Reply

Abstract

We agree that the disparities among published reports in infliximab therapy for patients with UC may be attributed by the phenomenon of nonplacebo, nonrandomized, controlled studies. The final results of the ongoing industry-sponsored multicenter, randomized, placebo-controlled trials (called “ACT I” and “ACT II”) should clarify this uncertainty once and for all by proving or disproving the efficacy of infliximab as a therapy for UC. The disparities in the data from various studies in this subject matter and the difficulty in interpreting these results raise an interesting dilemma that we often face. Because randomized-controlled trials (RCTs) are expensive and may even be difficult to complete, as illustrated by Dr. Sands’ experience with the initial trial (Inflamm Bowel Dis 2001;8:83–88), open-labeled or even observational studies are almost always a prerequisite for an RCT. An important step in this process is the interpretation of the preliminary results. Although a higher success rate in the preliminary study should translate to a greater efficacy and an increased likelihood of a “positive” study (i.e., drug more effective than placebo) in an RCT, such an assumption may require a leap of faith. Clinical studies may vary in patient selection, treatment and follow-up regimen, and definition of outcomes. An equally important consideration is that the placebo response has been reported to reach nearly 50% in some studies. It is also recognized that this problem plagues research performed in patients with CD as well. We have shown in a recent meta-analysis of previous RCTs for active CD that the placebo remission rate is approximately 18% and that the study duration is an important factor influencing the placebo remission rate (Gastroenterology 2002;122:A497). We are in full agreement that a placebo-controlled study is required to assess the efficacy of infliximab therapy in patients with UC, and that placebo-controlled RCTs are necessary to prove drug efficacy. At the same time, we argue that inflammatory bowel disease investigators need to understand better how features of study design may influence the study outcome and the placebo response. Such knowledge will enhance our ability to interpret preliminary data before committing enormous manpower and cost into an RCT. We agree that the disparities among published reports in infliximab therapy for patients with UC may be attributed by the phenomenon of nonplacebo, nonrandomized, controlled studies. The final results of the ongoing industry-sponsored multicenter, randomized, placebo-controlled trials (called “ACT I” and “ACT II”) should clarify this uncertainty once and for all by proving or disproving the efficacy of infliximab as a therapy for UC. The disparities in the data from various studies in this subject matter and the difficulty in interpreting these results raise an interesting dilemma that we often face. Because randomized-controlled trials (RCTs) are expensive and may even be difficult to complete, as illustrated by Dr. Sands’ experience with the initial trial (Inflamm Bowel Dis 2001;8:83–88), open-labeled or even observational studies are almost always a prerequisite for an RCT. An important step in this process is the interpretation of the preliminary results. Although a higher success rate in the preliminary study should translate to a greater efficacy and an increased likelihood of a “positive” study (i.e., drug more effective than placebo) in an RCT, such an assumption may require a leap of faith. Clinical studies may vary in patient selection, treatment and follow-up regimen, and definition of outcomes. An equally important consideration is that the placebo response has been reported to reach nearly 50% in some studies. It is also recognized that this problem plagues research performed in patients with CD as well. We have shown in a recent meta-analysis of previous RCTs for active CD that the placebo remission rate is approximately 18% and that the study duration is an important factor influencing the placebo remission rate (Gastroenterology 2002;122:A497). We are in full agreement that a placebo-controlled study is required to assess the efficacy of infliximab therapy in patients with UC, and that placebo-controlled RCTs are necessary to prove drug efficacy. At the same time, we argue that inflammatory bowel disease investigators need to understand better how features of study design may influence the study outcome and the placebo response. Such knowledge will enhance our ability to interpret preliminary data before committing enormous manpower and cost into an RCT.