Bimagrumab and the Lean-Mass Question: BELIEVE, and the Trial Lilly Stopped

BELIEVE showed that adding bimagrumab to semaglutide changed the composition of the mass lost as well as the amount. In the phase 2 trial, published in Nature Medicine in 2026, the high combination arm reached 22.1% body-weight change at 72 weeks, with 92.8% of the loss coming from fat and lean mass falling by 2.9%. Semaglutide alone reached 15.7%, with 71.8% of the loss coming from fat 1. What the trial did not show is that preserving lean mass on a scan preserves strength or function, or that the pattern carries over to other incretin agents. A planned phase 2b trial pairing bimagrumab with tirzepatide in type 2 diabetes was withdrawn before it enrolled anyone 5.

Bimagrumab is a monoclonal antibody — a protein dozens of times the mass of an incretin peptide — and it is not a peptide therapeutic. It appears on this site because of the question the incretin drugs created. Their pharmacology is covered elsewhere. This article is about body composition: why lean-mass loss became the field's problem, what BELIEVE measured, and what a DXA endpoint can and cannot prove.

The receptor bimagrumab blocks

Skeletal muscle mass is held in check by members of the TGF-β family, of which myostatin is the best known. Myostatin and related ligands, including activin A, signal through activin type II receptors, and that signal restrains muscle growth. Animals lacking functional myostatin carry conspicuously increased muscle mass. Bimagrumab is a monoclonal antibody that binds activin type II receptors and blocks signalling at the receptor, rather than neutralising any single ligand 2.

The fat effect was the surprise. In a 48-week phase 2 trial in adults with type 2 diabetes and obesity, bimagrumab alone reduced fat mass substantially, including visceral fat, while lean mass rose 2. The same trial recorded early, transient rises in serum lipase, amylase and liver enzymes in the bimagrumab group 2. Fat falling while lean mass rises is the reverse of the pattern that incretin agonists produce, and it is why the antibody was paired with one.

Why lean mass became the incretin field's problem

Every intervention that lowers body weight removes some lean tissue along with fat. The question is proportion. In the DXA subset of STEP 1, lean tissue accounted for roughly two-fifths of the mass lost with semaglutide 3. In the body-composition analysis of SURMOUNT-1, the lean share with tirzepatide was roughly a quarter 4. As incretin-driven body-weight change moved from single digits into the twenties, the absolute amount of lean tissue lost grew with it, and the concern sharpened for older adults and for anyone starting with low muscle mass.

The concern is real, but its clinical meaning is not settled. DXA lean mass includes muscle, organ tissue, connective tissue and water, and part of what is lost during large reductions in body weight is tissue that supported the extra mass in the first place. Whether the lean fraction lost on incretin therapy is disproportionate, and whether it causes functional harm, has not been established by any trial designed to answer that question. What is established is that the proportion can be measured. BELIEVE was designed to see whether it could be moved.

BELIEVE: design and the body-composition result

BELIEVE was a double-blind, placebo-controlled phase 2 trial that randomised 507 adults with obesity to nine groups — placebo, bimagrumab alone at a low or high level, semaglutide alone at a low or high level, and combinations of the two — for 72 weeks 1. Body composition was measured by DXA, with MRI used to assess skeletal muscle and visceral fat 1.

Three things follow from that table. The combination produced more body-weight change than semaglutide alone, so the antibody added to the amount as well as altering the composition. The lean share of the loss fell from a little over a quarter to under a tenth. And bimagrumab alone produced double-digit body-weight change made up entirely of fat while lean mass rose — the clearest demonstration so far that the two main components of body mass can be moved in opposite directions pharmacologically 1. The trial also reported falls in high-sensitivity C-reactive protein of up to 83% 1.

Adverse events, and what is missing

The adverse events specific to bimagrumab were acne and muscle spasms, mostly mild to moderate 1. The gastrointestinal profile of the semaglutide-containing arms matched the known profile of the drug. Discontinuation rates by arm were not available in the sources reviewed for this article and are not stated here. The transient pancreatic-enzyme and liver-enzyme rises seen in the earlier monotherapy trial are the other safety observation that any larger trial will need to track 2.

The main thing missing is function. BELIEVE measured tissue. The reports reviewed describe no strength, physical-performance or mobility result showing that preserved lean mass translated into anything a participant would notice 1. A body-composition result is a mechanism result until a functional result stands beside it.

What a DXA endpoint proves

The practical reading is simple. A DXA result can show that a treatment changed what was lost. It cannot show that the change matters. Hydration alone shifts lean soft tissue, which is why MRI was added in BELIEVE to look at muscle specifically 1. Neither instrument measures how well the retained muscle works. BELIEVE cleared the first bar convincingly for this antibody with this incretin partner. The second bar has not been attempted in the published phase 2 data.

The trial Lilly withdrew

A phase 2b trial of bimagrumab and tirzepatide, alone or in combination, in adults with obesity or overweight and type 2 diabetes was registered in 2025 and then withdrawn; the registry records the reason as strategic business reasons 5. Withdrawn, in registry terms, means halted before the first participant was enrolled. The trial therefore generated no data, and it should not be read as a safety or efficacy failure. Nothing about the combination was learned from it.

A separate phase 2 trial of bimagrumab and tirzepatide in 252 adults with obesity or overweight without type 2 diabetes continues. Its primary completion date was January 2026, and no results had been posted to the registry at the time of writing 6. That trial, not the withdrawn one, is the test of whether the BELIEVE pattern holds with a different incretin partner — one that already produces a smaller lean share on its own 4. Whether bimagrumab adds as much to tirzepatide as it did to semaglutide is an open question, with a real chance of a smaller answer.

Limits and open questions

• BELIEVE is a phase 2 trial. Nine groups shared 507 participants, so each arm is small.
• No functional endpoint result has been reported.
• Whether lean-mass preservation lasts after treatment stops, and what composition any regained mass has, is unknown.
• The long-term safety of chronic activin type II receptor blockade in people with obesity is unknown.
• The result applies to semaglutide as the partner. The tirzepatide trial has not reported.
• Whether lean-mass loss on incretin therapy causes harm at all remains the prior, unanswered question.

The lean-mass question was a fair one to raise, and BELIEVE is the first substantial answer to part of it: the composition of body-weight change can be altered, markedly, by blocking a separate pathway 1. The rest of the question — whether that alteration improves strength, mobility or long-term health — remains open. The trial designed to extend the answer to tirzepatide has not reported, and the one that would have tested it in type 2 diabetes never began 56.