confounders, the examined groups turned out to be significantly different in the following aspects: planned versus acute surgery; postoperative days of hospitalization; preoperative CRP status; the preoperative disease severity; and postoperative infectious complications of CS treated patients.

The proportion of planned surgery in the IFX Group was 10% and in the control group 37.3%, which proved to be a significant difference (P = 0.015). This difference suggests that the IFX-treated group as a whole had more active disease preoperatively. Disease activity assessment by the Mayo-score in the control-group strengthens this assumption, although the difference in absolute numbers is low. The IFX-treated patients had shorter length of postoperative hospitalization, may be interpreted as a result of a less complicated postoperative course in the IFX-group. This notion is strengthened by lower numbers of complications in the IFX-group, although this difference did not reach statistical significance. The modestly lower levels of CRP in the IFX-group, however, suggest less disease activity in IFX-treated patients. Although the IFX-treated group and the control group did differ in several areas, these differences were not uniform in direction. The overall implication is therefore that the IFX-treated patients had more active disease at the time of surgery than the controls, and for this reason had an increased risk of complications. Despite that the IFX group potentially was at greater risk due to higher disease activity, no increased risk in the IFX group could be demonstrated.

The results of this study showed no significant differences in terms of postoperative complications. This study is one of few studies that take duration of IFX biological activity into consideration. The studies that take durability of biological activity into account share the conclusion that IFX treatment preoperatively does not affect the rate of complications after surgery in the UC patient.

We did find an independent negative role for preoperative use of CS on the occurrence of short-term postoperative infectious complications, which is in agreement with previous studies 11. This finding is also in accordance with the data from both the TREAT 21 and the ENCORE 22 registry, which showed that CS use but not IFX use was associated with infectious complications in patients with CD.

In conclusion, the present study does not support the notion of IFX increasing the risk of short-term postoperative complications in ulcerative colitis.


Author Contributions: DB collected data, and drafted the paper. CM contributed to planning; design and data interpretation. FB conceived the study; contributed to the planning and data interpretation. All authors have been providing significant intellectual content to the final manuscript.

The authors would like to thank Karen Vibeke Jakobsen for excellent assistance in preparing the manuscript.


  • Disclosures: Dr. Bendtsen is a member of an advisory board for MSD, which markets infliximab in Denmark. Dr. Mortensen has received an unrestricted research grant from Schering-Plough, who previously marketed infliximab in Denmark.

Ulcerative Colitis
5-aminosalicylic acid
tumor necrosis factor-alpha
Crohn's Disease
Inflammatory bowel disease


View Abstract