Bimagrumab:
Can We Lose More Fat While Keeping More Muscle?

Bimagrumab is a monoclonal antibody that binds activin type II receptors A and B. Those receptors receive signals from several TGF-β-family ligands, including myostatin and activins, and transmit them through pathways that can restrain muscle growth. Blocking the receptor reduces downstream SMAD2/3 signaling in experimental muscle and can release part of that brake.^[3]

This is broader than selectively neutralizing myostatin. A myostatin-specific antibody intercepts one ligand. Bimagrumab occupies two receptors shared by multiple ligands. That broader biology may help explain effects on both muscle and adipose tissue, but it also means “myostatin inhibitor” is an incomplete label.

The most exciting proposition is therefore not simply “build muscle.” It is a partitioning hypothesis: during substantial weight loss, can the body lose a larger fraction from fat while defending lean tissue? BELIEVE was designed to test that proposition in people, not merely in cells or mice.

BELIEVE enrolled 507 adults with obesity, or overweight plus a weight-related complication, without diabetes. Mean baseline age was 47.5 years, weight 107.5 kg and BMI 37.3 kg/m²; mean DXA fat mass was 45.8 kg and lean mass 58.3 kg. Participants were randomized across placebo, two bimagrumab regimens, two semaglutide regimens and four combinations. Bimagrumab was infused every 12 weeks; semaglutide was injected weekly after dose escalation. The blinded primary period lasted 48 weeks, followed by an open-label extension to week 72.^[1][2]

The week-48 primary reporting used a treatment-regimen estimand: an effect across all randomized participants regardless of adherence or early discontinuation, with conservative placebo-based imputation for missing data. The extension emphasized an efficacy estimand: the hypothetical effect among treated participants under continued adherence, excluding data after specified intercurrent events. Those numbers answer different questions and should not be merged.^[1]

Only 377 participants—74.4% of those randomized—completed week 48. That makes the missing-data assumptions clinically important. Randomization still protects the primary comparison, but a striking completer-style or adherence estimand will generally look different from a treatment-policy result that includes discontinuations.

Under the treatment-regimen estimand, placebo participants lost 3.5% of body weight, high-dose bimagrumab 8.6%, semaglutide 2.4 mg 13.5%, and the high-dose combination 16.4%. The combination’s advantage over high-dose semaglutide was nominally significant; the trial did not adjust those secondary comparisons for multiplicity.^[1]

The more distinctive result was what composed that weight change. Fat mass fell 33.7% with the high-dose combination versus 21.1% with semaglutide 2.4 mg. Lean mass fell 2.3% with the combination versus 6.9% with semaglutide. Bimagrumab monotherapy reduced fat mass 18.9% while average lean mass increased by about 1.1%.^[1]

The investigators also reported a “fat-loss index”: the fraction of absolute weight loss attributed to fat. It was 71.1% with high-dose bimagrumab, 92.3% with high-dose semaglutide and 100% with the high-dose combination at week 48. A 100% index is a group estimate, not evidence that every participant lost zero lean tissue—or that DXA can precisely identify every tissue compartment.^[1]

In the extension’s efficacy estimand, high-dose bimagrumab, semaglutide and their combination produced estimated weight changes of −10.8%, −15.7% and −22.1%. Fat mass changes were −28.5%, −27.8% and −45.7%; lean mass changed +2.5%, −7.4% and −2.9%. Visceral adipose tissue fell 45.1%, 35.8% and 58.2%, respectively.^[1]

That is an exciting “if treatment is continued” efficacy picture. It should not replace the more conservative randomized treatment-regimen result. The extension was open label, and there was no true continuing placebo comparison at week 72. Its value is in showing durability and trajectory among those represented by the efficacy estimand—not in proving the exact effect every real-world user would experience.

DXA divides tissue into bone mineral, fat and lean soft tissue. Its lean compartment includes skeletal muscle but also water, organs and other non-fat tissue. An intervention can therefore improve “lean mass” without adding the same amount of contractile muscle protein. Hydration and glycogen can also move the number.

BELIEVE did not show a broad physical-function advantage at week 48. SF-36 physical-function scores were similar across groups. One combination group—bimagrumab 30 mg/kg plus semaglutide 1.0 mg—had a nominally larger grip-strength change than placebo, 4.8 kg versus 1.7 kg; the other groups were similar. With multiple comparisons and no consistent dose pattern, that isolated result is hypothesis-generating.^[1]

This distinction is not a reason to dismiss the result. Protecting lean tissue during a large caloric and weight-loss intervention is a worthwhile clinical target. It is a reason to describe the achievement accurately: BELIEVE demonstrated a favorable body-composition pattern, not a generalized enhancement of performance.

In a smaller 48-week trial of 75 adults with type 2 diabetes and obesity, bimagrumab produced a 20.5% reduction in fat mass and a 3.6% increase in lean mass among completers, versus little change with placebo. Body weight fell 6.5%, waist circumference 9.0 cm and HbA1c 0.76 percentage points. The study was small, used a complete-case analysis, had sex imbalance between groups and reported 80% confidence intervals, but it established that bimagrumab’s composition effect was not unique to BELIEVE.^[4]

Small single-dose studies also showed increases in thigh muscle volume or lean mass within weeks without matching strength gains. In 180 men aged 70 or older with sarcopenia, bimagrumab increased lean mass and reduced fat but did not significantly improve the primary Short Physical Performance Battery endpoint. A hip-fracture trial increased lean mass at higher doses without a significant gait-speed or SPPB advantage. In 251 people with sporadic inclusion body myositis, RESILIENT did not improve the primary six-minute walking endpoint.^[5][6][7][8]

A controlled immobilization and low-protein experiment adds a useful mechanistic bridge: bimagrumab helped preserve thigh muscle volume under acute disuse. That is evidence for tissue protection in a deliberately catabolic model, not evidence that a trained healthy person will gain performance-enhancing muscle.^[9]

Through week 48 in BELIEVE, treatment-emergent adverse events occurred in 91.1–98.2% of active-treatment participants versus 74.5% with placebo. Common bimagrumab-associated events included muscle spasms, diarrhea and acne. Adverse-event discontinuations occurred in 14.0–21.4% of bimagrumab-monotherapy participants, 3.6–8.8% with semaglutide, 5.3–12.5% with combinations and 3.6% with placebo. Serious adverse-event proportions varied by group; no deaths occurred.^[1]

Laboratory and imaging findings deserve follow-up. Magnesium declined but remained in the normal range; alkaline phosphatase and creatine kinase rose. Bimagrumab groups showed early increases in total and LDL cholesterol, while semaglutide partly offset that pattern in combination. At week 72, estimated LDL change was +17.6% with high-dose bimagrumab, −8.9% with semaglutide 2.4 mg and +0.1% with the high-dose combination. Several higher-dose groups had roughly 2.1–2.3% reductions in total hip bone-mineral density versus 0.8% with placebo by week 48.^[1]

Bimagrumab remains investigational and is not FDA-approved. As of October 11, 2026, one bimagrumab–tirzepatide phase 2 study in adults with obesity was active but not recruiting, and an academic three-arm body-composition study was recruiting. A separate type 2 diabetes trial was withdrawn before enrollment for strategic reasons. That mixed registry picture means development continues, but there is no marketed indication or established personal-use protocol.^[10][11][12]

Bimagrumab has moved beyond an attractive pathway diagram. Across randomized human studies, it repeatedly shifts body composition toward less fat and more—or less-lost—lean mass. Combined with semaglutide in BELIEVE, it produced a particularly compelling separation: more total weight loss, much more fat loss and substantially less lean-mass loss than semaglutide alone.

The constraint is equally clear. A scanner’s lean compartment is not a direct assay of contractile muscle, and preserving it has not yet translated into a consistent strength or function advantage. The next decisive trial should recruit people at high risk of muscle loss, use imaging that isolates skeletal muscle, prespecify strength and performance endpoints, report both treatment-policy and adherence estimands, and follow participants long enough to measure durability, bone and cardiometabolic safety.

So can we lose more fat while keeping more muscle? The best phase 2 evidence says bimagrumab can meaningfully improve the body-composition side of that equation when combined with semaglutide. Whether it preserves the muscle that matters most—the tissue that produces strength, mobility and long-term function—is the exciting question still waiting for a definitive answer.