CT-868
CT-868, once-daily biased dual GLP-1 and GIP receptor agonist
Written by Aaron CuhaReviewed Sep 2026
Also known as: Acmopatide, CT868
Its phase 2 in 103 adults with type 2 diabetes cut glycated haemoglobin 1.61 to 2.24 percentage points against placebo but produced only 2.9 percent weight loss. Roche discontinued it in July 2026 in favour of another compound from the same acquisition.
Overview
CT-868 came out of Carmot Therapeutics in Berkeley, which Roche bought at the end of 2023 in a deal worth up to 3.1 billion dollars. It later received the nonproprietary name acmopatide.
It is unusual in this class in two ways. It is a dual GLP-1 and GIP receptor agonist, like tirzepatide, but it is given once daily rather than weekly, and it is described as fully biased toward cyclic AMP signalling, meaning it is engineered to trigger one downstream arm of receptor signalling while avoiding the beta-arrestin arm that drives receptor internalisation.
The phase 2 read out well on glucose and poorly on weight. Over 26 weeks in 103 adults with type 2 diabetes and a body mass index of at least 27, glycated haemoglobin fell 1.61 to 2.24 percentage points more than placebo, which is a large effect. Body weight fell 2.9 percent against placebo on the top dose, which is small for a dual incretin agonist in 2026. The authors conclude by supporting investigation of higher doses to maximise weight loss.
That investigation will not happen. Roche discontinued the compound in July 2026, stating in its investor materials that other type 1 diabetes programmes were being prioritised, and directed its effort to enicepatide, the once-weekly dual agonist from the same acquisition which is also being developed for obesity.
The trial also carries an unusual protocol note: COVID-19 related supply constraints meant some participants randomised to 4.0 mg received no more than 3.25 mg and were analysed as a separate dose arm.
Mechanism of action
A dual agonist at the GLP-1 and glucose-dependent insulinotropic polypeptide receptors, described by its developers as fully biased toward cyclic AMP signalling. Signalling bias means the ligand preferentially activates the G protein and cyclic AMP arm of receptor signalling while producing little beta-arrestin recruitment, which reduces receptor internalisation and desensitisation. The theoretical attraction is sustained receptor responsiveness with potentially less of the receptor downregulation that accompanies conventional agonism. Whether that theory delivers anything clinically useful is exactly what this compound's record leaves unanswered, because the glycaemic effect was strong and the weight effect was not, and the programme stopped before higher doses were tested.
Human evidence
One published phase 2 trial in 103 adults with type 2 diabetes and obesity or overweight. No phase 3 was run and the programme has been discontinued.
- Glycated haemoglobin fell 1.61 to 2.24 percentage points against placebo over 26 weeks, a strong glycaemic effect.
- Body weight fell 2.9 percent against placebo on the highest completed dose, which is modest for a dual incretin agonist.
- No participant experienced hypoglycaemia and adverse events were mostly mild to moderate.
- Some participants randomised to 4.0 mg received a maximum of 3.25 mg because of COVID-19 related supply constraints and were analysed separately.
- The authors explicitly called for investigation of higher doses to maximise weight loss; those trials were not run.
What this does not tell you: 103 participants, 26 weeks, one trial. The dose that the authors thought would produce meaningful weight loss was never tested. Most authors are current or former employees of the sponsor. And because the programme was stopped, the question the trial raised, whether a biased once-daily dual agonist can match weekly agonists on weight at a higher dose, will not be answered by this molecule.
Reading the research record
This is the discontinuation entry in the batch, and discontinuations are worth documenting because they carry information that surviving programmes do not.
The phase 2 result splits cleanly. On glucose, CT-868 performed at the top of its class: a 1.61 to 2.24 percentage point placebo-adjusted reduction in glycated haemoglobin is comparable with what the weekly dual agonists achieve. On weight, 2.9 percent placebo-adjusted at the top dose is far below tirzepatide territory and below what most GLP-1 monoagonists manage.
The trial's own conclusion was to test higher doses. Roche instead stopped the compound in July 2026 and put its effort behind enicepatide, formerly CT-388, the once-weekly dual agonist acquired in the same transaction, which is being developed for obesity and diabetes together. Read commercially, that is a decision about where obesity value sits, not a safety finding: nothing in the published record suggests the compound was stopped for harm.
The signalling bias story is the part that stays unresolved. A fully cyclic AMP biased agonist was supposed to sustain receptor responsiveness by avoiding beta-arrestin driven internalisation. The one trial that tested the idea in people produced excellent glucose control and unimpressive weight loss, which is a genuinely interesting dissociation, and now nobody will find out whether it was a dose problem or something intrinsic to the bias.
A compound with one published phase 2 and a terminated programme is not a treatment. Anything sold under this code is not the trial material.
The evidence, charted
Fig. 1 · evidence composition
1of 2 citations (50%) are in people
Counted from the citation list on this page. The Human count is the same number shown in the badge at the top. Both understate any literature larger than the sources we cite.
Fig. 2 · evidence over time
Citations here span just two years, 2021 and 2026.
Too few distinct publication years on this page to plot as a timeline.
Fig. 3 · legal status at a glance
US
Not approved
UK
Not approved
AU
Not approved
CA
Not approved
Not approved in any of the four jurisdictions shown. A jurisdiction's classification is a regulatory fact, not a verdict on the science; see Legal status below for the full text and any notes.
Fig. 4 · dose response
No human dose response curve exists
We draw this figure where the data supports it. For this compound in humans it does not, so the panel stays empty rather than borrowing an animal curve and implying it transfers.
Key studies & citations
- Human2026
Efficacy and safety of CT-868, a novel, fully biased, dual glucagon-like peptide-1/glucose-dependent insulinotropic polypeptide receptor agonist, in type 2 diabetes: A double-blind, randomized placebo controlled phase 2 trial
26 week phase 2 in 103 adults with type 2 diabetes, body mass index at least 27 and glycated haemoglobin 7.0 to 10.0 percent, randomised to once-daily CT-868 1.75 mg, 4.0 mg or placebo. Glycated haemoglobin fell 1.61 to 2.24 percentage points against placebo. Body weight fell only 2.9 percent against placebo on 4.0 mg. Fasting glucose, self-monitored glucose and most lipid parameters improved. No participant experienced hypoglycaemia. COVID-19 supply constraints meant some participants assigned 4.0 mg received a maximum of 3.25 mg and were analysed as a separate arm. Most authors are current or former Roche or Carmot employees.
Diabetes, Obesity and Metabolism - Review2021
Emerging glucagon-like peptide 1 receptor agonists for the treatment of obesity
A review of compounds with GLP-1 receptor agonism then in clinical development for obesity, listing CT-868 alongside CT-388, tirzepatide, cagrilintide plus semaglutide, cotadutide, efinopegdutide and others. Useful here mainly as a record of where this compound sat in the pipeline five years before it was discontinued, and as a reminder of how many entries on that 2021 list never reached a market.
Expert Opinion on Emerging Drugs
Frequently asked questions
Is CT-868 still in development?
No. Roche discontinued it in July 2026, stating that other type 1 diabetes programmes were being prioritised, and focused on enicepatide, the once-weekly dual agonist acquired from Carmot in the same transaction.
Was it stopped for safety reasons?
Nothing in the published record suggests so. The phase 2 reported mostly mild to moderate adverse events and no hypoglycaemia in any participant. The stated reason was portfolio prioritisation.
Why did it work so well on glucose and so poorly on weight?
That dissociation is the most interesting fact about the compound and it was never explained. Glycated haemoglobin fell 1.61 to 2.24 percentage points against placebo while weight fell only 2.9 percent. The authors thought higher doses would fix it. Those doses were never tested.
What does fully biased mean?
It describes a ligand engineered to activate one arm of receptor signalling, here the cyclic AMP pathway, while producing little beta-arrestin recruitment, which is the arm that drives receptor internalisation. The theoretical benefit is sustained receptor responsiveness. This trial is the only human test of the idea for this molecule, and it left the question open.
Can I buy CT-868?
It was never approved anywhere and its development has stopped. Material offered under this code online has no connection to the clinical trial product and no verified identity, purity or dose.