Design of randomised controlled trials (RCTs) examining maintenance of clinical remission in inflammatory bowel disease (IBD) varies, with some trials re-randomising patients who have responded to active drug during induction to either active drug or placebo and others treating patients through with active drug or placebo from baseline. Whether this influences therapeutic gain of drug over placebo is unknown.
We searched the literature to January 2023 for maintenance of remission trials of biologics or small molecules versus placebo in IBD. We extracted maintenance of remission rates according to trial design; either trials re-randomising patients or trials treating patients through. We pooled data in a meta-analysis for all patients, and according to type of IBD. We calculated the number needed to treat (NNT), with a 95% confidence interval (CI), to assess therapeutic gain of active drug over placebo according to trial design.
We identified 37 maintenance of remission trials (12,075 patients). Rates of maintenance of clinical remission were higher (41.9% with active drug, versus 20.3% with placebo), and NNT lowest (5; 95% CI 4-6), in trials re-randomising patients compared with those treating through (maintenance of remission rate 30.9% with active drug versus 14.6% with placebo, NNT = 7; 95% CI 5-9). Results were similar when trials were analysed according to IBD type but were more marked in ulcerative colitis RCTs (maintenance of remission rates in re-randomised trials 39.4% with active drug versus 17.8% with placebo, NNT = 5; 95% CI 3-7; treat-through trials 27.3% with active drug versus 11.9% with placebo, NNT = 7; 95% CI 5-11.5).
Trials re-randomising patients had generally higher maintenance of remission rates, lower NNTs, and greater therapeutic gains over placebo.

Copyright © 2023 The Author(s). Published by Elsevier Ltd.. All rights reserved.