Colon-in-Continuity: Choosing a GLP-2 Analog Before There Is a Clear Winner
A single patient, eighteen months out from her last resection and finally stable enough to think past survival. The disagreement isn’t about whether she needs a GLP-2 analog — it’s about which one, when neither drug’s own trial was built to answer that question for her specific anatomy.
Renata O., a 44-year-old hairstylist, has run her own single-chair salon out of a converted garage for eleven years, booked solid most weeks by clients who’ve followed her through three different storefronts. Crohn’s disease, diagnosed at 26, took most of her jejunum and ileum across three resections, the last of them eighteen months ago after a fistulizing recurrence that nearly cost her the rest of her small bowel; what remains is roughly 85 centimeters of jejunum anastomosed to an intact colon. She is finally off the wound-care regimen and back behind the chair four days a week, standing through most of a shift on legs that still tire faster than they used to. She has been on 10 liters of parenteral support weekly since the last surgery, infusing five nights out of seven, and the infusion pole is the one part of her routine she has never stopped hoping to shrink.
Her surgeon's own colon-in-continuity anatomy is exactly what complicates the drug choice in front of her: teduglutide, the only GLP-2 analog actually approved for this indication, showed materially smaller benefit in its own pivotal trials among patients with a preserved colon than among those with an end jejunostomy — a real, measured gap, not a theoretical one, since a functioning colon already recovers some fluid and calorie absorption on its own, leaving less room for a drug built to enhance exactly that process to add further gain. Real-world cohort data reported at UEG Week since have pushed back against reading that gap as a ceiling, suggesting CIC patients can still respond meaningfully outside a trial's stricter entry criteria — but a cohort report is a weaker form of evidence than the randomized trial it's arguing against. The investigational alternative, apraglutide, met its overall primary endpoint in the phase 3 STARS trial, a genuinely larger and more current dataset than teduglutide's own approval trials — but its own CIC-specific comparison against placebo, the exact subgroup Renata sits in, did not reach statistical significance (-25.2% vs -17.6% weekly parenteral-support-volume reduction, p=0.179), even though the point estimate ran close to the same magnitude that was significant in patients without a colon. And apraglutide remains unapproved: the FDA required a confirmatory phase 3 trial in 2025 after its own pharmacokinetic review found the drug exposure actually delivered in STARS ran lower than planned, so access today means enrolling in the STARS Extend long-term study, not writing a prescription.
Intestinal failure clinic, eighteen months post-resection
I want to start teduglutide today. She is eighteen months out, stable, off wound care, and every week she waits is a week of five-night infusions and everything that comes with them — catheter risk, sleep disruption, a pole she has to plan her whole schedule around. The CIC efficacy gap in the pivotal trials is real, but it’s a gap, not an absence — the trials still showed benefit in colon-in-continuity patients, just a smaller one, and the UEG real-world cohort data since then are consistent with that smaller-but-real effect holding up outside trial conditions.
If she had an end jejunostomy instead of a preserved colon, I don’t think we’d be having three separate opinions about this — the trial data there is unambiguous. It’s specifically her anatomy that opens the door to the other two positions.
I’m not arguing teduglutide is wrong for her — I’m arguing we shouldn’t treat the apraglutide question as closed. Look at what the CIC subgroup in STARS actually showed: a 25.2% reduction versus 17.6% on placebo. That’s a real point estimate in the same range as the stoma subgroup’s significant result. It missed significance at p=0.179, in a subgroup of 83 patients, with what Ironwood’s own explanation calls an unusually strong placebo response — patients coached to reduce their own PS use as part of the trial protocol, which will always narrow a drug-versus-placebo gap in exactly the population that has the most residual gut left to work with.
A trial’s own p-value is not a verdict on biological truth, especially at this sample size — it’s a threshold. “Did not reach significance” and “doesn’t work” are not the same sentence, and I don’t think we should let them collapse into each other just because one is easier to act on.
Both of those arguments are about the drugs. I want to bring in what this actually costs her to live through. Enrolling in STARS Extend means a research-site relationship, visit schedules, and a drug she cannot simply fill at her local pharmacy if her life gets complicated — and she runs a one-person business where a missed Tuesday is a missed Tuesday’s income. Starting teduglutide today means a daily injection she administers herself and a drug that’s been through twelve years of ordinary clinical use, extension trial or not.
I don’t think the pharmacology alone should decide this. If she starts teduglutide now and it works even modestly, we haven’t closed any door — nothing about a confirmatory apraglutide trial succeeding later would make switching to it impossible. But if she waits for a drug that may not clear its own confirmatory bar, on a subgroup analysis that didn’t reach significance in exactly her anatomy, she’s spent that time on the infusion pole for a possibility, not a probability.
Agreed: teduglutide starts this month, with a formal reassessment of parenteral-support volume at 6 months against her own pre-treatment baseline, not against the pivotal trial’s pooled result.
Not agreed, and left explicit rather than smoothed over: whether this closes the apraglutide question or only defers it. The pharmacologist wants her flagged for the confirmatory apraglutide trial once it opens enrollment, regardless of how teduglutide performs; the nutrition-support physician sees no reason to plan around a trial that doesn’t exist yet.