Comparative efficacy and safety of biologic therapies for moderate-to-severe Crohn's disease: a systematic review and network meta-analysis

Siddharth Singh, M. Hassan Murad, Mathurin Fumery, Rocio Sedano, Vipul Jairath, Remo Panaccione, William J. Sandborn, Christopher Ma

Research output: Contribution to journalArticlepeer-review


Background: Data are needed to inform the positioning of biologic therapy in the treatment of moderate-to-severe Crohn's disease, both first line and after previous biologic exposure. We aimed to assess the comparative efficacy and safety of biologics in patients with Crohn's disease. Methods: We did a systematic review and network meta-analysis of phase 2 and phase 3 randomised controlled trials done in adults (≥18 years) with moderate-to-severe Crohn's disease (Crohn's Disease Activity Index [CDAI] 220–450) treated with tumour necrosis factor (TNF) antagonists, anti-integrin, anti-interleukin (IL)-12 and IL-23p40, or anti-IL23p19 agents, either alone or in combination with immunosuppressants, as their first-line biologic or after previous biologic exposure, compared with placebo or an active comparator. The minimum duration of therapy was 14 days for trials reporting induction of remission in active disease and 22 weeks in trials reporting maintenance of remission. We searched Medline, EMBASE, the Cochrane CENTRAL Register of Controlled Trials, conference proceedings, trial registries, and unpublished data from inception to June 3, 2021, without any language restrictions. Summary estimates of the primary and secondary outcomes were extracted from the published reports; individual patient-level data were not sought. The primary endpoint was induction of clinical remission in patients with active disease (CDAI <150) and maintenance of remission in patients with response to induction therapy, with data extracted from published reports. A network meta-analysis with multivariate consistency model random-effects meta-regression was done, with rankings based on surface under the cumulative ranking curve (SUCRA) values. Findings: The search strategy yielded 18 382 citations, of which 31 trials were eligible for inclusion. On the basis of 15 randomised controlled trials including 2931 biologic-naive patients, infliximab monotherapy (odds ratio [OR] 4·53 [95% CI 1·49–13·79]), infliximab combined with azathioprine (7·49 [2·04–27·49]), adalimumab (3·01 [1·25–7·27]), and ustekinumab (2·63 [1·10–6·28]) were associated with significantly higher odds of inducing remission compared to certolizumab pegol (all moderate confidence); infliximab and azathioprine combination therapy was also associated with significantly higher odds of inducing remission than vedolizumab (3·76 [1·01–14·03]; low confidence). On the basis of ten randomised controlled trials including 2479 patients with previous biologic exposure, adalimumab after loss of response to infliximab (OR 2·82 [95% CI 1·20–6·62]; low confidence), and risankizumab (2·10 [1·12–3·92]; moderate confidence), were associated with higher odds of inducing remission than vedolizumab. No differences between active interventions were observed in maintenance trials. Most trials were at low or uncertain risk of bias. Interpretation: Although biologic treatment choices in patients with moderate-to-severe Crohn's disease must be individualised for each patient, this analysis suggests that either infliximab with azathioprine or adalimumab might be preferred as a first-line therapy, and adalimumab (after infliximab loss of response) or risankizumab might be preferred as a second-line therapy, for induction of clinical remission. Funding: None.

Original languageEnglish (US)
Pages (from-to)1002-1014
Number of pages13
JournalThe Lancet Gastroenterology and Hepatology
Issue number12
StatePublished - Dec 2021

ASJC Scopus subject areas

  • Hepatology
  • Gastroenterology


Dive into the research topics of 'Comparative efficacy and safety of biologic therapies for moderate-to-severe Crohn's disease: a systematic review and network meta-analysis'. Together they form a unique fingerprint.

Cite this