Randomisation and masking

Patient details were entered onto a web-based password-protected site (hosted by Bangor University, UK), and allocated at random to infliximab or ciclosporin. A dynamic algorithm¹² was used to generate allocations on request, thus protecting against investigator preference or other subversion while ensuring that each trial group was balanced by centre, which was the only stratification used.

Since this was an open-label trial, local investigators and participants were aware of the treatment allocated, but the chief investigator and all analysts remained masked to allocation until the trial steering committee and data monitoring and ethics committee had reviewed and approved the analysis of the primary outcome.