Background Infliximab has been shown to be of benefit in the treatment of ulcerative colitis but long-term colectomy rates remain unknown.
Aims To review the rate of colectomy after infliximab for ulcerative colitis and to identify factors that might predict the need for colectomy.
Methods We conducted a retrospective cohort study of patients with active ulcerative colitis treated with infliximab between 2000 and 2006. The primary outcome was colectomy-free survival. Disease and treatment characteristics and complications were documented.
Results Thirty patients were treated with infliximab for refractory ulcerative colitis. Sixteen (53%) came to colectomy a median of 140 days after their first infusion (range 4–607). There was no difference in colectomy between those receiving infliximab for acute severe ulcerative colitis failing intravenous steroids (8/14) and out-patients with steroid-refractory ulcerative colitis (8/16). Only 17% (5/30) achieved a steroid-free remission after a median follow-up of 13 months (range 2–72). Univariate analysis showed that a younger age at diagnosis of colitis was significantly associated with an increased rate of colectomy (27.5 years vs. 38.7 years, P = 0.016).
Conclusion Over half the patients studied came to colectomy. Of those avoiding colectomy, only five (17%) sustained a steroid-free remission.
Ulcerative colitis (UC) has long been managed with 5-aminosalicylic acid (5-ASA) medications escalating to steroids and thiopurine analogues when disease is recurrent or severe. In a proportion of patients, flares can be acute and severe or be refractory to immunomodulators necessitating rescue medical therapy or colectomy. In these circumstances, ‘rescue therapy’ refers to second-line medical treatment to induce remission after unsuccessful treatment with steroids. Currently, two principal medications are used for rescue therapy in severe UC: ciclosporin (CsA) and infliximab (IFX). Tacrolimus and visilizumab are other options, but controlled trials are awaited.1
Ciclosporin has been used since 1984 as rescue therapy for intravenous steroid resistant UC, but its side effects and long-term outcomes are well recognized.2 Data from Oxford show a colectomy rate of 58% 7 years after CsA rescue therapy for severe UC, despite an initial remission of 74%. Importantly, 90% had relapsed requiring further treatment with steroids after 3 years of therapy.3 There is clearly a need for more effective rescue therapy for those patients with severe UC who do not respond completely to intravenous steroids.
Infliximab, a chimeric anti-tumour necrosis factor alpha (TNF-α) antibody of the IgG1 class has been used since 1999 in the treatment of Crohn’s disease. Early studies using IFX in UC had mixed results.4–6 Järnerot performed the main randomized controlled trial of IFX for severe UC refractory to intravenous steroids.7 Seven of twenty-four patients randomized to IFX (29%) underwent colectomy at 3 months compared with 14/21 (66%) in the placebo group. In this study, two different scores were used to randomize patients. Those randomized after 3-days of treatment according to the Sweden Index had more active disease than those randomized after 5–7 days. Paradoxically, it was those randomized later, with less active disease, and presumably lower TNF-α concentrations, who responded best to IFX. However, the long-term colectomy rate remains unclear. Data presented at UEGW 2006 showed that 5 years after Järnerot’s first patient received IFX 11/24 (46%) had undergone colectomy from the treatment group. The ACT 1 and 2 trials included out-patients with UC who had failed to respond to 5-ASA, corticosteroids or thiopurines. Overall, steroid-free remission rates of 21% at 7 months (30 weeks) were reported, but colectomy rates have not been presented.8
We aimed to determine the long-term colectomy rates after treatment with IFX for active UC failing conventional therapy in two groups: in-patients failing intravenous hydrocortisone and out-patients failing treatment with oral prednisolone.
Infliximab started to be used for UC at the John Radcliffe Hospital, Oxford in 2000. Therefore, we set up a database of all patients with UC who received IFX for active colitis between January 2000 and July 2006. Patients were sourced by cross-referencing pharmacy records, operating lists and attendances in the day unit to ensure all cases were captured.
Records were then reviewed to capture demographic data, disease extent and duration, medication history, dates, doses and indications for IFX and outcomes including colectomy and adverse events. The median follow-up was 13 months (range 2–72 months) (see Table 1). None of these patients have been reported previously, nor were any taking part in clinical trials with IFX.
Table 1. Patient characteristics
No colectomy group
Values within parentheses are expressed in %.
CsA, ciclosporin; IFX, infliximab.
Age at diagnosis
Time from Dx
No. of infusion (Av)
Follow-up from IFX
176 days (Av)
573 days (Av)
CsA pre IFX
T-W criteria at 1st infusion
The primary outcome measure was colectomy-free survival, with secondary measures of adverse events and need for further therapy. Further analysis was undertaken to ascertain whether any factors predictive of colectomy could be defined.
Severity of disease was determined according to the original Truelove and Witts’ criteria.9 Extent was defined as ‘pancolitis’ if disease was proximal to the splenic flexure, ‘left sided’ if distal to the splenic flexure and proximal to the sigmoid and ‘distal’ if distal to the sigmoid-descending junction. ‘Steroid-free remission’ was defined as normal stool frequency and no rectal bleeding whilst not taking steroids. A ‘sustained response’ to IFX was defined as steroid-free remission for greater than 3 months. ‘Relapse’ was defined as the need for further steroids after remission, being either infrequent (relapse ≤ 1 per year) or frequent (relapses >1 per year) as defined by Edwards and Truelove in 1963. ‘Chronic activity’ was defined as disease without remission despite treatment over 6 months. ‘Refractoriness’ was defined according to the response to steroids: intravenous steroid refractoriness was defined as CRP >45 with a stool frequency 3–8 on day 3 of intravenous steroids or continued rectal bleeding and stool frequency >3 per day after 7 days of in-patient treatment with steroids. Out-patient steroid-refractory UC was defined as continued rectal bleeding and stool frequency >3 per day with active colitis confirmed by sigmoidoscopy despite oral prednisolone in doses up to 0.75 mg/kg over a period of 4 weeks.
Administration of infliximab
In-patients with intravenous steroid-refractory UC received IFX as a single dose (5 mg/kg) a median of 4 days (range 1–14) after admission (similar to Järnerot). Out-patients with active oral steroid-refractory UC received a single dose (5 mg/kg) of IFX as a day patient. Clinical decisions to use IFX were not subject to prescribing restrictions. As treatment of most of these patients predated the ACT trials and 90% of the response to IFX occurs after the first dose,10 the response to a single infusion of IFX was evaluated and then repeated according to need. The decision to use IFX instead of CsA was based on patient choice, previous use of CsA and contraindications to CsA (hypomagnesaemia and low cholesterol).
Colectomy-free survival was analysed by a Kaplan–Meir curve; quantitative data were compared by a two-tailed Student’s t-test; qualitative variables and differences were analysed by chi-squared analysis or Fisher’s exact test.
Over the 6-year period (2000–2006), thirty patients were treated with IFX for active UC at the John Radcliffe Hospital. Forty-seven per cent (14/30) of patients had their first infusion for severe UC as an in-patient with the remainder having their infusion for out-patient moderate steroid-refractory or dependent disease. One patient was a current smoker (five cigarettes per day). All but three were British Caucasians, the others being Jewish (2) or Iranian (1). There was a slight preponderance of males in both the ‘colectomy’ group and the ‘non-colectomy’ group.
All patients had already been started on a thiopurine analogue by the time of IFX infusion except two. One was a newly diagnosed patient with severe disease requiring admission. The other was previously maintained on 5-ASA agents and then had a severe flare of colitis. Those on thiopurine analogues received 2–2.5 mg/kg of azathioprine or 1–1.5 mg/kg of mercaptopurine unless the dose was reduced because of bone marrow suppression. Ten patients had been unable to tolerate thiopurines and were on methotrexate 25 mg oral weekly. One patient had received long-term CsA therapy to maintain remission. Eleven patients had tried at least one course of CsA prior to IFX treatment. Two patients received CsA after IFX, one 40 days later and one 9 months later.
In-patients with intravenous steroid-refractory UC
Of the patients receiving IFX as an in-patient for intravenous steroid-refractory disease, 12/14 (85%) avoided colectomy during that current admission which is comparable with other studies.7, 11 After initial discharge, another six subsequently underwent colectomy a median of 274 days (range 40–343) after their first infusion. Overall, 8/14 (57%) patients receiving IFX for intravenous steroid-refractory UC ultimately underwent colectomy a median of 88 days (range 4–343) from their last infusion. Two of the fourteen (14%) achieved a sustained steroid-free remission.
Out-patients with oral steroid-refractory UC
Of the 16 out-patients receiving IFX for out-patient-refractory disease, eight (50%) ultimately underwent colectomy a median of 103 days (12–607) after their first IFX infusion. The median time to colectomy from their last infusion was 51 days (12–369). Three (19%) achieved a steroid-free remission with one relapsing during the study period. This is comparable with the ACT 1 and 2 studies8 and the outcomes are not significantly different to those receiving IFX for acute severe colitis.
Overall, 53% (16/30) underwent colectomy with a median time to colectomy of 140 days (range 4–607) after their first infusion (Figure 1). The median time from their last infusion to colectomy was 55 days (4–369). The most common indication for colectomy was intractable symptoms (88%). Overall, the colectomy rate was similar between intravenous steroid-refractory (8/14) and out-patient-refractory UC (8/16) patients. Of those avoiding colectomy (14/30) only five (17%) have achieved a steroid-free remission. Of the remainder, one is booked for a colectomy, two have been recommended colectomy but have refused, one is lost to follow-up and the rest have frequently relapsing disease.
All patients received infusions at 5 mg/kg (none at a higher dose), with a range from 1 to 7 infusions. No patients received the infusions in the 0-, 2- and 6-week induction regimen as specified in the ACT 1 and 2 trials8. Follow-up infusions in those that received them were given on an ‘as-needed’ basis as opposed to scheduled infusions, as neither maintenance therapy with IFX nor treatment of UC by IFX is yet approved by the National Institute for Health and Clinical Excellence (NICE) in the UK.
There were no documented infusion reactions or adverse events following IFX in these 30 patients. Of those who came to colectomy, there were five postoperative complications but no mortality. The median time from IFX to surgery was 40 days (5–286). Two patients suffered a small bowel obstruction, one suffered a simple perianal abscess, one had widespread skin candidiasis (colectomy 40 days post-IFX) and one had a postoperative fever thought to be portal bacteraemia on imaging (colectomy 286 days post last IFX). All patients were also on either immunomodulators, a minimum of 15 mg of prednidolone, or both up to the time of surgery. Given the concomitant medications and the length of time between IFX and surgery, a clear causation cannot be attributed.
Univariate analysis for preoperative factors predictive of need for colectomy identified a statistically significant difference in the age at diagnosis and subsequent operation rate. Those undergoing colectomy had a mean age at diagnosis of 27.5 years compared with 38.8 years for those avoiding colectomy (P = 0.016; CI: 2–19). There was, however, no significant difference in the age at the time of colectomy (37 years vs. 46 years, P = 0.083). However, the numbers of patients are small which cautions against over-interpretation of these differences.
There was no statistically significant difference between severity of disease at first infusion and colectomy (8/16 vs. 4/14, P = 0.28), nor between the two groups of patients (in-patient 8/14 vs. out-patients 8/16 underwent colectomy). There was no difference between numbers of patients receiving prior CsA (8/16 vs. 5/14, P = 0.48), disease extent, numbers of patient on immunomodulators or numbers of patients refractory or dependant on steroids or immunomodulators.
The management of intravenous steroid-refractory UC and oral steroid-refractory UC is becoming more complicated as therapeutic options increase.1 No evidence-based guidelines are available to clinicians on how best to use these new drugs and there is a real danger of drug-related mortality. As the mortality of severe UC in specialist units is now no more than 1%, including surgical mortality, an increased mortality resulting from drug therapy is unacceptable.12 Thus, new therapies for refractory UC need to be evaluated not only for their short-term efficacy and safety profile, but also on their effect on the long-term course of the disease.
Despite the use of CsA showing excellent initial remission rates of up to 80% in acute severe UC, the majority of those treated will come to colectomy over the next 7 years.2, 3 This implies that treatment is shifting the colectomy-free curve to the right, but not affecting the pattern of disease or changing the ultimate outcome. This may be a reasonable approach for some patients, especially those with distal disease, or a new presentation of UC, because it gives patients’ time to come to terms with their likely need for colectomy, and will spare some from this fate. But at what cost? The literature reports deaths in patients treated with CsA, in addition to the recognized risks of hypertension and renal impairment which may not be reversible. It is worth remembering that just one death per 100 treated patients will exceed the mortality shown for steroid therapy followed by surgery.12
This study is limited by being retrospective, having a selected patient group and not using a 0-, 2- and 6-week induction regimen. Patients received further doses on an as-needed basis instead of receiving scheduled maintenance therapy. However, its strength lies in it being a single-centre study with a predominantly secondary care population (<20% were tertiary referrals). The study represents everyday clinical practice from a centre which follows well-characterized procedures and has regularly published its experience with severe UC. The use of a single infusion of IFX is supported by data showing 90% of the response occurs after one dose and the message from the Swedish–Danish study indicates this may be sufficient.7, 10 Furthermore, the outcome as measured by steroid-free remission (17%) is not appreciably different from that reported in the ACT trials (21%), where patients received the induction regimen and then maintenance IFX.8
The need for better medical therapy for severe UC is pressing. IFX has now been studied in both acute severe UC and in moderate to severe out-patient UC. Järnerot enrolled 45 patients with severe or moderate to severe UC requiring intravenous steroids and showed a significant reduction of the 30-day colectomy rate in the moderate to severe cohort (characterized by the Seo index), but no significant difference in the more severe cohort (characterized by the Sweden Index).7 Of all those treated with IFX, 29% underwent colectomy at 3 months, but at 5 years that number had risen to 46% (unpublished data presented in Berlin, UEGW, 2006). An Italian multi-centre study by Kohn et al. presented at UEGW 2006, examined the long-term outcome of 83 patients treated with IFX for severe intravenous steroid-refractory UC.11 The colectomy rate at 60 days was 14% but by 3 years had risen to 48%. Of note there was a death from a lung abscess (mortality >1%). Recently, the Leuven group published their first 100 patients with UC treated with IFX.13 They reported a colectomy rate of 11%, but their patient group was significantly different to ours. They included patients enrolled in the ACT 1 study and only 63% of their patients were on an immunomodulator. Our patients were significantly more unwell with 11/30 having tried CsA and 10/30 were on methotrexate having been intolerant to thiopurines.
Our study confirms that there is a high colectomy rate despite IFX and suggests, much like CsA, that the time to colectomy can be delayed, but that at least half will ultimately come to colectomy (53% over 6 years). It is vitally important to look beyond the single outcome of colectomy and see how those who avoided it have fared. In our study, only five of the original 30 patients maintained a steroid-free remission at the end of follow-up (median 13 months, range 2–72). This implies that the remainder were receiving further courses of steroids and/or IFX or CsA. Interestingly, our steroid-free remission rate (17%) was almost similar to that seen in the ACT 1 and 2 studies (21%), although three of the five in remission in our study had severe, intravenous steroid-refractory disease at the first infusion.
Despite initial enthusiasm, IFX does not appear to prevent the need for colectomy in most patients with UC. It has a potentially valuable role in delaying the need for colectomy, but over half will require surgery over 6 years. Its drawback is a consistent serious adverse event incidence of about 5%, including opportunistic infections.14 Younger age at diagnosis may be a risk factor for predicting the need for colectomy.
The initial enthusiasm for IFX in the treatment of UC should be tempered by these findings. Potential side effects and expected outcomes, including the minority who achieve steroid-free remission, should be made very clear to any patient in whom its use is contemplated.
Declaration of personal and funding interests: S. Jakobovits has received no financial support. S. Travis has received unrestricted educational grants from, acted as advisor to, or lecturer for Centocor and Schering Plough. D. Jewell has acted as a consultant for Schering Plough.