Larazotide: what happened to the coeliac drug that reached Phase 3

Larazotide's flagship trial did not work. The Phase 3 was stopped at its interim analysis in June 2022, after more than 900 coeliac patients had been through controlled trials that produced a single positive primary result, at the lowest of three doses. Its developer filed for Chapter 7 liquidation a year later.

What it is

Larazotide acetate, first called AT-1001, is an eight-amino-acid peptide taken by mouth. Its structure was derived from a protein made by the cholera bacterium, Vibrio cholerae [1]. The idea behind it: in coeliac disease, gluten fragments slip between the cells lining the small intestine by loosening the tight junctions that seal them together, and the immune reaction follows. A drug that kept those junctions shut from inside the gut might blunt the reaction, for people who still get symptoms despite a gluten-free diet or who are accidentally exposed.

That's a mechanism theory. It is reasonable, it comes from years of work on intestinal permeability, and it's what the trials below set out to test.

The human evidence

Larazotide has more human data than almost any peptide we cover, so it's worth laying out in order.

The early promise

The 2007 study was small and short, but it did what a proof-of-concept study should: under controlled inpatient conditions, 14 patients given larazotide and 7 given placebo ate gluten, and only the placebo group's gut became measurably leakier [1]. Interferon-gamma, an inflammatory marker, rose in 4 of 7 on placebo versus 4 of 14 on the drug.

The problem with the outpatient trials

The next two trials used the same permeability test, a urine ratio of two sugars (lactulose and mannitol), as their main outcome, and it failed both times. In 2012, with 86 patients given 2.4 g of gluten a day for two weeks, the measurement was so noisy outside hospital that gluten itself didn't raise it significantly compared with a gluten-free control, so the trial couldn't test its own question [2]. In 2013, with 184 patients eating 2.7 g of gluten a day for six weeks, there was again no difference in the permeability ratio between drug and placebo [3]. Both trials reported better symptom scores at some lower doses, and in 2013 antibody rises against tissue transglutaminase were several times smaller on larazotide (a mean ratio over baseline of 19.0 on placebo versus 3.88 to 7.72 on the drug) [3].

The trial usually cited

The 2015 Phase 2b study is the one you'll see quoted as evidence that larazotide works. It randomised 342 adults who'd been on a gluten-free diet for at least a year but still had symptoms [4]. The lowest dose, 0.5 mg three times daily, met the primary endpoint of fewer gastrointestinal symptoms, and also reduced symptomatic days by 26%. The 1 mg and 2 mg doses did no better than placebo on any endpoint. Safety matched placebo.

The dose pattern is the most interesting thing in the programme. Symptom benefit kept turning up at the lower doses, 1 mg in 2013 [3] and 0.5 mg in 2015 [4], while the higher doses matched placebo. The authors themselves called the results mixed [4], and you cannot build a programme on a dose that only works when it is the smallest one tested. But that pattern is a real observation about this molecule, and it is the first thing a future trial should try to explain rather than repeat.

Phase 3: stopped at the interim

CeDLara was designed to settle it: a randomised, double-blind, placebo-controlled Phase 3 in coeliac patients with persistent symptoms, aiming for 525 people [5]. On 21 June 2022, the sponsor, 9 Meters Biopharma, announced that a planned interim analysis, run by an independent statistician after about half the target had finished the 12-week double-blind phase, showed the extra patients needed to detect a significant effect were too many to justify continuing [6]. The registry lists the trial as terminated with 307 people enrolled [5]. No results have been posted or published, so we don't know how close it came.

In July 2023, 9 Meters filed a voluntary Chapter 7 petition and ceased operations [7].

What the failure teaches

Larazotide's programme is a useful case study in how a sensible mechanism can fail to become a medicine. The drug was built to reduce gut permeability, yet the trials never managed to measure permeability reliably outside hospital [2,3]. That left the developers steering by symptom scores, which in coeliac disease move with diet slips, stress and placebo effects. A 2026 review in Gut still lists larazotide among experimental barrier therapies and calls for reliable biomarkers of permeability before such drugs can be judged properly [11].

It is also a reminder of why people with coeliac disease still have no approved drug. When the Phase 2b ran, the authors noted there was no approved or proven non-dietary treatment [4]. A strict gluten-free diet remains the only established therapy, and larazotide was never a replacement for it: every trial kept patients on the diet or deliberately gave measured gluten under supervision.

Beyond coeliac disease

Researchers at Massachusetts General Hospital tried larazotide in multisystem inflammatory syndrome in children (MIS-C), a rare complication after COVID-19. Four children treated alongside standard therapy cleared viral spike protein from their blood faster and had quicker resolution of gut symptoms than 22 historical comparison children [8]. A later randomised, placebo-controlled Phase 2a in 12 children saw no drug-related adverse events, with faster symptom resolution and spike clearance [9]. Twelve children is a narrow base for an efficacy claim, so treat the efficacy signal as provisional and the clean safety record as the part that travels; the registry records the trial as terminated at 12 participants because MIS-C cases declined [12].

In rats with experimental acute pancreatitis, larazotide given for a week beforehand reduced gut leakiness and cut bacterial spread to other organs from 100% of animals to 50% [10]. That is the tight-junction mechanism doing exactly what it was designed to do in a living animal, and it is why the compound keeps reappearing in conditions other than coeliac disease. The design that would matter most in the clinic, giving it after an injury rather than before one, is the version still waiting to be run.

Safety and side effects

This is where larazotide's record is strongest. Across the trials above, adverse event rates were similar to placebo [1,2,3,4], with headache and urinary tract infection the most common in the 2012 trial [2]. No serious drug-related safety problem emerged in the published studies, though the Phase 3 safety data were never published.

Legal and regulatory status

Larazotide isn't approved as a medicine anywhere we could find, and with its developer liquidated [7] no active coeliac programme is visible. It isn't available on prescription in the UK or US. Anyone selling "larazotide" for personal use is selling an unapproved research chemical whose clinical programme failed.

What is still open

• The actual Phase 3 numbers. No results were posted or published [5,6], so the size of any effect is unknown.
• Whether the lowest-dose signal in Phase 2b was real or chance.
• Whether a better way to measure gut permeability would have shown the mechanism working in outpatients.
• Whether the MIS-C results hold in a larger trial.
• Whether anyone now owns the rights to larazotide and plans to develop it further. After the 2023 liquidation [7], we found no active coeliac trial on the ClinicalTrials.gov registry.
• Whether larazotide would help as a rescue treatment for accidental gluten exposure, the scenario the gluten-challenge trials came closest to modelling but never tested directly.

What this means

Our position is that larazotide is the clearest case we cover of a mechanism being right while the medicine still failed. The tight-junction idea behaved the way it was designed to behave whenever it could be measured cleanly: gut permeability stayed flat on the drug in the 2007 inpatient study [1], and the same molecule cut bacterial translocation in an animal model of acute pancreatitis [10]. Nothing in the coeliac programme suggests the biology was wrong. What the programme could not survive was a main outcome it could not measure. We think that distinction matters, because it is the difference between abandoning a target and abandoning a way of testing it, and larazotide's failure was the second kind.

We also push back on the reading that larazotide simply "didn't work". One dose did work. In the 2015 Phase 2b trial, 0.5 mg three times daily met its primary endpoint, and it was the lowest of the three doses tested [4]. That dose pattern, benefit at the bottom of the range and nothing at the top, points at a narrow therapeutic window rather than an inert molecule, and it is the single most interesting thing in the whole programme. A future trial that ignored it and simply used a bigger sample or higher doses would be repeating the mistake rather than correcting it.

For anyone reading this with coeliac disease, our view is blunt: none of this changes how you eat today. Every trial here kept patients on a gluten-free diet or gave them measured gluten under supervision, and a strict gluten-free diet is still the only established therapy [4]. With the developer liquidated [7] and no active coeliac programme on the registry, larazotide is not a horizon to plan around. If it comes back, our expectation is that it comes back as an adjunct to the diet in a trial built on an endpoint that can actually be measured, not as a replacement for the diet.

Educational content only — not medical advice. Many peptides discussed on HPR are not approved for human use. Talk to a qualified clinician before making any decision about your health.