GEMINI 2CD · Induction and maintenance
Vedolizumab as induction and maintenance therapy for Crohn's disease
Sandborn WJ, et al.
N Engl J Med 2013 · PMID 23964933 (PubMed, opens in a new tab) · doi 10.1056/NEJMoa1215739 (full paper, opens in a new tab)
| Endpoint | Vedolizumab | Every 8 weeks | Every 4 weeks | Placebo | Difference |
|---|---|---|---|---|---|
| Clinical remission, Week 6 · induction, CDAI ≤150 (primary endpoint) | 14.5% | not in this comparison | not in this comparison | 6.8% | P=0.02 |
| CDAI-100 response, Week 6 · induction (primary endpoint) | 31.4% | not in this comparison | not in this comparison | 25.7% | P=0.23 (not significant) |
| Clinical remission, Week 52 · randomised-withdrawal maintenance (primary endpoint) | not in this comparison | 39.0% | 36.4% | 21.6% | P<0.001 (every 8 weeks) and P=0.004 (every 4 weeks) vs placebo |
The week-6 effect was small; maintenance, in patients who had already responded, showed the clearer benefit.
Source: Sandborn 2013, abstract. The abstract gives no counts or confidence intervals.
Structured abstract
BackgroundVedolizumab, a gut-selective anti-α4β7 integrin antibody, was being tested in UC in GEMINI 1; whether it worked in Crohn's disease was unknown.
MethodsIntegrated phase 3 study: randomised, double-blind, placebo-controlled induction trial plus randomised-withdrawal maintenance trial. 1,115 patients. Primary endpoint: Induction: clinical remission (CDAI ≤150) and CDAI-100 response at week 6. Maintenance: clinical remission at week 52.
ResultsPrimary result. Week 6 remission 14.5% vs 6.8% (P=0.02); CDAI-100 response 31.4% vs 25.7% (P=0.23, not significant). Week 52 remission 39.0% (every 8 weeks) and 36.4% (every 4 weeks) vs 21.6% with placebo (P<0.001 and P=0.004).
Take-homeVedolizumab works in Crohn's disease but slowly: modest at week 6, clearer by week 10 and in maintenance, so it suits patients who can wait for the effect.
| Patients | 1,115 |
|---|---|
| Population | Adults aged 18–80 with moderately to severely active Crohn's disease and inadequate response, loss of response or intolerance to immunomodulators, TNF antagonists or corticosteroids. Induction: 368 randomised (cohort 1) plus 747 open-label (cohort 2). Maintenance: 461 responders randomised. |
| Comparison | Vedolizumab 300 mg IV vs placebo for induction and maintenance |
| Drugs | vedolizumab |
| Classes | Anti-integrin |
| Published | Sandborn WJ, et al. N Engl J Med 2013;369:711–21. |
Key points
- Induction at weeks 0 and 2; responders were randomised to vedolizumab every 8 weeks, every 4 weeks, or placebo until week 52.
- Serious adverse events 24.4% vs 15.3%; infections 44.1% vs 40.2%; serious infections 5.5% vs 3.0% (vedolizumab vs placebo).
- Antibodies against vedolizumab developed in 4.0% of patients.
- Dosing every 4 weeks was no better than every 8 weeks in maintenance.
- GEMINI 3 (anti-TNF failures): week 6 remission was not significant (15.2% vs 12.1%; P=0.433) but became apparent by week 10 (26.6% vs 12.1%; nominal P=0.001) (PMID 24859203).
Limitations
- Short (6-week) induction assessment for a slow-acting drug.
- Symptom-based endpoints (CDAI) without endoscopic outcomes.
- Mixed population of anti-TNF-naive and anti-TNF-experienced patients.
Commentary
GEMINI 2 led to the approval of vedolizumab for Crohn’s disease. Its gut-selective mechanism (blocking α4β7 integrin) makes it attractive for older patients or those with infection or malignancy concerns, but the slow onset means a bridge (often steroids) is commonly needed.
Vedolizumab is also the only drug with a randomised trial in chronic pouchitis (EARNEST), and it reduced endoscopic recurrence after ileocolonic resection in REPREVIO.
Related. GEMINI 3 tested three-dose induction after anti-TNF failure, and VISIBLE 2 showed that, after two IV doses, subcutaneous vedolizumab 108 mg every 2 weeks maintains remission in Crohn’s disease (48.0% vs 34.3% with placebo at week 52).
Journal club
Design
Integrated phase 3 study: randomised, double-blind, placebo-controlled induction trial plus randomised-withdrawal maintenance trial. 1,115 patients.
Primary result. Week 6 remission 14.5% vs 6.8% (P=0.02); CDAI-100 response 31.4% vs 25.7% (P=0.23, not significant). Week 52 remission 39.0% (every 8 weeks) and 36.4% (every 4 weeks) vs 21.6% with placebo (P<0.001 and P=0.004).
Validity and applicability
- Short (6-week) induction assessment for a slow-acting drug.
- Symptom-based endpoints (CDAI) without endoscopic outcomes.
- Mixed population of anti-TNF-naive and anti-TNF-experienced patients.
Where it sits now
Vedolizumab works in Crohn's disease but slowly: modest at week 6, clearer by week 10 and in maintenance, so it suits patients who can wait for the effect.
- EARNEST (2023): Vedolizumab is the first advanced therapy shown in a randomised trial to induce remission in chronic pouchitis.
- REPREVIO (2025): Starting vedolizumab within 4 weeks of ileocolonic resection reduces early endoscopic recurrence in patients at risk.
- GEMINI 3 (2014): After anti-TNF failure in Crohn's disease, vedolizumab missed its week-6 remission endpoint but showed benefit by week 10, so judge response later than with faster-acting drugs.
- VISIBLE 2 (2022): After two IV doses, vedolizumab 108 mg SC every 2 weeks maintains remission in Crohn's disease better than placebo, so patients can switch to home injections.
Questions for discussion
- One week-6 co-primary endpoint was met and the other was not. How should you interpret a trial with split co-primary results?
- Induction was two doses judged at week 6. How did GEMINI 3 change the induction schedule and the time point for judging response?
- Serious adverse events were 24.4% vs 15.3%. How would you explain that to a patient, given vedolizumab's reputation for safety?