The Muscle-Preservation Race: Trevogrumab, Bimagrumab, Apitegromab, and the Pharmacology of Keeping Lean Mass on GLP-1s

The central critique of GLP-1 weight loss in 2024–2025 became this: a meaningful fraction of the weight patients lose isn't fat. It's muscle. The numbers vary by study, anywhere from 25% to 40% depending on population and measurement, but the signal is consistent. In the Phase 2 COURAGE trial run by Regeneron, semaglutide alone produced 6.5% lean mass loss at 26 weeks — roughly 33% of total weight lost was lean mass. That isn't a trivial body-composition side effect for a 55-year-old patient who will be on the drug for decades.

The nutrition-plus-resistance-training answer is real and works (covered in a separate piece), but even optimized behavioral interventions don't fully close the gap, especially in older adults with anabolic resistance. So the pharmacology race is on. Four companies now have obesity programs targeting the myostatin/activin signaling pathway — the endogenous brake on skeletal muscle growth — in combination with GLP-1 agonists.

Regeneron's trevogrumab (anti-myostatin/GDF8) ± garetosmab (anti-activin A) + semaglutide is in the COURAGE trial. Eli Lilly's bimagrumab (anti-ActRIIA/B receptor) + semaglutide is in BELIEVE. Scholar Rock's apitegromab (selective anti-pro/latent myostatin) + tirzepatide is in EMBRAZE. Biohaven's taldefgrobep (myostatin-activin pathway inhibitor) is in an unnamed Phase 2 as monotherapy.

All four produced their first-ever obesity-population clinical data between June 2025 and April 2026. All four work, in the sense of preserving lean mass. They differ in mechanism, efficacy, safety profile, and commercial position. The rest of this piece is what's in the data, what the mechanistic differences mean, and where the regulatory landscape is now.

Before the drug readouts, the biology. Myostatin — formally growth differentiation factor 8 (GDF-8) — was discovered in 1997 as an endogenous inhibitor of skeletal muscle growth. Mice with GDF-8 gene deletion develop dramatically increased muscle mass — individual muscles grow to roughly twice normal size. The 'double-muscled' cattle breeds (Belgian Blue, Piedmontese) carry natural myostatin loss-of-function mutations. A rare German boy born in 2004 with a homozygous myostatin mutation had striking muscle hypertrophy at birth and at age 4½.

The signaling cascade:

• Myostatin and the closely related activin A bind to activin type II receptors (ActRIIA and ActRIIB) on the muscle cell surface.
• Ligand binding recruits type I coreceptors (Alk-4, Alk-5).
• The receptor complex phosphorylates Smad2 and Smad3 intracellular signaling proteins.
• Smad2/3 complex with Smad4 and translocate to the nucleus, where they activate transcription of muscle-atrophy genes and suppress muscle protein synthesis.
• Downstream: suppressed Akt signaling (reducing protein synthesis) and upregulated ubiquitin ligases (increasing protein degradation). Net effect: less muscle.

Three drug-targeting strategies:

• Block the ligand — Monoclonal antibodies that bind myostatin directly (trevogrumab, apitegromab) or bind both myostatin and activin A (or block activin A separately, like garetosmab).
• Block the receptor — Bimagrumab binds ActRIIA/B, preventing any ligand (myostatin, activin A, and some related TGF-β family members) from activating the pathway.
• Decoy the ligand — Taldefgrobep is a soluble receptor-like fusion protein that binds myostatin (and some activins) in circulation before they reach the cell surface.

Each strategy trades off selectivity, potency, and safety. Receptor blockade (bimagrumab) is the broadest and most potent. Selective myostatin blockade (apitegromab specifically for pro/latent forms) is the narrowest. Ligand blockade targeting mature myostatin (trevogrumab) sits in between. The Phase 2 data below starts to show what these biological differences mean clinically.

Trevogrumab (REGN1033) is a fully human monoclonal antibody that binds and neutralizes GDF-8/myostatin. Regeneron's Phase 2 COURAGE trial tested it in combination with semaglutide, with and without garetosmab (an anti-activin A antibody Regeneron already has in development for fibrodysplasia ossificans progressiva).

Five arms ran over 26 weeks: semaglutide alone, sema + trevogrumab 200 mg, sema + trevogrumab 400 mg, sema + trevogrumab 200 mg + garetosmab (the triplet), and sema + trevogrumab 400 mg + garetosmab. Interim data published June 2025; full 26-week data presented at EASD September 2025.

The 26-week results: semaglutide alone produced −6.5% lean mass from baseline (~33% of total weight lost was lean mass). Sema + trevogrumab 200 mg cut that to −3.3% lean mass — roughly half the loss. Sema + trevogrumab 400 mg landed at −3.8%. The sema + trevogrumab + garetosmab triplet pushed lean mass loss down to −2.0%, with up to 92.6% of weight loss coming from fat mass vs ~67% on semaglutide alone. Fat loss was greater in all combination arms compared to semaglutide alone, both in absolute fat mass and fat-to-total-weight ratio. Metabolic markers (waist circumference, blood pressure, lipids, A1C) improved in all arms.

On safety: the doublet (sema + trevogrumab, either dose) was generally well-tolerated, with AE rates comparable to semaglutide alone. The triplet had substantially higher discontinuations for tolerability and other AEs. Adding the anti-activin A arm appears to cost more than it adds.

Commercially, Regeneron is positioning the sema + trevogrumab doublet as its lead candidate. Full trial completion is expected late 2026. This is the most complete Phase 2 GLP-1 + muscle-preservation dataset to date and the first to be presented at a major diabetes meeting.

Bimagrumab is a human monoclonal antibody that binds the activin type II receptors (ActRIIA and ActRIIB), blocking both myostatin and activin A from activating the pathway. Originally developed by Novartis, it was acquired by Eli Lilly in a 2021 deal worth up to $1.9 billion. The BELIEVE Phase 2b trial (published in Nature Medicine, March 2026) was the first large randomized comparison of this approach.

The trial enrolled 507 adults with obesity across nine arms, with a 48-week primary period and 72-week open-label extension. Arms included placebo; bimagrumab monotherapy at 10 and 30 mg/kg IV; semaglutide monotherapy at 1.0 and 2.4 mg; and four combination arms crossing the two bimagrumab doses with two semaglutide doses.

At 48 weeks, bimagrumab monotherapy at 30 mg/kg produced 10.8% weight loss — 100% from fat mass — with lean mass actually increasing by 2.5%. That's the most striking signal in the dataset: a drug that reduces body weight primarily by increasing fat loss without any lean mass loss. Semaglutide monotherapy at 2.4 mg produced 15.7% weight loss, 71.8% from fat mass (28.2% from lean mass). The combination of bimagrumab 30 mg/kg + semaglutide 2.4 mg produced 22.1% weight loss, 92.8% from fat mass, with lean mass largely preserved. Combination arms also showed up to 83% decrease in hsCRP and significant increases in adiponectin, suggesting metabolic improvement beyond composition. In participants with prediabetes, some combination arms showed 100% reversion to normoglycemia.

Safety was the wrinkle. Discontinuations due to AEs were higher in bimagrumab arms (14.0–21.4%) than semaglutide (3.6–8.8%), combination (5.3–12.5%), or placebo (3.6%). Characteristic bimagrumab AEs included mild-to-moderate acne, muscle spasms, and eyelid edema — all consistent with broad activin receptor blockade. Combination AEs included nausea, diarrhea, constipation, and fatigue (the standard semaglutide profile).

The strategic twist: in June 2025, Lilly terminated its Phase 2b trial of bimagrumab + tirzepatide (Zepbound) in patients with obesity and type 2 diabetes for 'strategic business reasons' — before enrollment was complete. The non-diabetic BELIEVE trial ran to completion and produced the Nature Medicine data. The discontinuation is widely read as Lilly narrowing its bimagrumab investment: continuing the lead program while deprioritizing the diabetes comorbidity population, likely because the FDA has signaled it wants to see additional weight loss plus preserved lean mass, not just preserved lean mass on top of standard GLP-1 weight loss. Bimagrumab's profile (acne, muscle spasms, higher discontinuations) may make the weight-loss-plus-muscle value proposition harder to defend than Lilly originally expected.

Apitegromab is a differentiated myostatin inhibitor — it selectively binds only the pro- and latent forms of myostatin, not the mature active ligand. The rationale: selective pro/latent targeting prevents myostatin activation at the source while theoretically preserving some physiological regulation of mature myostatin. Scholar Rock originally developed apitegromab for spinal muscular atrophy (SMA), where the Phase 3 SAPPHIRE trial was positive and the drug is now approved.

The EMBRAZE Phase 2 proof-of-concept trial tested apitegromab as a lean-mass-preservation adjunct during tirzepatide-induced weight loss.

Key 24-week results:

• Tirzepatide alone: 70% fat / 30% lean composition of weight lost.
• Apitegromab 10 mg/kg + tirzepatide: 85% fat / 15% lean.
• Apitegromab + tirzepatide preserved an additional 4.2 lbs (1.9 kg) of lean mass compared to tirzepatide alone — a 54.9% relative preservation (p=0.001).
• Safety was consistent with apitegromab's established SMA safety profile; no unexpected signals in the obesity population.

Next-generation asset:

Scholar Rock is simultaneously advancing SRK-439, a next-gen selective myostatin inhibitor specifically designed for the obesity/metabolic indication. Preclinical data shows:

• Lean mass preservation at doses as low as 0.3 mg/kg; lean mass gain at higher doses.
• Dose-dependent enhancement of fat mass loss when combined with tirzepatide.
• Significant lean mass preservation and attenuation of fat mass rebound following GLP-1 receptor agonist withdrawal — a particularly interesting finding given how much of the GLP-1 story is now about post-discontinuation regain.

Commercial position:

Scholar Rock is the smallest company in this race and arguably the most focused. The EMBRAZE data is cleaner and more specific than COURAGE or BELIEVE (single tirzepatide dose, single apitegromab dose, single primary endpoint). If selectivity matters clinically — which is an open question — apitegromab may differentiate on safety. SRK-439 is positioned for a longer-term play.

Taldefgrobep alfa is a myostatin-activin pathway inhibitor originally developed by Bristol Myers Squibb, licensed to Biohaven in 2022. It's a soluble receptor-Fc fusion protein that acts as a decoy, binding circulating myostatin before it reaches cell-surface receptors.

The SMA failure and pivot:

Biohaven's Phase 3 RESILIENT trial of taldefgrobep in spinal muscular atrophy missed its primary endpoint in late 2024. Encouragingly for the obesity program, subgroup analyses showed the drug significantly reduced total body fat mass versus placebo at 48 weeks — a finding that, while bad news for SMA, pointed toward a potential obesity indication.

Biohaven pivoted. The Phase 2 obesity trial (NCT07281495):

• Randomized, double-blind, placebo-controlled, dose-ranging.
• ~150 participants with obesity.
• 24-week double-blind treatment period + 24-week open-label extension.
• Tests monotherapy taldefgrobep — both once-weekly and once-monthly dosing arms.
• Primary endpoint: percent change in total body weight at week 24.
• Secondary endpoints: total body fat mass, total body lean mass.

Enrollment completed March 2026. Topline data expected 2H 2026.

Why it's different:

Taldefgrobep is testing myostatin inhibition as a standalone obesity treatment — not adjunctive to a GLP-1. This is conceptually closer to the bimagrumab-monotherapy arm of BELIEVE, where blocking the myostatin/activin pathway alone drove 10.8% weight loss (all from fat). If taldefgrobep replicates that signal at a different mechanism point (ligand decoy vs receptor block), the implication is substantial: myostatin inhibition alone may be sufficient for meaningful weight loss in some patients.

Once-monthly dosing is also commercially differentiating. GLP-1 weekly injections are already a compliance barrier; a monthly muscle-sparing biologic might fit patients who can't tolerate weekly GLP-1s.

The four Phase 2 programs used different comparators, doses, endpoints, and populations, so any head-to-head numerical comparison is rough. With that caveat:

Bimagrumab monotherapy (30 mg/kg, BELIEVE) showed lean mass +2.5% — the only drug so far to increase lean mass while reducing body weight. Bimagrumab 30 mg/kg + semaglutide 2.4 mg (BELIEVE) produced a ~92.8% fat / ~7% lean loss ratio at 22.1% total weight loss. The semaglutide + trevogrumab + garetosmab triplet (COURAGE) produced 92.6% fat loss / 2.0% lean mass loss, with high discontinuation for AEs. The semaglutide + trevogrumab 200 mg doublet (COURAGE) cut lean mass loss to 3.3% vs 6.5% for semaglutide alone (~49% prevention) and was well-tolerated. Tirzepatide + apitegromab 10 mg/kg (EMBRAZE) shifted the ratio from 70/30 fat/lean (tirzepatide alone) to 85/15 (~55% relative lean mass preservation). Taldefgrobep monotherapy still has to read out — H2 2026.

The broad pattern: receptor-level blockade (bimagrumab) produces the most dramatic lean-mass gain and the most side effects (acne, muscle spasms, higher discontinuations). Selective myostatin ligand inhibition (trevogrumab, apitegromab) preserves roughly half the lean mass that would otherwise be lost, with cleaner safety profiles. Adding activin A blockade (garetosmab) to myostatin blockade pushes lean-mass preservation marginally further but costs tolerability. Fat loss is typically enhanced when muscle-preservation mechanisms are added to a GLP-1 — not just lean mass preserved — so both directions of the body-composition ratio shift the right way.

A hand-wavy per-percent-prevention ranking: bimagrumab > trevogrumab ≈ apitegromab > trevogrumab doublet without garetosmab. The tolerability and safety ranking runs nearly inverse: apitegromab ≈ trevogrumab > bimagrumab. The commercial question, and the FDA regulatory question, is what balance point matters.

The June 2025 termination of Lilly's bimagrumab + Zepbound type 2 diabetes Phase 2b trial wasn't just a corporate decision. Industry reporting linked it directly to evolving FDA signals about what these combination products need to show to get approved. The read:

• Body composition improvement alone is not enough. Preserving lean mass while matching GLP-1's weight loss isn't a sufficient clinical benefit on its own. The FDA wants to see incremental weight loss (beyond GLP-1 alone) combined with preserved lean mass. Pure 'same weight loss, more of it fat' isn't a clear regulatory win.
• Safety bar is elevated. Chronic anti-activin-receptor therapy hasn't been evaluated for decades-long use. Acne, muscle spasms, and eyelid edema are manageable in a 48-week trial but raise questions about chronic use in people who may take these drugs for 20+ years.
• Outcome endpoints matter increasingly. Pure DXA composition data won't be enough. Future trials will need functional endpoints (strength tests, sit-to-stand, gait speed), hard outcomes (fractures, falls in older populations), or cardiovascular data.

The BELIEVE combination data (22.1% weight loss, 92.8% fat) is probably the strongest single dataset on FDA's desk because it shows both greater weight loss AND better composition simultaneously. But the discontinuation rates in bimagrumab arms (14–21%) are a problem. The COURAGE doublet (trevogrumab + semaglutide, ~3.3% lean mass loss) doesn't show the same incremental weight-loss magnitude over semaglutide alone — it's a composition improvement, not a weight-loss improvement. That's exactly the profile FDA has signaled isn't sufficient.

The likely approval path forward: combinations that show both better total weight loss and better composition, with acceptable tolerability. Some of these programs will get there; others are likely to be discontinued or repositioned.

Myostatin and activin A aren't just brakes on muscle growth. They're part of a conserved biological system with roles beyond skeletal muscle. Long-term inhibition could have unexpected effects:

• Bone and connective tissue: The activin pathway participates in bone remodeling. Early signals from bimagrumab trials showed small increases in fractures, though the data is noisy. Apitegromab's SMA trials — a population that already has compromised musculoskeletal systems — have not shown clear bone concerns, but the obesity population would need separate longitudinal data.
• Cardiac muscle: Myostatin is expressed in cardiac muscle too. Preclinical data suggest inhibition may cause cardiac hypertrophy. Clinical trial cardiac monitoring hasn't flagged this, but follow-up is short.
• Tendon and ligament stiffness: Myostatin regulates tendon biology alongside muscle. Tendon injury rates haven't been a clean safety signal yet, but chronic use in athletic or highly active populations hasn't been studied.
• Fertility and reproduction: Activin A has roles in ovarian function and pregnancy. Anti-activin therapies are contraindicated in pregnancy; long-term effects on fertility in younger patients are unknown.
• Immunity: Myostatin has immunomodulatory effects in some tissues. Long-term inhibition's effect on immune function hasn't been systematically characterized.
• Post-discontinuation effects: What happens when people stop these drugs? Do they lose the preserved muscle rapidly? Do they gain fat faster? The SRK-439 preclinical data showing attenuation of fat rebound after GLP-1 withdrawal is provocative and the most interesting signal in this direction so far, but human data is absent.

These aren't reasons to avoid the drug class. They're the open questions that Phase 3 and long-term real-world evidence will need to answer. The pattern that has emerged with GLP-1s — strong weight-loss data leading to widespread use, followed by slower accumulation of long-term signals (muscle loss, bone density, cardiac remodeling, etc.) — will likely repeat with myostatin inhibitors. Expect a 5-year period after initial approval during which the safety picture sharpens.

Handicapping the race to market:

• Apitegromab (Scholar Rock): Already FDA-approved for SMA (SAPPHIRE Phase 3). Has regulatory history, manufacturing scale, established safety in a muscle-adjacent population. EMBRAZE Phase 2 obesity data positive. Phase 3 in obesity likely to initiate 2026–2027. Earliest obesity approval: possibly 2028–2029.
• Bimagrumab (Lilly): BELIEVE Phase 2b published. Phase 3 would need to overcome the FDA's incremental-weight-loss bar and the AE discontinuation rates. Lilly's decision to terminate the T2D trial suggests they're being selective about which Phase 3 programs to pursue. Earliest obesity approval: 2028–2030.
• Trevogrumab (Regeneron): COURAGE Phase 2 complete data presented EASD September 2025, full 26-week readout and trial completion late 2026. Phase 3 initiation likely 2027. Earliest approval: 2029–2030.
• Taldefgrobep (Biohaven): Phase 2 monotherapy obesity data H2 2026. If positive, Phase 3 in 2027. Monotherapy + monthly dosing is differentiated. Earliest approval: 2029–2031.
• SRK-439 (Scholar Rock): Still preclinical; expected to enter Phase 1 in 2026. The preclinical data on GLP-1 withdrawal attenuation is compelling. Earliest approval: 2030+.

If this plays out as expected, the obesity market will see its first muscle-preserving adjunct sometime in 2028–2029. By the early 2030s, patients on GLP-1 combinations may routinely be offered a myostatin-pathway drug as a second agent — similar to how statin patients today are offered ezetimibe or PCSK9 inhibitors when LDL targets aren't met.

The bigger picture: If GLP-1s become a standard 20–30 year medication for chronic obesity (as cardiovascular medications have for chronic cardiovascular disease), body-composition preservation will matter enormously over that horizon. Pharmaceutical companies are positioning for a potential $15–30 billion combination-therapy market by 2035.

None of these drugs are available for the lean-mass-preservation indication today. Apitegromab is approved only for SMA. Bimagrumab, trevogrumab, and taldefgrobep are in clinical trials. So what does this research mean for someone currently on semaglutide, tirzepatide, or retatrutide who is worried about muscle loss?

Practical implications:

• The current best-practice protocol remains nutrition + resistance training. Adequate protein (1.2–1.5 g/kg ideal body weight/day, leucine-rich), 2–3 resistance training sessions per week with progressive overload, measurement of body composition at baseline and every 6 months.
• Don't wait for the combination drugs. Even if apitegromab is approved for obesity in 2028, you've already lost three years of lean mass on your current GLP-1 regimen if you're not actively preserving it. The behavioral intervention is available now and has strong clinical evidence.
• Get a DXA scan. If lean mass loss is a concern, knowing your baseline and 6-month trajectory helps decide whether you're on track or need to intensify protein/training. This applies even more if you're older or post-menopausal.
• Stay connected to the clinical trial landscape. The COURAGE, EMBRAZE, and BELIEVE data are available now, and qualifying patients can enroll in ongoing and upcoming Phase 2/3 trials of these drugs. Scholar Rock, Regeneron, and Biohaven all have active recruitment.
• Don't self-source muscle-sparing peptides from compounding or gray-market sources. Follistatin-related peptides, unapproved myostatin inhibitors, and other experimental muscle-growth compounds are circulating in the fitness and peptide wellness markets. These have none of the safety data of the clinical-stage antibodies discussed here, and the potency and purity of compounded or research-grade products are unreliable. This is one of the places where the evidence gap between clinical development and consumer peptide channels is widest and most dangerous.

The realistic near-term answer: sustained GLP-1 therapy + protein + strength training produces excellent outcomes for most patients. Pharmacological muscle preservation is coming, but it's a 2028+ story for most people. The trials are exciting because they prove that pharmacological muscle preservation is achievable — not because they change what you should do this month.

The myostatin/activin pathway is real, drug-targetable, and the most active area of obesity-adjacent pharmacology development in 2026. Four drugs are in the race.

Bimagrumab (Lilly, receptor blockade) has the most potent muscle-preservation signal — BELIEVE showed 22.1% weight loss with 92.8% from fat, and monotherapy actually increased lean mass. But it has the highest AE rates and the most complex safety picture. Lilly terminated one of its Phase 2b arms in 2025.

Trevogrumab (Regeneron, anti-myostatin) has strong Phase 2 data from COURAGE / EASD 2025. Doublet with semaglutide cuts lean mass loss roughly in half with clean safety. The triplet with garetosmab pushes further but costs tolerability.

Apitegromab (Scholar Rock, selective pro/latent myostatin) has the cleanest selectivity profile. EMBRAZE showed 54.9% relative preservation of lean mass in a tirzepatide-background population. Already approved for SMA.

Taldefgrobep (Biohaven, myostatin decoy receptor) is testing monotherapy and monthly dosing — the most differentiated commercial position. Phase 2 data H2 2026.

The common thread: all four significantly preserve lean mass during GLP-1 therapy, all four enhance fat loss, and all four are potentially approvable if Phase 3 clears the elevated FDA bar (must show incremental weight loss and composition improvement, with acceptable tolerability).

The outstanding questions: long-term safety (bone, cardiac, tendon, fertility, immune), post-discontinuation trajectories, whether selectivity actually matters clinically, and whether the commercial calculus for a combination therapy priced on top of a GLP-1 holds up.

Timeline: first muscle-preservation-in-obesity approval likely 2028–2029. Widespread adoption by the early 2030s. Until then, the boring answer still applies. Protein + resistance training is the muscle-preservation strategy available to every GLP-1 patient today, and it's well-validated. The pharmacology is coming. The behavioral approach is what you can do now, and it's still the foundation these drugs will augment rather than replace.