These authors contributed equally to this work.
Infliximab in severe ulcerative colitis: short-term results of different infusion regimens and long-term follow-up
Article first published online: 21 JUN 2007
Alimentary Pharmacology & Therapeutics
Volume 26, Issue 5, pages 747–756, September 2007
How to Cite
KOHN, A., DAPERNO, M., ARMUZZI, A., CAPPELLO, M., BIANCONE, L., ORLANDO, A., VISCIDO, A., ANNESE, V., RIEGLER, G., MEUCCI, G., MARROLLO, M., SOSTEGNI, R., GASBARRINI, A., PERALTA, S. and PRANTERA, C. (2007), Infliximab in severe ulcerative colitis: short-term results of different infusion regimens and long-term follow-up. Alimentary Pharmacology & Therapeutics, 26: 747–756. doi: 10.1111/j.1365-2036.2007.03415.x
- Issue published online: 21 JUN 2007
- Article first published online: 21 JUN 2007
- Publication data Submitted 16 March 2007 First decision 6 April 2007 Resubmitted 15 June 2007 Accepted 18 June 2007
Severe ulcerative colitis is a life-threatening disorder, despite i.v. glucocorticoids treatment. Infliximab has been proposed as a safe rescue therapy.
To evaluate short- and long-term effectiveness and safety of infliximab in severe refractory ulcerative colitis.
Eighty-three patients with severe ulcerative colitis (i.v. glucocorticoids treatment-refractory) were treated with infliximab in 10 Italian Gastroenterology Units. Patients underwent one or more infusions according to the choice of treating physicians. Short-term outcome was colectomy/death 2 months after the first infusion. Long-term outcome was survival free from colectomy. Safety data were recorded.
Twelve patients (15%) underwent colectomy within 2 months. One died of Legionella pneumophila infection 12 days after infliximab. Early colectomy rates were higher in patients receiving one infusion (9/26), compared with those receiving two/more infusions (3/57, P = 0.001, OR = 9.53).
Seventy patients who survived colectomy and did not experience any fatal complications were followed-up for a median time of 23 months; 58 patients avoided colectomy during the follow-up. Forty-two patients were maintained on immunosuppressive drugs. No clinical features were associated with outcomes.
Infliximab is an effective and relatively safe therapy to avoid colectomy and maintain long-term remission for patients with severe refractory ulcerative colitis. In the short term, two or more infusions seem to be more effective than one single infusion.
Ulcerative colitis (UC) is a relapsing and remitting inflammatory disorder commonly managed with aminosalicylates, glucocorticoids and immunosuppressive drugs. Fifteen to nineteen per cent of UC patients1, 2 have severe attack at some time during their illness requiring hospitalization and intensive i.v. glucocorticoid treatment. About 30–40%3, 4 of the patients with severe attacks fail to respond to i.v. glucocorticoids and require colectomy, and several clinical markers of poor response to steroids have been proposed until now5–7, 21. Although surgery is considered to be curative for patients with UC, the quality of life after restorative proctocolectomy is generally poorer than in patients who respond to medical therapy and thus avoid surgical treatment.8
Intravenous ciclosporine has been proposed as rescue therapy for severe UC.9, 10 However, in spite of high remission rates, i.v. ciclosporine has a limited use because of risks of major toxicity, including occasional fatalities, and high rate of relapse after discontinuation.11
Infliximab is an anti-tumor necrosis factor α (TNFα) chimeric monoclonal antibody that has recently been tested in two large multicentre, placebo-controlled studies (ACT 1 and ACT 2) in patients with active UC despite treatment with conventional therapy. Using an induction regimen of three doses of infliximab, followed by maintenance infusions every 8 weeks, significant efficacy was demonstrated in terms of clinical response and mucosal healing.12 In addition, the results of a double-blind placebo-controlled trial performed in severe and moderately severe UC refractory to i.v. glucocorticoids have recently been published.13 This study confirmed previous data supporting the efficacy and the safety of infliximab as rescue therapy in severe UC.14–18 Long-term follow-up in these patients seems to indicate infliximab as an effective agent for long-standing remission that may act as a ‘bridge therapy’ for immunomodulators in long-term maintenance treatment.19
We report here the results of an Italian multicentre open-label study evaluating the effectiveness and the long-term outcome of infliximab as a rescue therapy in patients hospitalized for an acute attack of UC and candidates for colectomy, because of their being resistant to intensive i.v. glucocorticoid treatment (IIVT).
From May 2000 to January 2006, 83 in-patients (49 men, 34 women) admitted to 10 Italian referral centres [Rome ‘San Camillo’ (n = 17), Turin ‘Mauriziano’ (n = 21), Rome ‘Università Cattolica’ (n = 14), Palermo ‘Università di Palermo’ (n = 10), Rome ‘Tor Vergata’ (n = 5), Palermo ‘V. Cervello’ (n = 4), Rome ‘Università La Sapienza’ (n = 4), San Giovanni Rotondo ‘Casa Sollievo della Sofferenza’ (n = 4), Naples ‘Seconda Università’ (n = 3), Como ‘Valduce’ (n = 1)] were treated with infliximab 5 mg/kg of body weight for acute severe or moderately severe attack of UC. All patients were candidates for colectomy because of resistance to IIVT (i.v. methylprednisolone 50–60 mg daily, or equivalent) administered for 7 or more days. This study includes retrospective analysis of clinical records (n = 46) as well as prospective data (starting from 2003) obtained using a shared clinical form (n = 37). All patients had UC diagnosis established by commonly accepted clinical, endoscopic and histological criteria. Abdominal X-ray films were obtained to exclude possible toxic megacolon or perforation and to establish the approximate extent of colitis.20 A flexible proctosigmoidoscopy was usually performed to evaluate disease activity at hospital admission, based on usual policy adopted in different centres.
Endoscopic severity was considered as both presence and absence of deep colonic ulcer, based on Carbonnel et al.21 criteria, and the overall observers’ global evaluation. All patients were tested for enteric stool pathogens and Clostridium difficile toxin. A chest radiograph was performed in all patients and intradermal test with PPD, together with accurate personal history evaluation was carried out since 2001 for excluding ongoing or past tuberculosis infection. No patient had received infliximab in the past.
Each patient provided written, informed consent to treatment with infliximab. All patients received a single i.v. infusion of infliximab at a dose of 5 mg/kg. Different infusion protocols were used, with either one single, two or three infusions, based on individual treating physicians’ preferences, not on the ground of clinical response. Serum levels of C-reactive protein (CRP) were evaluated before and three and 7 days after infusion of infliximab or before colectomy in patients who failed to respond.
All concomitant medications were continued, if clinically indicated.
Patients were recruited according to severe flare as defined by Truelove and Witts22 and modified by Chapman et al.23– six or more bloody motions per day and at least one of the remaining criteria: fever (mean evening >37.5°C or > 37.8°C for 2 days out of four), tachycardia (mean pulse rate >90 per minute), anaemia (decrease in haemoglobin levels greater than 75%), erythrocyte sedimentation rate (ESR) > 30 mm/h. To investigate if patients with severe disease, as originally defined by Truelove and Witts by all five criteria22, may show a different response to treatment, two subgroups were formed: one including patients who fulfilled all the criteria (subgroup S1) and one including patients who met all but 1 or 2 criteria (subgroup S2). Clinical activity was measured by the Lichtiger Clinical Activity Score for acute UC (CAI).9 The CAI takes into account symptoms – diarrhoea, nocturnal diarrhoea, visible blood in stools, faecal incontinence, abdominal pain, general well-being, abdominal tenderness, need for anti-diarrhoeal drugs – with a maximal score of 21; severe clinical activity matches with a score ≥12. A score of <10, on two consecutive days, was considered as clinical response; clinical remission was defined by a CAI score of 4 or less.24 All patients had a score ≥12 calculated prior to the first infliximab infusion.
Patients whose condition worsened or failed to respond to treatment underwent colectomy. In the individual patient, to perform colectomy was a joint medical–surgical decision.
Patients were followed-up after the first infusion by serial clinical evaluation. Clinical activity was evaluated by using CAI. If there was no recent contact (<2 months), the patient was evaluated by telephonic interview.
The primary endpoint was survival free from colectomy or death, within 2 months from the first infliximab infusion.
Secondary endpoints were: (i) clinical remission at 1 month after the first infliximab infusion and during the long-term follow-up, and (ii) time to clinical relapse defined as the need for a new steroid course and/or infliximab or surgery.
Colectomy performed within 2 months from the first infusion of infliximab was defined as early colectomy, while any colectomy performed during the follow-up period was considered late colectomy.
Statistical analysis was carried out using medcalc (version 9.0, Mariakerke, Belgium) statistical software for Windows.
Descriptive statistics was used to summarize the data; frequencies and median with inter-quartile ranges (IQR) for categorical and continuous variables were used, respectively, as appropriate. Univariate analysis was carried out to explore the differences between groups: chi-squared, Fisher’s exact test and the Mann–Whitney test for categorical and continuous variables were used, as appropriate.
Stepwise logistic regression was used to test the independent association of different clinical characteristics to the outcome variables, i.e. early or delayed colectomy and clinical relapse (which were set as dependent variables).
Survival analysis (for 60-day colectomy, late colectomy and clinical relapse) was carried out using Cox proportional-hazards regression, that allows analysing the effect of several risk factors (covariates) on survival, and calculating the relative risk attributable to significantly associated covariates.25
P -values <0.05 were considered statistically significant; whenever appropriate, 95% CI were reported to aid the understanding of magnitude of differences or variations.
Population demographics and disease characteristics are summarized in Table 1.
|Variable||All patients (n = 83)||Subgroup S1 (n = 66)||Subgroup S2 (n = 17)||P-value|
|Age (years), median (IQR)||36 (27–50)||36 (26–51)||37 (35–46)||0.502|
|Disease duration, months median (IQR)||37 (12–72)||28 (12–72)||50 (37–82)||0.055|
|First year of disease, n (%)||18 (22)||16 (24)||2 (12)||0.339|
|Disease extent, n (%)|
|Total colitis||56 (67)||42 (64)||14 (82)|
|Left sided colitis||23 (28)||20 (30)||3 (18)|
|Distal colitis||4 (5)||4 (6)||0||0.285|
|Smokers, n (%)||6 (7)||6 (9)||0||0.444|
|Previous smokers, n (%)||14 (17)||12 (18||2 (12)||0.789|
|Concomitant medications, n (%)|
|Aminosalicylates||47 (57)||38 (58)||9 (53)||0.788|
|Antibiotics||52 (63)||42 (64)||6 (35)||0.053|
|Azathioprine/6MP||20 (24)||14 (21)||6 (35)||0.535|
|Cyclosporin||9 (11)||9 (14)||0||0.192|
|Parenteral nutrition, n (%)||27 (32)||22 (33)||5(29)||0.986|
|Blood transfused patients, n (%)||15 (18)||14 (21)||1 (6)||0.266|
|Hb g/dL, median (IQR)*||10.6 (8.9–12)||10.2 (8.7–11.4)||12.2 (11.4–12.9)||<0.001|
|CRP mg/L, median (IQR)*||45 (30–72)||48 (32–78)||30 (17–43)||0.019|
|CAI, median (IQR)*||14 (12–15)||14 (13–16)||11 (11–13)||<0.001|
|Endoscopy, n (%):||79 (95)||65 (98)||14 (82)||0.026|
|Deep ulcers||47 (59)||39 (60)||8 (57)||0.918|
|Duration of IIVT before infliximab, days, median (IQR)||8 (6–13)||9 (6–13)||6 (1–12)||0.116|
|Dose of methylprednisolone or equivalent, mg/day, median (range)||60 (20–80)||60 (20–80)||60 (30–80)||0.489|
Among the 83 patients included in the study, 66 fulfilled all five criteria for severity (group S1) and 17 fulfilled all but one or two criteria (group S2). Twenty-six, 14 and 43 patients underwent one (at week 0), two (at week 0, and after 2 or 4 weeks) and three (at week 0, after 2 and 6 weeks) infliximab infusions, respectively. The infliximab schedule (1, 2 or 3 infusions) did not differ between patients of the S1 and S2 groups.
Short-term response and early colectomy
Seventy of the initial 83 patients (84%) avoided colectomy at 2 months. Twelve patients (15%) were operated for the absence of clinical response; among them, two underwent surgery within 4 days because of clinical deterioration. All 12 patients who underwent surgery belonged to S1 group (12/66, 18%). All the 17 patients of the S2 group avoided colectomy (P = 0.114). The median time to operation after infliximab infusion was 27 days (95% CI: 8–53); four and seven patients underwent colectomy within 15 and 30 days, respectively. Sixty-one patients (73%) reached clinical remission (CAI ≤ 4) at 1 month. No UC-related death was reported. One patient died because of Legionella pneumophila infection, 11 days after infliximab administration. Outcome measures within 2 months from the first infusion of infliximab are summarized in Table 2.
|Patients||Operated n (%)||p||OR (95% CI)||Death n (%)||p||OR (95% CI)||Severe n (%)|
|Overall||83||12 (15)||NA||NA||1 (1)||NA||NA||66 (80)|
|Number of infusions|
|1 infusion||26||9 (35)||0.001||9.53 (2.31–39.26)||1 (4)||0.313||1.04 (0.96–1.12)||22 (85)|
|≥2 infusions||57||3 (5)||0||44 (77)|
|Subgroup S1||66||12 (18)||0.114||1.22 (1.09–1.37)||1 (1)||1.000||1.02 (0.99–1.05)||NA|
No clinical characteristic was significantly associated with 2 months colectomy rate, including basal bowel motions, clinical activity index, endoscopic findings, CRP or ESR levels. The only variable significantly associated with the risk of colectomy was the number of infliximab infusions (P = 0.001, Figure 1). Cox proportional hazards regression confirmed that the number of infliximab infusions was the only significant predictor of short-term colectomy (P = 0.005, Figure 2). The clinical characteristics of patients in the three infliximab dosing groups were not significantly different (Table 3).
|Variable||1 infusion (n = 26)||2 infusions (n = 14)||3 infusions (n = 43)||P-value|
|CAI, median (IQR)*||15 (13–16)||13 (12–14)||14 (12–15)||0.110|
|First year of disease, n (%)||4 (15)||2 (14)||17 (39)||0.114|
|Deep ulcers at endoscopy, n (%)||16/24 (67)||8/12 (67)||23/43 (53)||0.494|
|Mean duration of IIVT before infliximab, days (IQR)||11 (7–14)||6 (4–8)||8 (6–13)||0.089|
|CRP mg/L, median (IQR)*||49 (32–80)||45 (36–66)||40 (29–76)||0.684|
|Blood transfused patients, n (%)||7 (27)||1 (7)||7 (16)||0.273|
|Pancolitis, n (%)||19 (73)||7 (50)||30 (70)||0.298|
Seventy-nine out of 83 patients (95%) underwent endoscopic evaluation during the acute attack; sixty-two (79%) had severe endoscopic activity with deep ulcer in 47 (60%) patients. The endoscopic severity did not seem to affect the rate of colectomy (Fisher’s exact test P = 0.510 and P = 1.000, respectively). Serum levels of CRP, measured at baseline, after 3 and 7 days following the first infliximab infusion, were correlated with the clinical outcome (Figure 3); CRP levels did not differ significantly basally, while they were significantly lower in patients avoiding colectomy compared with those who underwent colectomy within 2 months, both at 3 and 7 days after infusion (P = 0.598, P = 0.006 and P = 0.001, respectively).
Long-term response and late colectomy
Seventy patients, who attained the primary outcome avoiding colectomy or death, were followed-up for a median time of 23.4 months (IQR 15–40.3); among them, 59 patients (84%) had a follow-up longer than 1 year. During the follow-up period, 68 (97%) patients discontinued glucocorticoids. Fifty-four (77%) patients continued on immunosuppressive treatment with azathioprine or 6 mercaptopurine (6MP), alone (n = 3) or in combination with infliximab every 8 weeks (n = 12), or with 5-ASA derivatives (n = 39); three patients were on methotrexate, while 14 patients were on 5-ASA derivatives alone. During the follow-up period, 27 patients (39%) relapsed after a median interval of 13.5 months (IQR 5–23). The relapse rate was not associated with any maintenance regimen. Among relapsing patients, 13 were treated successfully with a course of oral/parenteral glucocorticoids, two received further infliximab infusions, and 12 required surgery. Cox proportional-hazards regression survival analysis for follow-up colectomies is shown in Figure 4; no covariate was significantly associated with the outcome (P = 0.693). Among patients who underwent surgery, three were on maintenance treatment with aminosalicylates, eight with immunosuppressants, and one patient on re-treatment with infliximab.
Survival analysis for remission free from colectomy and from disease relapse during the follow-up period, as well as logistic regression analysis, did not show any effect of the initial number of infliximab infusions or of any other clinical variable on the outcomes.
Nine out of 83 patients (11%) reported severe adverse events (five infections and four infusion reactions).
One patient died of a pulmonary abscess, 11 days after the first infliximab infusion. The patient was a 71-year-old male with a 2-year history of UC, on mesalazine maintenance when he experienced a severe steroid-refractory relapse of UC, which responded successfully to infliximab after 12 days of IIV steroid treatment. Two days after hospital discharge, on glucocorticoids, he developed pneumonia complicated by pulmonary abscess. Legionella pneumophila serogroup 1 was isolated from sputum samples by polymerase chain reaction analysis. Despite treatment with antibiotics, the patient died of a septic shock 8 days later.
One patient developed pulmonary tuberculosis 2 months after the second infliximab infusion (this patient has been treated before 2001); one patient suffered from pneumonia, one had an herpes simplex virus infection with fever and headache and one patient had a Candida albicans sepsis that occurred 3 months after the last infusion, and therefore was probably not related to infliximab. Infusion reactions occurred in one patient after the second infusion and in three patients after the third infusion of infliximab.
No newly diagnosed malignancies or dysplasia were reported in any of the 83 patients included, although no specific investigations were performed in this regard. No post-operative complications were reported.
Postsurgical morbidity and mortality
Overall colectomy rate (early and delayed, 12 cases each) was 29% (24/84) among IIVT-refractory severe UC cases treated with infliximab.
Short-term postsurgical morbidity in this series was 33% (8/24), with two infectious complications (8%), two early anastomotic leaks treated conservatively (8%) and three pouch-related adverse events (13%), the remaining adverse event was non-fatal pulmonary thromboembolism. Postsurgical mortality was 0% in this series.
This is the largest series evaluating the effectiveness of infliximab in patients with severe ulcerative colitis refractory to glucocorticoids. The results of this study confirm the efficacy of infliximab in acute severe UC. The short-term survival rate free from colectomy (84%) in this group of patients with severe disease is similar to the figure of 71% recently reported by Järnerot et al.13 Our study population, similar to that of Järnerot trial, enrolled the hospitalized patients who failed intensive glucocorticoids treatment and were candidates for surgery.
The primary endpoint of avoiding colectomy or death in this high-risk group of patients has been achieved. In fact, only 15% of the patients were operated in concomitance with the acute episode, while, of those who escaped from the acute attack, 17% were colectomized during the follow-up period after the disease relapsed. The short-term evaluation time for early colectomy was 2 months from the first infusion of infliximab; all early colectomies were performed within this time interval, while late colectomies were carried out when the disease relapsed after a clinical remission of at least 4 months. In this study, the clinical records of 46 patients were retrospectively analysed. They had been treated with infliximab between May 2000 and December 2002 when no treatment protocol was designed. In each centre, all patients were treated by the same doctors who thereafter retrospectively analysed the clinical records. We are aware that a mix of retrospective and prospective data could be a methodological bias that may reduce the strength of evidence, but we think that the rate of colectomy should not have been influenced by retrospective data. There was no predefined infusion regimen and patients received different number of infusions at different interval times following the decision of the physicians in charge of each of the 10 centres that participated in the study.
Early-colectomy rates seem to indicate that two or three infusions work better than one single infusion of infliximab. This finding could explain the overall lower rate of early colectomy reported in our study compared with that of the Järnerot trial in which only one infusion of infliximab was employed.13 No significant differences in baseline disease characteristics were observed among the three groups of patients treated with different number of infusions.
The chance to receive the second and third infusion of infliximab did not depend on the clinical response to treatment, because a standard protocol for treatment of UC with infliximab has been proposed only very recently.12 Therefore, the decision was influenced by either fear of adverse events or lack of scientific evidence that subsequent infusions may induce a clinical advantage.
However, we evaluated the possibility that the worsening of clinical conditions or the failure to respond within 2 weeks may have precluded an eventual second infusion at week 2, thus increasing the colectomy rate in the group of patients receiving a single infusion of infliximab. But, when individual data were subsequently analysed and the four patients operated within 2 weeks from the first infusion were excluded, early-colectomy rates were still significantly higher in the group of patients who received one single infusion of infliximab (23%) compared with the group of patients who had undergone two or more infusions (5%; P = 0.034; OR = 5.29, 95% CI: 1.14–24.49). This finding would suggest that failure to respond to infliximab, which may have limited the treatment to a single infusion before colectomy, did not influence the better results obtained with one or two subsequent infusions. After a median follow-up of more than one-and-a-half year, 70% (58/83) of patients have avoided colectomy so far and 58% remained in remission. Colectomy rates and time to colectomy during the follow-up period were not affected by the initial infusion protocol.
In our study, the proportions of patients with serious adverse events (11%) were similar to the figure of 11% reported in the ACT 2 study.12
The overall early postsurgical morbidity of proctocolectomy for ulcerative colitis was reported to be 50–60%41, with a slight worse prognosis for some surgical outcomes in patients operated after an infliximab rescue course. In our series, the occurrence of early morbidity was lower than previously described41, but still about one-third of patients who underwent colectomy reported peri-operative adverse events. Our figures do not differ from those reported in patients not receiving infliximab before surgery26, 27, confirming that infliximab does not seem to affect the post-operative course of severe UC.13
The case of fatal L. pneumophila infection, possibly related to infliximab, is of concern. Legionella pneumophila infection is a relatively common but under-diagnosed cause of both community- and hospital-acquired pneumonia with a 25% mortality rate, which may increase up to 40% in nosocomial cases.28
The risk of legionellosis and the likelihood of death are markedly increased in severe immunocompromized patients.28, 29 Four cases of L. pneumophila infection were reported during treatment with infliximab in three patients with rheumatoid arthritis30–32 and in one with psoriasis.33 To our knowledge, this is the first case of fatal L. pneumophila infection reported in a patient with UC treated with infliximab. Studies on macrophage cultures34 and on experimental animals infected with L. pneumophila and treated with TNFα antibodies35 suggest an essential role for TNFα in the defence against intracellular bacteria. Is worth noting that the patient was also treated with glucocorticoids, for total 23 days, which may be a further independent risk factor for Legionella infection. The length of glucocorticoid treatment and the interval between admission and rescue therapy have to be accounted for the reason that inflximab was used out of label under personal responsibility of the treating physician. An earlier use of rescue therapy with infliximab may reduce the infection-related mortality.
Our data on short-term success rate and long-term effectiveness with infliximab are similar to those obtained in severe UC treated with i.v. ciclosporine9–11, 36. However, infliximab seems to have a better safety profile with respects to infections, mortality and major toxicities.11, 37
Safety issues remain a main concern and should be counterbalanced with the risk of surgery38–41 in this critical situation, with post-operative complications and with the better quality of life without colectomy.37–42
In conclusion, our data strengthen the concept that infliximab is an effective rescue therapy for severe ulcerative colitis refractory to conventional treatments. Two or three infusions seem to be more effective than a single one for preventing early colectomy. However, because of methodological limits and small number of patients, a dose-finding study should be performed to confirm this finding. An earlier recourse to rescue therapy with infliximab could increase the safety.
Authors’ declaration of personal interests: None. Declaration of funding interests: Marco Daperno was funded by Fondazione IBD Onlus, with a grant by Compagnia di San Paolo (project ref. 2002.117).
- 2Inflammatory bowel disease. In: YamadaTed. Textbook of Gastroenterology: A Community Study, 2nd ed. Philadelphia: JB Lippincott Company, 1995: 1748–806.
- 14Infliximab for refractory ulcerative colitis. Am J Gastroenterol 2001; 96: 2373–81., , , et al.Direct Link:
- 25An Introduction to Medical Statistics, 3rd edn . Oxford, UK: Oxford University Press, 2000..
- 36Cyclosporine A for induction of remission in severe ulcerative colitis. Cochrane Database Syst Rev 2005; 1: CD004277., , , et al.
- 37Intravenous cyclosporin in ulcerative colitis: a five-year experience. Am J Gastroenterology 1999; 94: 1587–92., , .Direct Link: