highly significantp=0.0008
Reassuringly, the sensitivity analysis in which all the patients who were randomised and dosed (ie, the conventional intention-to-treat population) were analysed (with patients excluded from the mITT population set to non-responder imputation) still showed a highly significant effect of budesonide MMX 9 mg versus placebo on combined clinical and endoscopic remission rates (15.1% vs 3.1%, respectively, p=0.0008).