Enicepatide is an experimental once-weekly injection from Roche, and it belongs to the same family as tirzepatide: it switches on the receptors for two gut hormones, GLP-1 and GIP, that help control blood sugar and appetite. On September 22, 2026, Roche said its top dose lowered A1c by 2.65 points and body weight by 15.5% over 48 weeks in people with type 2 diabetes. It is not approved anywhere, nobody can prescribe it, and those numbers can't yet be fairly set beside tirzepatide's.
What Roche announced about enicepatide
As of September 23, 2026, the day this article was written, the only source for these results is Roche's release of September 22. It is a company summary of the headline numbers, not a peer-reviewed paper. The trial, CT-388-104, enrolled 447 adults with type 2 diabetes and overweight or obesity and gave them weekly enicepatide or a placebo shot for 48 weeks (ClinicalTrials.gov NCT06628362).
At the highest dose, 24 mg:
- A1c fell 2.65 points, from an average of 8.1%. A1c is a blood test that reflects your average blood sugar over the past two to three months. 6.5% is the line used to diagnose diabetes.
- About 9 in 10 people (90%) finished at 6.5% or lower, and 62% got below 5.7%, the non-diabetic range.
- Weight fell 15.5% on average, and Roche says it had not levelled off by week 48.
- 2.0% of people on enicepatide stopped because of side effects, against none on placebo. The most common side effects were mild-to-moderate stomach and gut problems, the same pattern seen across this class of drugs.
Among people who started with an A1c above 8.5%, the average drop was 4.13 points. Those figures sit close to what people know from tirzepatide, which is why "is it better?" comes up first.
What enicepatide is, and where CT-388 came from
CT-388 is the lab code enicepatide carried at Carmot Therapeutics, a small company in Berkeley, California. Roche agreed to buy Carmot in December 2023 for $2.7 billion upfront and called CT-388 the lead drug in the deal (Roche, December 4, 2023).
The design idea is called biased signalling. When a hormone switches on a receptor (the docking point on a cell that the hormone fits), the cell responds, then pulls some receptors inside, which turns the signal down. Enicepatide was built to trigger the main signal at both receptors while barely setting off that pull-inside step, so the effect should last longer (Chakravarthy et al., Molecular Metabolism 2026). Tirzepatide does something similar, but only at the GLP-1 receptor. At the GIP receptor it acts much like the natural hormone (Willard et al., JCI Insight 2020). Whether that difference matters in people is what later trials have to show. Our guide to how tirzepatide works explains the two receptors.
The first human data, in that same paper, came from a small study: after four weekly doses, people with overweight or obesity lost 4.7% to 8.0% of their weight, against 0.5% on placebo. Then, in January 2026, Roche reported a 48-week trial in 469 people with obesity but no diabetes. At 24 mg, average weight loss was 22.5 percentage points bigger than on placebo when counting only time on the drug, or 18.3 points when counting everyone as randomised, including people who stopped (Roche, January 27, 2026).
Enicepatide vs tirzepatide: reading the numbers fairly
Nobody knows yet whether enicepatide is better, worse or about the same, because the two drugs have never been tested in the same trial. What we can do is put the published numbers side by side, clearly labelled as different studies.
| Enicepatide (CT-388-104) | Tirzepatide (SURPASS-2) | Tirzepatide (SURMOUNT-2) | |
|---|---|---|---|
| Who took part | 447 adults with type 2 diabetes and overweight or obesity | 1,879 adults with type 2 diabetes on metformin | 938 adults with type 2 diabetes and overweight or obesity |
| Length | 48 weeks | 40 weeks | 72 weeks |
| Starting A1c | 8.1% | 8.28% | 8.02% |
| A1c change, top dose | −2.65 points (24 mg) | −2.30 points (15 mg); semaglutide 1 mg −1.86 | Secondary measure, not shown |
| Weight change, top dose | −15.5% (24 mg); placebo figure not released | Secondary measure, not shown | −14.7% (15 mg) vs −3.2% on placebo |
| Source | Company release, not yet peer-reviewed | Frías et al., NEJM 2021 | Garvey et al., Lancet 2023 |
Read across and enicepatide looks competitive. Four things stop that from being a verdict. The trials enrolled different people on different background medicines. They ran for different lengths. Roche hasn't released the placebo group's result, so there's no way to know how much of the 15.5% came from the drug rather than from being in a trial with regular check-ins; in SURMOUNT-2, the placebo group lost 3.2%. And the full analysis, including how people who dropped out were counted, isn't public yet.
Semaglutide sits one step back in the same family. It works on the GLP-1 receptor alone, and in SURPASS-2 tirzepatide lowered A1c more than semaglutide 1 mg did. The next wave adds a third kind of signal: Roche plans to pair enicepatide with petrelintide, a drug based on amylin (a fullness hormone released alongside insulin). Viking Therapeutics' VK2735 is another dual GLP-1/GIP drug still in trials.
What the enicepatide results do not mean
It is not something anyone can prescribe. Enicepatide is investigational and not approved in any country. The only people receiving it are volunteers in Roche's trials. A vial sold online under the name enicepatide or CT-388 has no approval behind it, no prescriber and no pharmacy of record.
It has not beaten tirzepatide. No head-to-head trial exists, and the table above shows why the numbers can't settle it.
A normal-range A1c is not the same as the diabetes being gone. 62% of people reached below 5.7% while taking the drug. What happens after stopping wasn't part of this report.
"No plateau" doesn't mean weight keeps falling forever. It means weight was still going down when the trial ended, so nobody knows yet where it levels off.
Phase 2 is a middle step. A few hundred people for under a year can show whether a drug is worth pursuing. It can't reliably pick up rarer side effects, which is what much larger phase 3 trials are for.
If you're prescribed tirzepatide or semaglutide today
Nothing about your treatment changes because of this news. Tirzepatide is already sold as Mounjaro for type 2 diabetes and Zepbound for weight management, and tirzepatide for type 2 diabetes covers what its own trials found. Enicepatide can't reach a pharmacy before its phase 3 trials finish, and their registry entries estimate those finish in 2028.
Through Promise, tirzepatide is dispensed as a compounded medication, which is different from an FDA-approved product: the formulation offered here is not FDA-approved. Compounded tirzepatide explains the rules around it. A licensed provider may still prescribe a compounded formulation where they judge it appropriate; that decision is between you and your doctor. A licensed provider reviews every request, and not everyone qualifies.
What to watch next
The full paper. It should show the placebo figures, how dropouts were counted and side effects by dose. As of September 23, 2026, the day this article was written, the trial's registry entry lists its two main measurements at week 36, while Roche's release reports week 48; the full paper should explain the difference.
Phase 3. ENITH-1 (about 2,000 people without diabetes) and ENITH-2 (about 1,600 people with type 2 diabetes) started in March 2026 and measure weight at 72 weeks. Their registry entries estimate primary completion in July and August 2028 (ENITH-1, ENITH-2). Roche says a separate blood-sugar program and heart-outcome trials are planned to start in the first half of 2027.
The combination. A phase 2 study pairing enicepatide with petrelintide, called ZYNERGY, is listed as not yet recruiting, with an estimated start of September 30, 2026 (NCT07589686).