DIVERSITYCD · Induction and maintenance
Efficacy and safety of filgotinib as induction and maintenance therapy for Crohn's disease (DIVERSITY): a phase 3, double-blind, randomised, placebo-controlled trial
Vermeire S, et al.
Lancet Gastroenterol Hepatol 2025 · PMID 39637881 (PubMed, opens in a new tab) · doi 10.1016/S2468-1253(24)00272-3 (full paper, opens in a new tab)
| Outcome | Filgotinib 200 mg | Placebo | Estimate (confidence interval) | P |
|---|---|---|---|---|
| A: PRO2 remission, Induction study A · week 10 (primary) | 32.9% | 25.7% | +6.9 (95% CI −1.4 to 15.2) | P=0.0963 |
| A: endoscopy, Induction study A · week 10 · endoscopic response, SES-CD (primary) | 23.9% | 18.1% | +5.5 (95% CI −2.0 to 12.9) | P=0.1365 |
| B: PRO2 remission, Induction study B · week 10 (primary) | 29.7% | 17.9% | +11.9 (95% CI 3.7–20.2) | P=0.0039 |
| B: endoscopy, Induction study B · week 10 · endoscopic response, SES-CD (primary) | 11.9% | 11.4% | +0.1 (95% CI −6.5 to 6.6) | P=0.9797 |
| Maintenance: PRO2 remission, Week 58 (primary) | 43.8% | 26.4% | +16.8 (95% CI 2.0–31.6) | P=0.0382 |
| Maintenance: endoscopy, Week 58 · endoscopic response, SES-CD (primary) | 30.4% | 9.4% | +20.6 (95% CI 8.2–33.1) | P=0.0038 |
Three of the four induction intervals cross zero; both maintenance intervals, in patients who had already responded, sit above it.
Source: Vermeire 2025, abstract. PRO2 remission: abdominal pain subscore ≤1 and liquid or very soft stool frequency ≤3. Endoscopic response: SES-CD down ≥50%. Study A: biologic-naive and later biologic-experienced; study B: biologic-experienced.
Structured abstract
BackgroundFilgotinib, a JAK1-preferential inhibitor, works in UC; whether it induces and maintains both symptomatic and endoscopic improvement in Crohn's disease had not been tested in phase 3.
MethodsPhase 3, double-blind, randomised (1:1:1), placebo-controlled: two 11-week induction studies and a maintenance study to week 58 with 2:1 re-randomisation (371 centres, 39 countries). 1,372 patients. Primary endpoint: Co-primary at week 10 (induction) and week 58 (maintenance): PRO2 clinical remission (abdominal pain subscore ≤1 and liquid or very soft stool frequency subscore ≤3) and endoscopic response (SES-CD decrease ≥50%).
ResultsPrimary result. Filgotinib 200 mg at week 10: PRO2 remission 32.9% vs 25.7% in induction A (difference 6.9%, 95% CI −1.4 to 15.2; P=0.0963) and 29.7% vs 17.9% in induction B (11.9%, 95% CI 3.7–20.2; P=0.0039); endoscopic response 23.9% vs 18.1% (5.5%, 95% CI −2.0 to 12.9; P=0.1365) and 11.9% vs 11.4% (0.1%, 95% CI −6.5 to 6.6; P=0.9797). Week 58: PRO2 remission 43.8% vs 26.4% (16.8%, 95% CI 2.0–31.6; P=0.0382); endoscopic response 30.4% vs 9.4% (20.6%, 95% CI 8.2–33.1; P=0.0038).
Take-homeFilgotinib failed its week-10 induction endpoints in Crohn's disease, unlike upadacitinib in its own phase 3 trials, so JAK inhibitors should not be assumed to work equally well in Crohn's disease.
| Patients | 1,372 |
|---|---|
| Population | Age 18–75 with moderately to severely active Crohn's disease for at least 3 months. Induction A: biologic-naive and later biologic-experienced patients (707). Induction B: biologic-experienced patients (665). 481 entered the maintenance study. |
| Comparison | Filgotinib 200 mg or 100 mg once daily vs placebo; filgotinib patients with week-10 PRO2 remission or endoscopic response were re-randomised to their induction dose or placebo |
| Drugs | filgotinib |
| Classes | JAK inhibitor |
| Published | Vermeire S, et al. Lancet Gastroenterol Hepatol 2025;10:138–53. |
Key points
- Filgotinib 200 mg did not meet the co-primary endpoints at week 10 but did meet both at week 58.
- Filgotinib 100 mg met no co-primary endpoint in any study.
- Serious adverse events in induction A: 8% (200 mg), 7% (100 mg) and 6% (placebo); no deaths.
- Filgotinib is not available in Canada (see SELECTION).
Limitations
- Maintenance used randomised withdrawal of filgotinib responders, which flatters maintenance results.
- Long recruitment (2016–2022).
- Not available in Canada.
Commentary
Context. DIVERSITY is the Crohn’s disease programme for filgotinib, a JAK1-preferential inhibitor that works in UC (SELECTION). Induction did not separate clearly from placebo, especially on endoscopy, although maintenance results were positive in patients who had already responded.
Why it matters in Canada. The drug itself is not available here, but the trial is a useful contrast with upadacitinib, which met its clinical and endoscopic co-primary endpoints in Crohn’s disease in U-EXCEL, U-EXCEED and U-ENDURE. Comparing the two is cross-trial, but it is a reminder that efficacy in UC does not guarantee efficacy in Crohn’s disease, and that endoscopic response is a harder target in Crohn’s disease.
Journal club
Design
Phase 3, double-blind, randomised (1:1:1), placebo-controlled: two 11-week induction studies and a maintenance study to week 58 with 2:1 re-randomisation (371 centres, 39 countries). 1,372 patients.
Primary result. Filgotinib 200 mg at week 10: PRO2 remission 32.9% vs 25.7% in induction A (difference 6.9%, 95% CI −1.4 to 15.2; P=0.0963) and 29.7% vs 17.9% in induction B (11.9%, 95% CI 3.7–20.2; P=0.0039); endoscopic response 23.9% vs 18.1% (5.5%, 95% CI −2.0 to 12.9; P=0.1365) and 11.9% vs 11.4% (0.1%, 95% CI −6.5 to 6.6; P=0.9797). Week 58: PRO2 remission 43.8% vs 26.4% (16.8%, 95% CI 2.0–31.6; P=0.0382); endoscopic response 30.4% vs 9.4% (20.6%, 95% CI 8.2–33.1; P=0.0038).
Validity and applicability
- Maintenance used randomised withdrawal of filgotinib responders, which flatters maintenance results.
- Long recruitment (2016–2022).
- Not available in Canada.
Where it sits now
Filgotinib failed its week-10 induction endpoints in Crohn's disease, unlike upadacitinib in its own phase 3 trials, so JAK inhibitors should not be assumed to work equally well in Crohn's disease.
- SELECTION (2021): Filgotinib 200 mg induces and maintains remission in UC, but it is not available in Canada; it matters here mainly as evidence for the JAK class.
- U-EXCEL, U-EXCEED and U-ENDURE (2023): Oral upadacitinib induces and maintains clinical remission and endoscopic response in Crohn's disease, including after biologic failure, with JAK-class safety issues to monitor.
- RELIEVE UCCD (CD) (2026): Duvakitug, an anti-TL1A antibody, gave a dose-related endoscopic response signal in Crohn's disease in phase 2b; phase 3 trials will decide whether TL1A becomes a new Crohn's disease target.
- GALAXI-2 and GALAXI-3 (2025): IV guselkumab induction followed by SC maintenance is effective in Crohn's disease and outperformed ustekinumab at week 48 on several endpoints.
Questions for discussion
- Filgotinib works in UC but not convincingly as Crohn's induction. What biological or trial-design reasons might explain the difference?
- Maintenance success came in a randomised-withdrawal population of week-10 responders. How should that be read alongside the failed induction results?
- Endoscopic response in induction study B was 11.9% vs 11.4%. What does that say about relying on symptom endpoints alone in Crohn's disease?