Sirs, We read the 3-year follow-up to the randomized controlled trial (RCT) of infliximab rescue therapy in severe to moderately severe ulcerative colitis by Gustavsson et al. with interest.1 The authors reported that the impressive initial effect of infliximab rescue therapy compared with placebo was still present, in terms of a lower rate of colectomy. This is an important study, and the authors are to be commended for the completeness of the dataset they present. However, there are some limitations.
First, the choice of subsequent maintenance therapy between 3 and 36 months was undefined and nonstandardized. As the authors conceded in their discussion that ‘the 3-year colectomy rate is more dependant on the maintenance therapy than on a single infusion of infliximab as rescue therapy’, their conclusion that ‘the favourable short term response (of infliximab rescue therapy) remains in a long term perspective’ is only true in so far as it is still possible after 3 years to discern the initial impact of infliximab. All of the benefit of infliximab compared with placebo was realized in the initial RCT,2 with identical proportions of patients requiring colectomy over the subsequent 33 months (29% from each treatment arm). The study population clearly comprised patients with severe disease, as in total 62% underwent colectomy during the entire 3 years of the study; yet there was no standardized use of immunomodulators after rescue therapy. In fact, there was a trend towards a higher use of azathioprine in the infliximab arm compared with the placebo arm (odds ratio = 2.25). This was not anticipated or controlled for in the study design, and may have altered the outcomes between the two arms of the trial during follow-up.
Secondly, two instances of unreported data represent unfortunate missed opportunities. The study found that presence of mucosal healing at 3 months predicted future risk of colectomy, a finding that is corroborated by a previous population-based study.3 It would therefore have been valuable to the clinician to report mucosal healing data by original treatment arm, as healing appears to be associated with outcome, and may therefore be used to predict the likelihood of colectomy among patients who receive infliximab who do not achieve this goal. In addition, validated quality of life questionnaires were completed, providing broader outcome measures than simply colectomy vs. no colectomy, with an excellent response rate (41/45). The authors reported that these showed no difference between the two treatment arms, but the opportunity was missed to present the data by colectomy status at end of follow-up. This would have relevance for both clinicians and patients deliberating over whether further attempts at medical therapy or a colectomy is best in initial responders to infliximab who develop subsequent subacute ulcerative colitis, who would not meet criteria for further infliximab in the UK currently,4 in terms of holistic health benefit.