ADMIRE CD IIPerianal CD · Other
Darvadstrocel in Patients With Crohn's Disease With Complex Perianal Fistulas: The ADMIRE CD II Phase 3 Randomized Trial
Colombel JF, et al.
Gastroenterology 2026 · PMID 41790076 (PubMed, opens in a new tab) · doi 10.1053/j.gastro.2025.12.033 (full paper, opens in a new tab)
| Outcome | Darvadstrocel | Placebo | Estimate (confidence interval) | P |
|---|---|---|---|---|
| ADMIRE CD II, Primary endpoint · 568 patients (primary) | 48.8% (138/283) | 46.3% (132/285) | +2.4 (95% CI −5.8 to 10.6) | P=0.571 |
| ADMIRE-CD (2016), Earlier trial · 212 patients | 50% (53/107) | 34% (36/105) | +15.2 (97.5% CI 0.2–30.3) | P=0.024 |
The earlier result sat just above zero; in the larger trial nearly half of placebo patients reached remission and the difference is centred near zero.
Source: Colombel 2026, abstract (ADMIRE CD II, 95% CI); Panés 2016, abstract (ADMIRE-CD, intention to treat, 97.5% CI). Both arms of ADMIRE CD II had curettage and closure of the internal opening.
Structured abstract
BackgroundADMIRE-CD had suggested that darvadstrocel, an allogeneic stem-cell injection, heals complex perianal fistulas; this confirmatory trial tested it in a larger, broader population that included North America.
MethodsPhase 3, double-blind, randomised, placebo-controlled trial (Europe, Israel and North America). 568 patients. Primary endpoint: Combined remission at week 24: closure of all treated external openings and no collections larger than 2 cm.
ResultsPrimary result. 48.8% (138/283) vs 46.3% (132/285); estimated treatment difference 2.4% (95% CI −5.8 to 10.6); P=0.571.
Take-homeIn the larger confirmatory trial, darvadstrocel was no better than placebo for complex perianal fistulas when both arms had surgical preparation.
| Patients | 568 |
|---|---|
| Population | Adults with Crohn's disease and complex perianal fistulas (no more than 2 internal and 3 external openings) with inadequate response to immunosuppressants or biologics. |
| Comparison | Darvadstrocel (120 × 10⁶ cells in 24 mL) vs placebo; both groups had curettage and closure of the internal opening |
| Drugs | darvadstrocel |
| Classes | Cell therapy |
| Published | Colombel JF, et al. Gastroenterology 2026;170:1562–70. |
Key points
- Key secondary endpoints were also negative: clinical remission at week 24 (P=0.515) and time to clinical remission (P=0.374).
- Treatment-emergent adverse events were similar: 73.0% (203/278) vs 73.4% (201/274).
- Nearly half of placebo patients reached combined remission, showing that curettage and closure of the internal opening drive much of the healing.
- The EU marketing authorisation for darvadstrocel was withdrawn on 13 December 2024; it is not listed in Health Canada's Drug Product Database.
Limitations
- 24-week endpoint.
- Both arms had active surgical treatment, so the trial tests the added value of the cells, not of surgery.
Commentary
ADMIRE CD II is a good teaching example of a promising phase 3 result (ADMIRE-CD: combined remission 50% vs 34%, difference 15.2%, 97.5% CI 0.2–30.3) not holding up in a larger, broader trial. Its high placebo response reinforces the value of good surgical management of perianal fistulas, which is also the lesson of PISA-II.
Journal club
Design
Phase 3, double-blind, randomised, placebo-controlled trial (Europe, Israel and North America). 568 patients.
Primary result. 48.8% (138/283) vs 46.3% (132/285); estimated treatment difference 2.4% (95% CI −5.8 to 10.6); P=0.571.
Validity and applicability
- 24-week endpoint.
- Both arms had active surgical treatment, so the trial tests the added value of the cells, not of surgery.
Where it sits now
In the larger confirmatory trial, darvadstrocel was no better than placebo for complex perianal fistulas when both arms had surgical preparation.
- ADMIRE-CD (2016): ADMIRE-CD suggested benefit from stem-cell injection for complex perianal fistulas, but the larger ADMIRE CD II was negative and the product has been withdrawn in the EU and was never approved in Canada.
- PISA-II (2022): For a high perianal fistula with a single internal opening, surgical closure after seton and short anti-TNF gives more MRI healing and fewer reinterventions than anti-TNF alone, and MRI healing predicts durable closure.
Questions for discussion
- ADMIRE-CD's lower 97.5% confidence bound was 0.2 points. What does a result that close to zero predict about replication?
- Nearly half of placebo patients reached combined remission. What does that say about surgical preparation, and how should it shape what you offer a patient with a complex fistula?
- Combined remission needed both clinical closure and MRI. Which matters more to patients, and how should fistula trials define success?