CT-388: Roche's Signal-Biased Dual GLP-1/GIP Agonist Heads Into Phase 3 for Obesity (July 8, 2026)
Table of Contents
The setup: Roche buys its way into the obesity race
For most of the GLP-1 era, the obesity story has belonged to two companies - Novo Nordisk (semaglutide) and Eli Lilly (tirzepatide). Roche, historically an oncology and diagnostics powerhouse, was a conspicuous latecomer. It fixed that in one move: in December 2023 it agreed to acquire California-based Carmot Therapeutics for about $2.7 billion (closing in January 2024), buying a ready-made incretin portfolio. The crown jewel of that deal is CT-388, an investigational once-weekly, subcutaneously injected peptide that activates both the GLP-1 and GIP receptors - the same dual-incretin mechanism that made tirzepatide a blockbuster. Roche now carries CT-388 under the internal code RO7795068, and in the first half of 2026 it pushed the drug into Phase 3, the last stage before a regulatory filing. The question the field is now asking is simple: is CT-388 just another 'me-too' dual agonist, or does it have a genuine edge?
What actually makes CT-388 different: biased signaling
CT-388 hits the same two receptors as tirzepatide, so the headline mechanism is not new. The interesting part is how it turns those receptors on. When a hormone or drug activates a G-protein-coupled receptor like GLP-1R or GIPR, two things can happen downstream: G-protein signaling (the therapeutic effect) and beta-arrestin recruitment. Beta-arrestin is a molecular 'off switch' - it pulls the activated receptor inside the cell (internalization) and desensitizes it, which over time can blunt a drug's effect (a phenomenon called tachyphylaxis). CT-388 was engineered as a 'signal-biased' (G-protein-biased) agonist: it potently activates GLP-1 and GIP receptors while recruiting little to no beta-arrestin. The hypothesis is that by avoiding the internalization/desensitization arm, CT-388 keeps its receptors on the cell surface and responsive, producing more sustained, durable activity across the once-weekly dosing interval. If that translates to the clinic, it could mean deeper or more persistent weight loss - though 'biased signaling advantage' remains a hypothesis that head-to-head trials will have to prove.
The Phase 1b signal: 18.8% in 24 weeks
The first serious human read on CT-388 came in May 2024. In a Phase 1b study, adults with obesity received once-weekly subcutaneous CT-388, titrated over 24 weeks, versus placebo. The result was a placebo-adjusted weight loss of roughly 18.8% at 24 weeks - a strong number for a Phase 1b, and one that put CT-388 immediately in the conversation with the best-in-class incretins. The response-rate breakdown was striking: 100% of treated participants lost more than 5% of their body weight, 85% lost more than 10%, 70% lost more than 15%, and 45% lost more than 20%. There was also a metabolic bonus - every participant who had prediabetes at baseline became normoglycemic by week 24. Tolerability was consistent with the class: mostly mild-to-moderate gastrointestinal side effects (nausea, vomiting, diarrhea).
The Phase 2 readout: 22.5% and no plateau
In January 2026, Roche reported positive topline results from CT388-103, a 48-week, randomized, double-blind, placebo-controlled Phase 2 dose-finding trial in 469 adults with obesity or overweight plus at least one weight-related comorbidity. The study tested several dose-escalation regimens up to 24 mg once weekly. It met its primary endpoint with a placebo-adjusted mean weight loss of 22.5% using the efficacy estimand (18.3% using the more conservative treatment-regimen estimand, which better reflects real-world adherence) at the 24 mg dose at week 48. Two details stood out. First, there was a clear dose-response and the 24 mg curve had not plateaued by week 48 - meaning longer or higher-dose treatment might push the number even higher. Second, the response rates were deep: 95.7% of participants on 24 mg lost at least 5%, 87% at least 10%, 47.8% at least 20%, and 26.1% at least 30%, and 54% achieved 'resolution of obesity' (dropping below a BMI of 30) versus just 13% on placebo. Safety was again in line with the incretin class, with predominantly mild-to-moderate GI events and no new or unexpected signals.
How it stacks up against tirzepatide - and where combos come in
A placebo-adjusted 22.5% at 48 weeks lands CT-388 squarely in tirzepatide territory. In the SURMOUNT-1 obesity trial, tirzepatide produced roughly 20.9% total weight loss at the top dose over 72 weeks. Cross-trial comparisons are treacherous - different populations, durations, and estimands make head-to-head claims unreliable - so the honest read is that CT-388 looks competitive, not clearly superior, on the current data. That is why many analysts think CT-388's real test lies in combinations. Because Roche inherited a broader Carmot pipeline, it can pair CT-388 with an amylin analog such as petrelintide (also in the Roche/Carmot portfolio); amylin-plus-incretin combinations are one of the most promising routes to bariatric-surgery-like weight loss. Roche has signaled interest in exactly this kind of combination development, which could ultimately differentiate CT-388 more than the biased-signaling mechanism alone.
What to watch, and the caveats
CT-388 is investigational. It is not approved anywhere, it is not available by prescription, and anything marketed online as 'CT-388' is not the clinical drug - it is unverified material of unknown identity and purity, and should not be used. The efficacy figures so far are topline/press-release stage for Phase 2 and Phase 1b; full peer-reviewed publications and, critically, Phase 3 outcomes over longer durations are still needed. Durability after stopping the drug - a real weakness of the incretin class, where weight tends to return - has not been characterized for CT-388. And the biased-signaling advantage, while mechanistically appealing, has not yet been shown to produce a clinical benefit over conventional dual agonists in a head-to-head setting. The near-term milestones to watch are the design and enrollment of Roche's Phase 3 program (initiated in H1 2026), any combination-therapy trials with an amylin analog, and full publication of the CT388-103 data. As always, this article is educational and not medical advice; decisions about obesity or diabetes treatment should be made with a qualified clinician.
Ready to track your peptide research?
Open PepTracker ProPepTracker Pro Research Team
The PepTracker Pro Research Team is an editorial group of science writers, pharmacologists, and clinical researchers dedicated to making peptide science accessible. Every article is reviewed for accuracy against peer-reviewed sources and updated as new evidence emerges.
Citations
- [1] Roche announces positive Phase II results for its dual GLP-1/GIP receptor agonist CT-388 in people living with obesity (Roche, January 2026) Source
- [2] Roche reports positive Phase Ib results for its dual GLP-1/GIP receptor agonist CT-388 in people with obesity (Roche, May 2024) Source
- [3] Investigational Dual GLP-1/GIP Agonist CT-388 Demonstrates 22.5% Placebo-Adjusted Weight Loss in Phase 2 Obesity Trial (Patient Care Online, 2026) Source
- [4] Roche moves obesity asset CT-388 into Phase III (European Biotechnology, 2026) Source
- [5] Roche, trailing in obesity, showcases new data for GLP-1 shot (BioPharma Dive) Source
Related Articles
A comprehensive guide to GLP-1 receptor agonists — how they work, what sets them apart, and where the science is heading.
Insulin Efsitora Alfa: The Once-Weekly Basal Insulin Fc-Fusion Heading for an FDA Decision (July 7, 2026)
Basal insulin has been a daily shot for a century. Insulin efsitora alfa fuses a single-chain insulin to an antibody Fc fragment so one injection lasts a week - a ~17-day half-life and a flat, stable profile. Here's how the Fc-fusion design works, what the Phase 3 QWINT trials showed in type 2 and type 1 diabetes (including a real hypoglycemia caveat), and why its FDA decision this quarter matters.