Tirzepatide vs Retatrutide: Is Adding Glucagon Worth It?

Here is a small fact about the two most-discussed weight-loss molecules in the pipeline: as of today, no published trial has given them to the same people at the same time. Tirzepatide is a dual agonist that hits the GIP and GLP-1 receptors. Retatrutide is a triple agonist that hits those two plus the glucagon receptor. On the numbers that exist, retatrutide loses more weight. In its TRIUMPH-1 trial, the 12 mg dose produced 28.3% average weight loss at 80 weeks. In SURMOUNT-1, tirzepatide produced 19.5% at 10 mg and 20.9% at 15 mg.

So the honest answer to tirzepatide vs retatrutide is that retatrutide looks bigger. The direction is probably real and the size is not settled. The head-to-head trial meant to settle it, TRIUMPH-5 (NCT06662383), has an estimated primary completion of November 2026. That is this month. The comparison everyone is making is, in some sense, a comparison of two different trials’ results, and the rest of this article is about why that is both useful and a little dumb.

What glucagon is doing in a weight-loss drug

Adding glucagon is a strange choice on its face, because glucagon is the hormone that raises blood sugar. That is roughly the opposite of what the diabetes half of this field has spent decades trying to do. The drug designers’ answer is to pair it with GLP-1 activity. A review titled “Shared mechanistic pathways of glucagon signalling” says glucagon-based multi-agonists like retatrutide include GLP-1 receptor-activating pieces “to counteract glucagon’s hyperglycaemic effects.” You add the thing that raises blood sugar, and then you add the thing that cancels it, and you keep whatever else it does.

What else it does, per the literature, is spend energy. GLP-1/glucagon agonism is described as increasing energy expenditure by stimulating fat oxidation, which distinguishes it from the more appetite-focused GLP-1/GIP approach that tirzepatide takes. Tirzepatide mostly makes you want less food. Glucagon, in theory, also makes the body burn more of what it has. (Why GIP helps at all is its own puzzle, since both stimulating and blocking the GIP receptor seem to add weight loss on top of GLP-1. A different article on this site gets that one.)

There is a cost, at least in the data so far. A network meta-analysis of glucagon receptor agonists found that retatrutide showed the largest weight reduction against placebo (a mean difference of -13.44 kg) and the largest HbA1c reduction, but it and cotadutide “were associated with comparatively lower tolerability” than mazdutide. More receptors, more effects. Some of them you wanted.

A tirzepatide vs retatrutide comparison you could trust, and then worse ones

Let’s build up from the good version. Step one: you take a few thousand similar people, randomly assign half to each drug, run them for the same number of weeks with the same diet counseling and the same measurement rules, and compare. That is a head-to-head randomized trial, and it is the only kind of comparison where the difference between two numbers is mostly the difference between two drugs. SURMOUNT-5 did this for tirzepatide against semaglutide, with 751 participants without diabetes. TRIUMPH-5 is doing it for retatrutide against tirzepatide. Fine.

Step two: you run two separate trials, one per drug, each against placebo, in different years. The people are similar-ish. You compare the drug arms anyway. Step three: the trials are different lengths, so one drug got more weeks to work than the other. Step four: the trials enrolled different people (people being, inconveniently, the main ingredient). Step five: you pool 26 trials of 12 drugs into a meta-analysis and rank them (this is considered rigorous, and it mostly is, which is the annoying part). Step six, the absurd endpoint: you just compare the best number each drug has ever had printed next to its name, in any trial, at any dose, in anyone.

This is a bit like comparing your commute to a coworker’s when they live farther away, drive a different road, and left the house at a different hour. You learn something. You do not learn whose car is faster. (Sports fans have the same argument about whether a hitter from one decade was better than a hitter from another, and it has never been resolved, and the stakes there are lower.)

“Please don’t put these two in the same chart,” is a thing a trial statistician might say, right before someone puts them in the same chart. I write articles that put numbers from different documents next to each other for a living, so I am not throwing stones here.

How far down that road the real numbers are

Pretty far, it turns out. Consider tirzepatide alone, the same molecule in different rooms. In SURMOUNT-1 it got 20.9% at 15 mg. In the SURMOUNT-5 subpopulation meeting Japanese obesity criteria, it got a -20.1% change at week 72. In CagriSema’s phase 3 trial it served as the comparator and got 25.5% at 84 weeks, while CagriSema got 23% and failed to show non-inferiority. Same drug. Several numbers.

Retatrutide moves around too. TRIUMPH-1 enrolled people with a baseline average weight of 248.5 lb and a BMI of 40.0, ran 80 weeks, and reported 28.3% (70.3 lb) by the efficacy estimand, with 2,339 participants in that analysis. In TRANSCEND-T2D-1, which enrolled people with type 2 diabetes and ran 40 weeks, the 12 mg dose produced 15.3% weight loss. People with diabetes tend to lose less on these drugs, and 40 weeks is half of 80, so this is not a contradiction. It is just step four in action.

Then there is step five. A review in Annals of Internal Medicine looked at 26 randomized trials involving 15,491 adults without diabetes and found retatrutide led at 24.2% at 12 mg, ahead of tirzepatide at 17.8% and subcutaneous semaglutide at 13.9%. Notice that retatrutide’s number there is lower than its TRIUMPH-1 number, and tirzepatide’s is lower than its SURMOUNT-1 number. Pooling pulls everyone toward the middle. The ranking held.

So here is the question, and it has two readings. Retatrutide’s TRIUMPH-1 figure is about 7 points above tirzepatide’s best SURMOUNT-1 figure. One reading is that this is glucagon: the energy-expenditure mechanism working on top of appetite suppression, exactly as designed. The other reading is that TRIUMPH-1 enrolled heavier people, ran a long time, and reported the flattering estimand, and that tirzepatide in that same trial might have closed a chunk of the gap (its own 25.5% at 84 weeks suggests it can stretch). Both readings fit the published numbers. I don’t know which is closer. Neither does anyone outside the TRIUMPH-5 data room, I guess.

The pooled-ranking world is also getting crowded. A 2024 systematic review counted 14 ongoing phase 3 obesity trials. UBT251, another triple GLP-1/GIP/glucagon agonist, showed -13.6% to -19.7% at 24 weeks in a phase 2 trial of 205 adults. A 2026 review on central appetite regulation and peripheral energy expenditure said glucagon-containing dual and triple agonists like retatrutide “have exhibited unprecedented weight-loss efficacy in clinical trials.” That is true, and also every one of those trials had its own patients, its own length and its own estimand. The leaderboard is mostly step five and step six, lol.

The trial that is supposed to settle it

TRIUMPH-5 is the step-one version: retatrutide against tirzepatide, same people, same protocol. If it reads out roughly on schedule, a lot of the cross-trial arithmetic above becomes trivia. It will answer how much more weight loss the third receptor buys, in a population where both drugs ran under identical rules. It will also produce the comparison of tolerability that the network meta-analysis could only estimate from separate trials.

It does not answer everything, for instance outcomes. Tirzepatide has cardiovascular data: SUMMIT showed a 38% reduction in cardiovascular death or worsening heart failure, and SURPASS-CVOT found it noninferior, but not superior, to dulaglutide. Retatrutide so far has things like a post hoc analysis of lipoprotein and inflammatory biomarkers from its phase 2 trials. Biomarkers are stuff that might predict outcomes. They are not outcomes, and that gap is not closed by a weight-loss trial, however clean.

And I suppose there is the naming. A head-to-head trial carries real risk for whichever drug is supposed to win, and it must be hard to call your program TRIUMPH and then learn the answer in public. I’m sure that consideration never entered anyone’s planning at all.

Until the data is out, the cleanest summary is the boring one: three receptors have beaten two in every comparison so far, and none of those comparisons was a fair fight.