Roche's peptide pushed 62% of trial patients to normal blood sugar

On September 22, Roche announced that in a mid-stage trial of adults with type 2 diabetes and overweight or obesity, 62% of the people on its highest dose of an experimental peptide called enicepatide brought their blood sugar down into the non-diabetic range within 48 weeks — a threshold usually reached through years of treatment, lifestyle change, or not at all. Ninety percent reached a less strict target, an HbA1c of 6.5% or below. The company released the numbers three months after analysts had dismissed the same drug as a "me-too" also-ran.

What the trial found

The trial, internally named CT-388-104, randomized 447 adults with type 2 diabetes and overweight or obesity to weekly injections of enicepatide or placebo for 48 weeks, at doses that were increased gradually over the study (Roche, September 22, 2026). At the highest dose, average HbA1c — a blood test that reflects blood sugar control over roughly three months — fell by 2.65 percentage points, from a starting average of 8.1%. Among the subset of participants who started with the worst-controlled diabetes, HbA1c above 8.5%, the average drop was 4.13 points. Average body weight fell 15.5% at 48 weeks, and Roche said the weight-loss curve was still trending down with no sign of leveling off. The most common side effects were mild-to-moderate gastrointestinal problems — nausea, the usual complaint with this drug class — and 2% of people on enicepatide stopped the drug over side effects, versus none on placebo. Roche's chief medical officer, Levi Garraway, called the company "highly encouraged" by a "meaningful proportion of patients reaching normalised glucose levels within less than a year," and the company says two Phase 3 weight-management trials are already underway, with a Phase 3 diabetes program and cardiovascular-outcomes trials planned to start in the first half of 2027.

How it's built to work

Enicepatide is a peptide that activates two separate gut-hormone receptors at once — GLP-1 and GIP — the same pair of targets Eli Lilly's tirzepatide already works through as an FDA-approved treatment for type 2 diabetes and for chronic weight management. Both hormones normally get released after eating; they slow digestion, dial down appetite, and help the pancreas release insulin at the right moments. What Roche says is different about enicepatide is how it engages those receptors: it's designed to trigger the useful signaling pathway on each receptor while minimizing a secondary process called beta-arrestin recruitment, which normally causes a receptor to be pulled inside the cell and desensitized after use. Less of that internalization, in theory, means the receptor stays responsive for longer between doses — a design idea Roche picked up when it acquired the peptide's originator, Carmot Therapeutics, in 2023.

What's still missing

Roche's announcement was a topline summary, not a peer-reviewed paper, and it left out a number that matters: how much of the effect showed up in the placebo group too. People in diabetes trials often see real HbA1c and weight improvements on placebo alone, from closer monitoring and dietary counseling, so the untreated comparison is what turns a raw result into a measure of the drug's actual effect. Roche didn't publish it this time. An analysis in Clinical Trial Vanguard flagged the same gap, noting the company "has not yet released the specific HbA1c and weight-loss figures" beyond the topline numbers, and that dropout patterns at each dose — not just the pooled 2% figure — will matter for how the Phase 3 program is designed (Clinical Trial Vanguard, September 22, 2026).

That gap is a bigger deal for enicepatide specifically because of how differently an earlier readout landed. In June, at a separate trial in people with obesity but not necessarily diabetes, Roche did publish a placebo-adjusted number — weight loss 18% to 22.5% above placebo, depending on how it was measured — and the reaction from Wall Street was lukewarm. Jefferies' Michael Leuchten wrote that the data did "little to differentiate enicepatide from its peers," and RBC's analyst called it a "late-comer, me-too option if it ever reaches the market," next to Eli Lilly's three-hormone peptide retatrutide, which posted roughly 28% weight loss in earlier trials (BioPharma Dive, June 8, 2026). September's diabetes-focused numbers look considerably more striking on their face — but without the placebo-adjusted figures this time, there's no clean way to check whether the underlying drug effect actually improved, or whether a sicker, more diabetes-focused trial population simply had more room to move. Coverage from Fierce Biotech pointed to one more wrinkle: Roche measured its endpoints at week 48, eight weeks later than Lilly measured retatrutide's, which further complicates any head-to-head reading of the numbers as they stand (Fierce Biotech, September 22, 2026).

What this means for researchers

The interesting question enicepatide raises isn't really about Roche's stock chart — it's whether re-engineering how a peptide binds its receptor, rather than adding a third receptor target the way retatrutide does, is a separate, viable route to the same clinical result. The biased-signaling idea behind enicepatide is a genuinely different mechanism than stacking more hormone targets, and a large, randomized, 48-week trial reporting a real reduction in HbA1c and weight is a legitimate data point either way.

But it's one topline announcement, not a finished case. What would resolve the open questions is what's still absent: the placebo-adjusted numbers, a full per-dose safety breakdown, and eventually a peer-reviewed publication that outside statisticians can check against Roche's summary. Until then, September's results sit next to June's more skeptical reception as two readings of the same drug that are hard to fully reconcile — which is itself a useful reminder that a single company announcement, however large the trial, is rarely the last word. Background on how GLP-1 and GIP-targeting peptides are studied more broadly is in Torrey Labs' compound guide; the Journal covered a related head-to-head trial result, CagriSema beating tirzepatide on weight loss the same week Novo's stock fell, in a separate piece.