Disease indication
For two decades the primary endpoint asked how patients felt. It took an endoscope to find out whether the gut had actually healed.
Disease-modifying: Stop the transmural inflammation that erodes the bowel wall and eventually causes strictures and fistulas, judged on endoscopic healing rather than on how the patient reports feeling. Five distinct mechanisms have reached approval since 1998 — a real, if imperfect, translational success story.
TNF, leukocyte-integrin trafficking, IL-12/23 (p40), IL-23 (p19), and JAK inhibition. Fewer than rheumatoid arthritis' seven; far more than Parkinson's or DM1's zero.
Every disease-modifying programme below with a completed readout, not a cherry-picked subset. One of the four did not just fail to help — it made the disease worse.
Modern biologics can produce objective mucosal healing, and regulators increasingly require it as a co-primary alongside — not instead of — the symptom score.
A separate induction trial, then a re-randomised maintenance trial in induction responders, run out to a year — increasingly with an endoscopic co-primary layered onto the symptom score.
The Crohn's Disease Activity Index was the registrational endpoint for two decades, and it conflates two different questions: does the patient feel better, and has the gut actually healed. A drug can move the first without touching the second — which is exactly why regulators now require an endoscope to look.
| Endpoint | Type | Read at | Standing |
|---|---|---|---|
| Crohn's Disease Activity Index (CDAI) | Composite | 8–16 weeks | Contested |
| Endoscopic response / remission | Imaging | 12–52 weeks | Accepted |
| Histologic remission | Laboratory | 12–52 weeks | Exploratory |
| Fecal calprotectin | Laboratory | 2–8 weeks | Surrogate |
| Fistula response / closure | Clinician-rated | 10–54 weeks | Accepted |
| Steroid-free remission | Composite | 26–52 weeks | Accepted |
No model here gets a clean 'high' for predicting disease modification — the closest matches to human ileitis and colitis are only moderately predictive. That is a third pattern next to Parkinson's (models predict nothing) and rheumatoid arthritis (one model predicts almost everything): moderately good models, real approvals, and real high-profile failures alongside them.
| Model | FaceDoes it look like the disease? | ConstructDoes it arise the same way? | PredictiveHeal the gut |
|---|---|---|---|
Five distinct mechanisms reached approval — fewer than rheumatoid arthritis, far more than Parkinson's or DM1. The failures are not obscure: one of them made the disease worse.
| Intervention | Rested on | Primary endpoint | Reached | Outcome |
|---|---|---|---|---|
Upadacitinib AbbVie · JAK1-selective inhibition, blocking intracellular signalling downstream of several inflammatory cytokines at once | IL-10 KOT-cell transfer colitis | Clinical remission and endoscopic response, co-primary, at week 12 (induction) and week 52 (maintenance) | Approved U-EXCEL / U-EXCEED / U-ENDURE · 2023 | Approved |
Risankizumab AbbVie · Selective neutralisation of the interleukin-23 p19 subunit | IL-10 KO | Clinical remission (CDAI below 150) and endoscopic response, co-primary, at week 12 (induction) and week 52 (maintenance) | Approved ADVANCE / MOTIVATE / FORTIFY · 2022 | Approved |
Etrolizumab Genentech · Blockade of the β7 integrin subunit, affecting both α4β7 and αEβ7 | T-cell transfer colitis | Clinical remission and endoscopic improvement, co-primary | Phase 3 BERGAMOT · 2021 | Failed |
Ontamalimab (PF-00547659) Pfizer / Shire · Blockade of MAdCAM-1, the adhesion molecule that α4β7-expressing leukocytes bind to enter the gut | T-cell transfer colitis | Clinical remission and endoscopic response | Phase 3 2020 | Failed |
Ustekinumab Janssen · Neutralisation of the shared p40 subunit of interleukin-12 and interleukin-23 | IL-10 KOT-cell transfer colitis | Clinical response (CDAI reduction) at week 6 (induction) and remission at week 44 (maintenance) | Approved UNITI-1 / UNITI-2 / IM-UNITI · 2016 | Approved |
Vedolizumab Takeda · Blockade of α4β7 integrin, selectively disrupting leukocyte trafficking into the gut only | T-cell transfer colitis | Clinical remission (CDAI below 150) at week 6 (induction) and week 52 (maintenance) | Approved GEMINI 2 · 2014 | Approved |
Secukinumab Novartis · Neutralisation of interleukin-17A | TNBS colitis | Clinical response (CDAI reduction) at week 6 | Phase 2a Hueber et al. 2012 · 2012 | Failed |
Natalizumab Biogen · Blockade of α4 integrin, disrupting leukocyte trafficking into the gut and the central nervous system | T-cell transfer colitis | Clinical response (CDAI reduction) at week 10 | Approved ENACT-1 / ENACT-2 · 2008 | Approved |
Adalimumab Abbott · Neutralisation of tumour necrosis factor | TNBS colitisT-cell transfer colitis | Clinical remission (CDAI below 150) at week 4 (induction) and week 26 (maintenance) | Approved CLASSIC-I / CHARM · 2007 | Approved |
Fontolizumab PDL BioPharma · Neutralisation of interferon-gamma | TNBS colitis | Clinical response (CDAI reduction) | Phase 2 2006 | Failed |
Infliximab Centocor · Neutralisation of tumour necrosis factor | TNBS colitisT-cell transfer colitis | Clinical response (CDAI reduction) at week 4; later, fistula closure on compression in a dedicated fistula trial | Approved 1998 | Approved |
A composite of patient-reported symptoms and clinician-assessed general wellbeing
The conflation built into this one number — symptomatic relief and disease modification, added together and called a single score — is the reason the field now insists on looking at the bowel directly rather than trusting the diary alone.
A common disease, but not a cheap one to prove a claim in: modern programmes need an induction trial and a re-randomised maintenance trial, increasingly with a central-read endoscopic co-primary layered on top of the symptom score.
A separate induction programme, then a re-randomised maintenance trial in the responders, run out to a year — increasingly with a central-read endoscopic co-primary layered onto the symptom score. The structural readout is what costs the money, and it costs less than Parkinson's disease-modifying archetype because it reads out in weeks to a year, not years to years.
663 patient-years against 250 for a symptomatic programme. Cost grows more slowly than exposure because the fixed enrolment cost of each participant — screening, baseline imaging — is paid once however long they are then followed. The goal, not the molecule, still sets the budget.
Citedenrolment, duration and study counts, from named trials.Assumptionthe two cost parameters below.Computedeverything else.
| Phase | Design | N × studies | Duration | Patient-years | Modelled cost |
|---|---|---|---|---|---|
| Phase 1 | SAD / MAD, healthy volunteers Reads: Safety, tolerability and pharmacokinetics Biologic and JAK mechanisms carry infection-risk monitoring, so first-in-human work is not abbreviated. Conventional first-in-human designmoderate confidence | 50 | 9 mo | 38 | $1.4M–$2.2M |
| Phase 2 | Dose-ranging, clinical and early endoscopic signal Reads: Clinical response, with an early endoscopic read as a supporting signal Modern programmes increasingly look at the endoscope even this early, rather than waiting for Phase 3 to find out the symptom score was misleading. Biologic and JAK inhibitor Phase 2 programmes in Crohn's diseasemoderate confidence | 250 | 6 mo | 125 | $5.5M–$8.9M |
| Phase 3 · Induction | Two identically-designed induction pivotals Reads: Clinical remission and endoscopic response, co-primary, at 12 weeks Sized after ADVANCE/MOTIVATE and U-EXCEL/U-EXCEED: two induction trials run in parallel, both required to clear the co-primary bar before a patient can enter maintenance. D'Haens et al., Lancet 2022 (ADVANCE/MOTIVATE); Loftus et al., NEJM 2023 (U-EXCEL/U-EXCEED)moderate confidence | 400 × 2 | 3 mo | 200 | $14M–$21M |
| Phase 3 · Maintenance | Re-randomised maintenance trial in induction responders Reads: Clinical remission and endoscopic response, co-primary, at 52 weeks Only induction responders are re-randomised, so this trial is smaller than induction despite running four times as long — the year of follow-up, not the headcount, is what this phase is paying for. Sandborn et al., NEJM 2013 (GEMINI 2); FORTIFY maintenance designmoderate confidence | 300 | 1 yr | 300 | $9.6M–$16M |
| Total | 2.5 yr of clinical development, treating phases as sequential | 663 | $30M–$48M |
Model assumptions, not citations — the only free parameters here. The fixed component covers screening, baseline imaging and randomisation, incurred once per participant; the running component covers visits, monitoring and site fees for each year on study. Splitting them matters: a single rate per patient-year would price short symptomatic trials far below what pivotal trials actually cost.
Calibration: priced this way a single pivotal trial in this programme comes to $5.4M–$8.6M, against published pivotal trial costs of $12–33M (median $19M) across all indications. A single symptomatic induction trial here is short and small by design — steroids have never needed more than a few months to prove themselves — so it lands below the published range entirely. That range is set by the much larger, longer biologic and JAK programmes the disease-modifying archetype represents. Moore et al., JAMA Internal Medicine 2018.
Programme cost divided by a 13% likelihood of approval from Phase 1 in immune-mediated inflammatory disease (gastroenterology) — what one success costs once the failures are paid for. BIO / Informa / QLS, Clinical Development Success Rates — inflammatory bowel disease sits below the broader immunology average.
Read this one carefully: This page's ledger shows why: three late-stage failures on mechanisms one step removed from an already-approved one (etrolizumab, ontamalimab) plus one mechanism that actively backfired (secukinumab). A crowded field of validated targets has not made picking the next one much easier.
These are portfolio figures spanning all of R&D and already carry the cost of failure, so they are not comparable with the per-programme clinical costs above — they are the order of magnitude those costs roll up into.
The field is trying to retire the endpoint that got it this far, on the grounds that it was measuring the wrong thing for twenty years.
The CDAI is no longer accepted alone: FDA and EMA guidance now expects a clinical outcome plus objective endoscopic evidence of healing as co-primary endpoints. Trials designed on the old symptom-only standard, in retrospect, may have killed viable drugs on placebo-response noise rather than genuine inefficacy.
FDA 2022 guidance for industry, ulcerative colitis and Crohn's disease drug development
With TNF, integrin, IL-12/23, IL-23-selective and JAK blockade all approved, the field increasingly resembles rheumatoid arthritis' problem of matching mechanism to patient rather than finding a mechanism at all. Newer targets — anti-TL1A, S1P receptor modulation — are being tested against that established baseline rather than against nothing.
Late-stage Crohn's pipeline disclosures
Every approved mechanism targets inflammation; none is proven to reverse the fibrostenotic strictures that eventually send a large fraction of Crohn's patients to surgery regardless of how well their inflammation is controlled. The SAMP1/YitFc model's stricture phenotype points at a problem the approved drug class has not yet solved.
Surgical-recurrence literature in Crohn's disease; anti-fibrotic pipeline disclosures