Roche's Enicepatide Cut HbA1c 2.65 Points, Placebo Unseen

On September 22, Roche announced topline results from CT-388-104, a Phase II trial of enicepatide, its once weekly injection that activates both the GLP-1 and GIP receptors. The trial enrolled 447 adults with type 2 diabetes and overweight or obesity. At the highest dose, 24 mg, average HbA1c fell by 2.65 percentage points over 48 weeks, from a starting level of 8.1%. Mean body weight fell 15.5%. Sixty two percent of people on that dose finished with an HbA1c below 5.7%, the range doctors call normoglycemia. Roche's own headline phrase was "best in disease glycemic control."

Those are large numbers, and they arrive in a crowded field. Our look at the race to follow tirzepatide shows how quickly the bar keeps moving. What the Roche release does not contain matters as much as what it does.

What the trial was

CT-388-104 is a randomized, double blind, placebo controlled, parallel group, multicenter trial. Adults received enicepatide by weekly subcutaneous injection for 48 weeks. The two primary outcomes were the change from baseline in HbA1c and in body weight at week 48, and Roche says both were met. The company describes the effects as dose dependent, which is what a Phase II dose finding study is built to show.

Enicepatide is a "dually biased" agonist. According to Roche, it was designed to activate both receptors strongly while recruiting very little beta arrestin, a protein that helps pull receptors off the cell surface and dampen their signal. The idea is that less internalization means less desensitization and a longer lasting effect. That is a design rationale, not a result, and the topline release offers no data that tests it.

The numbers, and how to read them

At 24 mg, HbA1c dropped by 2.65 points. In the subgroup that began with poor control, baseline HbA1c above 8.5%, the drop was 4.13 points. Ninety percent of people on the top dose reached an HbA1c of 6.5% or lower, the diagnostic threshold for diabetes, and 62% went below 5.7%. Simple arithmetic puts the average end point near 5.5%, which fits those proportions. Weight loss averaged 15.5% at 48 weeks, and Roche says the curve had not plateaued.

Safety looked ordinary for the class. The most common adverse events were gastrointestinal and mostly mild to moderate. Discontinuation because of adverse events was 2.0% across the enicepatide arms and 0.0% in the placebo arm. That is a low number, and it comes with the usual caution that Phase II trials are small, titrate carefully, and often report the tolerability of the people who stayed.

The missing placebo

In a diabetes trial, the placebo arm is not a formality. People in a placebo group usually improve, partly because they are being watched, partly because background medicines get adjusted, and partly because HbA1c drifts back toward the middle when you enroll people at a high point. A 2.65 point fall is the change from baseline on drug. The number that estimates what the drug itself did is the difference against placebo, and Roche did not publish it.

The release also reports the share reaching an HbA1c below 5.7% for the 24 mg group alone. Without the same figure for placebo, and without the lower doses, there is no way to see how steep the dose response curve is or how much of the effect depends on the top dose. The study did include a placebo arm, so the data exist. Until Roche shows them, the sensible reading is that enicepatide is clearly active and that its size relative to its competitors is unknown.

What comes next

Roche says two Phase III studies in chronic weight management, ENITH-1 and ENITH-2, are already running. It plans to start a Phase III glycemic control program and cardiovascular outcomes trials in the first half of 2027. That sequence matters. Weight management trials will answer whether the drug competes on weight. The diabetes program and the outcomes trials will answer the questions that decide whether it earns a place in diabetes care, much as the pill tested against insulin had to make its case head to head.

Comparisons across different trials are tempting and unreliable. Populations differ in baseline HbA1c, duration of diabetes, background medicines and how results are counted, and a drug that starts from a higher HbA1c can show a larger absolute fall for that reason alone. The 4.13 point drop in the worst controlled group is partly that effect. The strongest part of this release is the share of people reaching non diabetic glucose levels within a year, because it is a threshold and not an average. The weakest part is the silence about the control group. I would hold any ranking of enicepatide against tirzepatide or orforglipron until the placebo numbers, the analysis population and the full dose range are on the table.