Infliximab in ulcerative colitis: is there a placebo (effect) in the house?

Abstract

Su C, Salzberg BA, Lewis JD, Deren JJ, Kornbluth A, Katzka DA, Stein RB, Adler DR, Lichtenstein GR (The University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania; Atlanta Gastroenterology Associates, Atlanta, Georgia; Mount Sinai School of Medicine, New York, New York; Lutheran General Hospital, Park Ridge, Illinois). Efficacy of anti-tumor necrosis factor therapy in patients with ulcerative colitis. Am J Gastroenterol 2002;97:2577–2584. Infliximab has recently emerged as an effective agent in the treatment of patients with Crohn’s disease (CD). Overlap of the diagnosis and treatment of the inflammatory bowel diseases has generally allowed for application of therapies designed for either CD or ulcerative colitis (UC) to be used interchangeably; whether this holds true for infliximab remains to be seen. Inadequate placebo-controlled data has resulted in reliance on anecdotal reports and very small open-label experiences. A mixed group of academicians and private practitioners recently published their open label experience with infliximab in 27 patients with active UC. Patients were identified by their physician, medical records reviewed, and patients were telephoned to obtain follow-up data. The diagnosis of UC was based on standard clinical, endoscopic, and histological findings. Disease activity was graded by a modified index (DAI) consisting of stool frequency, rectal bleeding, and physician rating of disease activity (scored 0–4 for each, 0 implying “normal”). Patients were categorized as “steroid-refractory” if no response after 7 days to 300 mg/day hydrocortisone (or equivalent), and “steroid-dependent” if symptoms worsened when prednisone was tapered to <10 mg/day (or equivalent). Infliximab infusions were at 5 mg/kg (one was 3 mg/kg), and reinfusions were at the discretion of the physician. Remission (complete response) was defined as complete resolution of all symptoms, with DAI = 0. Partial response was defined as a substantial but incomplete DAI improvement. All others were categorized as no response. Seventy-four percent of the patients were men, with median age 40 years (range, 20–78 years). Sixty-seven percent had extensive UC, and 89% severe disease activity. Eighty-nine percent were concurrently on corticosteroids or another immunomodulator. Sixty-three percent were outpatients, 33% were steroid-refractory, and 37% were steroid-dependent. Five patients had failed to respond to double-blinded experimental therapies given within 2 days to 9 weeks (median, 4 weeks) before infliximab. Approximately half of the patients received more than 1 infusion; 89% had received up to 3 infusions. Median follow-up was 4 months (4 days to 16 months; >1 month in 24 of 27). Response was seen in 2 of 3 patients; remission was seen in 44%. At 6 months, remission was seen in 5 of 10 patients, and a partial response in a sixth patient. Most nonresponders subsequently underwent colectomy. Median time to response was 4 days (1 day to 3 weeks), and median duration was 8 weeks (2 weeks to 15 months). Half of the patients relapsed, and all but 1 responded to repeat therapy. Response was lowest in patients with steroid-refractory disease (1 of 3) and was not affected by demographics or concomitant immunomodulators. Steroids were reduced in 9 of 10 patients and completely withdrawn in 3 patients. Serious adverse events were seen in 2 patients, both of whom had received 2 infliximab infusions and then proceeded to colectomy: a postoperative death from line sepsis and subacute bacterial endocarditis in 1 patient and candidemia that resolved with therapy in another. The authors concluded that infliximab therapy is beneficial in UC patients, even those with severe disease refractory to mesalamine and immunosuppressives, but not to corticosteroids. The startling success of infliximab for the treatment of CD has understandably led to the application of this therapy in UC. There has been much speculation as to whether tumor necrosis factor (TNF) is as important a cytokine in UC as it is in some patients with CD. Earlier investigations suggested that the TNF might be more critical in Th-1 immunological responses (such as CD), whereas the Th-2 immunological conditions such as UC were reliant on other cytokines (Gastroenterology 1997;113:118–126, Clin Exp Immunol 1998;111:532–535). Nevertheless, the disconnect between theory and practice, coupled with the hope that UC patients might benefit from this therapy, has led to initial investigations. The first trial was an industry sponsored, double-blinded, placebo-controlled trial of infliximab in patients with severe UC (Inflamm Bowel Dis 2001;8:83–88). Doomed to fail because of poor enrollment, 4 of 8 infliximab patients receiving either 5, 10, or 20 mg/kg responded by week 2 to a single dose of drug, versus none of the 3 placebo-treated patients. All 3 placebo and 4 of the 8 infliximab-treated patients subsequently underwent colectomy. A second randomized, placebo-controlled trial has appeared to date only in abstract form (Gastroenterology 2002;122:A99). Of the 42 patients with steroid-resistant UC, 8 of 22 (36%) patients receiving infliximab 5 mg/kg at weeks 0 and 2 were in remission at week 6, compared with 6 of 20 (30%) of those receiving placebo. Similar results were also seen in changes of quality of life scores between the 2 groups. Whether this failure of infliximab to perform much better than placebo was due to the drug or the patient group remains to be seen. Open-label published experience with infliximab in UC is also scarce but growing in number. Chey et al. reported an initial 8-patient experience (Am J Gastroenterol 2001;96:2373–2381) and then extended it to a total of 16 patients (Inflamm Bowel Dis 2001;7:S30–S33) with severe UC refractory to medical therapy. The initial enthusiasm over nearly “perfect” results, with clinical, endoscopic, and/or histological responses seen in all of the first 8, and subsequently 14 of 16 patients after a single infusion of infliximab at 5 mg/kg, were tempered by concerns over the patient population. Considerably older (mean age 60 years) than most IBD trials, the group included patients with previous colonic resections for diverticulitis, a sigmoid stricture, and in 1 patient, “Crohn’s-like” serology positive for anti-saccharomyces cerevisiae antibody (ASCA). Notable open-label results were also reported in a 6-patient experience of severe steroid-refractory UC, in which 4 patients enjoyed long-term (median follow-up 5.5 months) remissions after just a single 5 mg/kg infliximab infusion (Wien Klin Wochenschr 2001;113:930–933). Colonoscopic healing and improvement of histological parameters were also noted. In a separate study of 13 steroid-refractory (60 mg methylprednisolone daily ≥7 days), 10 of 13 (77%) patients responded to single-infusion 5 mg/kg of infliximab, and 90% maintained remission off steroids at a mean of 10.1 months (range, 5–12 months) (Dig Liver Dis 2002;34:626–630). A leading pediatric group reported short-term improvement in 7 of 9 (77%) children with moderate to severe UC, with 6 of 9 discontinuing steroids (J Pediatr Gastroenterol Nutr 2002;34:307–311). This is contrasted by our own unpublished experience at the University of Chicago, where none of 12 hospitalized, steroid-refractory severe UC patients treated soon after the commercial release of infliximab had a convincing response to the agent. A small open-label trial of infliximab therapy (5 mg/kg) in 8 uncontrolled UC patients (6 refractory to parenteral steroids) failed to prevent urgent colectomy in 4 (50%), and subsequently a fifth after a second infusion at week 5 (Dig Liver Dis 2002;34:631–634). How could there be such a disparity between experiences with this agent? It all may come back to the big “P”… placebo. One thing had become very clear in the age of biologics: the need for adequately powered, placebo-controlled trials. Perhaps in part caused by patient (and physician) expectations of dramatic responses, recent trials of biologics have been plagued by placebo-response rates approaching 50% (Gastroenterology 1997;113:383–389, Gastroenterology 2001;121:1088–1094). Placebo responses in UC trials tend to diminish over time (J Clin Gastroenterol 1993;16:215–218). Many of the fore-mentioned studies showed far less impressive numbers when the response rates were followed for greater than 6 months; the 25% sustained responses reported in the aforementioned studies (Inflamm Bowel Dis 2001;7:S30–S33, Dig Liver Dis 2002;34:631–634) are similar to those seen in the placebo-arm of the earlier Crohn’s infliximab retreatment trial (Gastroenterology 1999;117:761–769). Other therapies targeting TNF-α (CDP-571; thalidomide) have also had underwhelming efficacy in open-label treatment of UC (Aliment Pharmacol Ther 1997;11:1031–1035, Gastroenterology 2000;118:A582). The “power of the placebo,” together with the retrospective nature of many open-label experiences, the failure to use conventional endpoints to measure and define efficacy in others, nonstandardization of infusion doses and schedules, and the small sample sizes that accompany such trials, leaves us all suspended in an “uncontrolled purgatory.” Hopefully, we will get our “final answers” when the large multicenter, randomized, placebo-controlled trial of infliximab in UC that was recently started is completed. Identification of UC patient subsets (steroid-refractory, steroid-naı̈ve, etc.) may also be important predictors of long-term response to this agent. Stay tuned! ReplyGastroenterologyVol. 124Issue 7Preview Full-Text PDF