A ceiling that looked like physics
Every drug class eventually runs into a number it cannot beat. For a long time, selective GLP-1 receptor agonism had one.
SUSTAIN-1 tested semaglutide 1.0 mg as monotherapy in people with type 2 diabetes. Baseline HbA1c of 8.1 percent. At 30 weeks, mean HbA1c landed at approximately 6.6 percent.
That is a genuinely strong result. It is also, across trial after trial in this class, close to where selective GLP-1 agonism stops. Push the dose, extend the trial, and you tend to get diminishing returns past roughly 6.5 percent. The incretin pathway alone appears to have a floor, and 6.6 percent sat right on it.
Adding a second receptor changed that. SURPASS-1 tested tirzepatide, which pairs GIP with GLP-1, at 15 mg monotherapy. Baseline HbA1c of 7.9 percent. At 40 weeks, mean HbA1c reached approximately 5.9 percent, under the efficacy estimand.
Below six. That is a number GLP-1 alone had not managed to reach as monotherapy. Adding GIP co-agonism appeared to move the floor.
Where retatrutide landed
TRANSCEND-T2D-1 started from a baseline HbA1c of 7.9 percent, the same starting point as SURPASS-1. At 40 weeks, the same duration as SURPASS-1. Retatrutide 12 mg brought mean HbA1c to approximately 6.0 percent, under the efficacy estimand.
Read only that number and the story looks unremarkable. Tirzepatide got to 5.9. Retatrutide got to 6.0. A rounding difference, arguably a wash.
Now put the other number back in the frame, the one from Part 2. Tirzepatide 15 mg reached that 5.9 percent with 11 percent weight loss. Retatrutide 12 mg reached its 6.0 percent with 17 percent weight loss, and the paper notes there was no evidence of a plateau in the weight loss trajectory by week 40, meaning the curve was still descending when the trial measurement window closed.
Nearly identical glycaemic control. Six points more weight loss, on a curve that had not finished.
The question this answers
If you read the first series in this publication, you know the shape of this question already, because it was left open on purpose.
Retatrutide's third receptor is glucagon. Glucagon's literal, textbook function is to raise blood glucose. It is the hormone your body releases to correct a low. Building a diabetes drug around a hormone whose primary job is raising blood sugar is, on its face, a strange decision, and the earlier series in this publication covered why the engineering makes it work: the two incretin arms, GLP-1 and GIP, handle glucose control, which frees the glucagon arm to do the things glucagon does well elsewhere in the body, namely drive lipolysis and raise energy expenditure.
What remained unclear, and what the original review this publication covered said explicitly, was whether adding that arm cost anything on the glycaemic side. A hormone that raises blood sugar, even a deliberately underpowered version of it, sitting inside a diabetes drug, is not a free addition on paper.
This trial is the answer, and it is a clean one. It did not cost anything. Retatrutide reached essentially the same HbA1c as the strongest dual-agonist monotherapy data available, at the same baseline, over the same duration, while producing considerably more weight loss.
The authors put it plainly: selective GLP-1 agonism appears to reach an apparent glycaemic floor around 6.5 percent, GIP co-agonism drives that floor down below 6 percent, and the added weight reduction in retatrutide is potentially attributable to the additional glucagon receptor activation, without giving back any of the glycaemic gain.
The numbers underneath the headline figure
The mean tells one story. The distribution tells another, and it is worth sitting with.
At week 40, HbA1c below 7.0 percent, the American Diabetes Association's general target, was reached by 82 to 89 percent of participants across the three retatrutide doses. HbA1c at or below 6.5 percent, the tighter target recommended for people with shorter disease duration and good overall health, which describes this population almost exactly, was reached by 75 to 83 percent.
Then the number that should stop you. HbA1c below 5.7 percent, which is a normal, non-diabetic value, was reached by 35 to 40 percent of participants treated with retatrutide.
More than a third of a population that entered the trial as diagnosed type 2 diabetics left it, at 40 weeks, with bloodwork indistinguishable from someone who does not have diabetes. Fasting serum glucose fell by roughly 30 to 32 mg/dL across doses, against essentially no change on placebo.
The caveat that has to sit right next to all of it
None of the tirzepatide or semaglutide numbers above came from a randomised comparison against retatrutide. They come from different trials, run at different times, with different populations, different baseline characteristics beyond HbA1c, and different statistical designs. The paper says this itself, in careful language: cross-trial comparisons are restricted by differences in study design, and differences in diabetes duration, baseline bodyweight and study design preclude definitive conclusions.
That is not a technicality to skim past. It is the difference between a hypothesis and a proof. Two trials landing on similar HbA1c numbers by different roads is suggestive. It is not the same as putting both drugs in the same participants under the same protocol.
The real answer is coming. Two head-to-head trials are already registered: NCT06260722 against semaglutide, and NCT06662383 against tirzepatide. Until those read out, everything in this section is an informed comparison, not a verdict.
What's left
A drug that appears to hold glycaemic control at the strongest level this class has produced, while not paying the usual weight loss penalty for diabetes, in a population that had never taken a diabetes medication before, is a genuinely unusual combination of results.
It is also, so far, a combination without any deaths that could be blamed on it, without severe hypoglycaemia, and delivered through 40 weeks of self-injection that the large majority of participants stuck with.
That last sentence deserves its own scrutiny, not a victory lap. Part 4 goes through exactly what did happen to the people in this trial, including the numbers that do not make it into a topline summary.
This series covers a peer-reviewed phase 3 clinical trial published in The Lancet. Retatrutide has not been approved by the FDA for type 2 diabetes or any other indication. Nothing here is medical advice.
