Apitegromab is approved as ISEMBYLD — and the muscle question everyone is asking about GLP-1s is still open

This is general education, not medical advice, and not a suggestion to obtain or use anything. Apitegromab is an approved prescription medicine in the United States, sold as ISEMBYLD and given as an infusion by health professionals. We give no doses anywhere in this article, including for the approved use. The use it spends most of its length on, preserving lean mass during weight loss, is approved nowhere and has been applied for nowhere.

On 11 September 2026 the US Food and Drug Administration approved apitegromab, sold as ISEMBYLD by Scholar Rock, for spinal muscular atrophy. It is the first drug that works by blocking myostatin — the body's own brake on muscle growth — to be approved anywhere, for anything, after two decades in which every attempt in this class failed. It binds the inactive precursor of myostatin rather than the mature molecule, which buys it the selectivity earlier attempts lacked.

That is a landmark, and it is not what most people read about it. Within a day the approval was circulating as the muscle drug, the thing that stops you losing muscle on tirzepatide or semaglutide. There is a real trial behind that framing, and a gap between what it measured and what the headline says. The gap is the whole story.

What the FDA actually approved

The indication is narrow: ISEMBYLD is indicated for spinal muscular atrophy in adults and paediatric patients 2 years of age and older who are currently receiving an SMN2-targeted treatment — on top of nusinersen or risdiplam, not instead of them. In this genetic disease the nerve cells that drive muscle die off; the SMN2-targeted drugs address the cause but do not rebuild what is already lost, and that is the gap apitegromab fills.

The approval came after a Complete Response Letter in September 2025 — a refusal in the form submitted — which Scholar Rock stated concerned only a routine inspection of a third-party fill-finish site, expressly not the efficacy or safety data. Europe does not have the drug: the EU application was withdrawn by the applicant on 13 August 2026, again over manufacturing, not over efficacy or safety. Switzerland has neither an approval nor a pending application, Canada no approval on record, and a Japanese submission is announced for the end of 2026.

The SMA evidence, in plain terms

The approval rests on SAPPHIRE (NCT05156320), a phase 3 trial at 48 clinics across nine countries, published in The Lancet Neurology in 2025. It enrolled 188 people aged 2 to 21 with type 2 or non-ambulatory type 3 SMA, all already on nusinersen or risdiplam and all unable to walk at the start; 128 received apitegromab and 60 placebo over 52 weeks, and the main analysis covered the 156 participants aged 2 to 12. The measured outcome, the Hammersmith Functional Motor Scale Expanded, is built from 33 observed tasks worth 0, 1 or 2 points each: it records what a person can actually do — precisely the kind of endpoint the obesity trial lacked.

All three figures are correct and answer different questions; quoting one as if it were the others is the most common error in coverage of this drug. In absolute terms the approved arm gained 1.0 points (n = 53) while placebo lost 1.2 (n = 50). "Nominal" means not corrected for testing several things at once, as the FDA notes in the label.

The label carries no boxed warning and no contraindications. It carries one warning, and it is the counterweight to any casual talk about using this class for cosmetic purposes: fractures. In the pivotal trial, 5 of 53 patients (9%) in the arm that was later approved and 3 of 53 (6%) in the other arm had a fracture, against 1 patient (2%) on placebo; in the open-label extension, 22 patients (9%) had fractures including 5 serious femur ones. Most of those affected already had low bone density or earlier fractures, and animal studies showed harmful effects on bone.

The question everyone is actually asking: EMBRAZE

GLP-1 and GIP/GLP-1 medicines take off a lot of weight, and not all of it is fat. In a body-composition analysis from SURMOUNT-1, participants on pooled tirzepatide lost 5.6 kg of lean mass at 72 weeks against 1.2 kg on placebo (difference −4.4 kg, 95% CI −5.6 to −3.2), alongside 15.9 kg of fat against 3.6 kg; across trials and methods, roughly 25 to 40% of the weight lost is lean tissue. So the question writes itself: if you add something that blocks the muscle brake, does the body give up less of it? EMBRAZE (NCT06445075) was that experiment — a phase 2 study, called proof of concept by its own authors, published in Nature Medicine in 2026.

One technical point shapes how every number below should be read. The trial worked to a two-sided significance level of 20% and therefore reports 80% confidence intervals, not the customary 95% ones, and nothing was corrected for multiple testing, so all p-values are descriptive — the authors say so themselves. An 80% interval is a narrower, more forgiving window: the setting of an exploratory study, not a confirmatory one.

What EMBRAZE showed

It showed a difference on the scan, and the difference was not small.

Taken on its own terms, the first row is a real finding: 1.9 kg more lean tissue kept over six months, 4.2 lb in the company's pounds (80% CI 2.55 to 5.84). Safety was unremarkable and short — adverse events in 39 of 51 on apitegromab (76%) and 36 of 51 on placebo (71%), mostly gastrointestinal and mild, one serious event in each arm (2%), no deaths. Fractures are not reported as a finding here, but this trial ran 24 weeks, while the fracture signal comes from a 52-week trial and a multi-year extension in people whose bones were already fragile.

What EMBRAZE did not show

Strength and physical function, the two things a person would actually notice, were measured — grip force in the dominant hand, and how many times someone can stand up from a chair — but only as exploratory endpoints, the lowest tier of a trial's hierarchy, without the power or the correction that would let them settle anything. From the publication, unedited:

"For additional exploratory endpoints at weeks 24 and 32, no notable differences were observed between apitegromab and placebo in change from baseline in cardiometabolic parameters or physical function, respectively."

And the numbers behind it, from the paper's extended data:

Look at the spread rather than the middle: in every row the standard deviation is several times the gap between the groups, so the variation between individuals swamps any difference between treatments. No confidence intervals and no p-values were reported, precisely because these were exploratory. This is not a trial showing that strength was unaffected; it is a trial that was not built to answer the question. The honest summary: a body-composition measurement improved, and the measurements of what people could do did not.

Lean mass is not muscle

A DXA scan sorts the body into a few buckets by how it absorbs two X-ray energies, and one of them is lean mass: everything that is not fat and not bone mineral. Muscle is in there, but so are your liver, kidneys, gut and connective tissue, and so is water, which moves around a great deal during rapid weight loss. Lean mass is a quantity of tissue, not a capacity. The authors put it in their own paper: "DEXA scan has a limitation in differentiating compartments within lean mass, with accuracy impacted by hydration levels."

That does not make 1.9 kg meaningless. It means the chain has three links — more lean mass, therefore more muscle, therefore more strength and better function — and only the first has been demonstrated. The class has a cautionary precedent for exactly that chain: bimagrumab, an antibody aimed at the receptor, increased thigh muscle volume in older adults in an earlier randomised trial while the paper recorded "no change in muscle strength". Our follistatin-344 page covers that history.

The three findings the headline leaves out

Fat loss was the same. In kilograms, 8.5 against 8.0 — half a kilogram, smaller than its own standard error of 0.86. The much-quoted contrast between "85% fat, 15% lean" with the antibody and "70 / 30" without it is entirely a difference in the ratio, and the ratio changed because the lean part of the denominator shrank, not because more fat came off.

Weight loss was slightly lower with the add-on. 11.2 kg against 12.5 kg, a gap of 1.3 kg in favour of tirzepatide alone. The authors describe weight loss as similar between the groups, which is fair, but it is certainly not the "more weight, better weight" story that circulates.

Most of the advantage faded within eight weeks of stopping. At week 32, eight weeks after both treatments ended, the lean-mass difference was 0.9 kg (80% CI 0.3 to 1.5; nominal p = 0.0480) — more than half the gap gone in two months. Whether the remainder persists nobody knows; the trial ended there.

The "54.9% muscle preservation" headline, decoded

The figure doing the most work in coverage is 54.9%, and it is a relative number: the lean-mass loss was 54.9% smaller with the add-on than without it — 1.6 kg instead of 3.5 kg. It does not mean that 54.9% of anyone's muscle was preserved, and it is not a share of body composition. The absolute number is the one that tells you the size of the effect: 1.9 kg over 24 weeks, in people who lost around 12 kg in total. Relative figures are not dishonest, they are conventional; they simply sound much larger. Ask, every time: a percentage of what, and how much is that in kilograms?

What an approval for a rare disease does not mean

Same molecule, two completely different situations. In spinal muscular atrophy there is a reviewed benefit-risk balance, a binding label, follow-up obligations, and a phase 3 trial in 188 people whose endpoint was a functional scale, in a group losing motor function without treatment. For lean-mass preservation during weight loss there is one phase 2 proof-of-concept study in generally healthy adults, and its endpoint was a number from a scan. A fracture rate of 9% against 2% weighs differently against a disease that dismantles independence than against a cosmetic goal.

A ClinicalTrials.gov query on 21 September 2026 returns seven registered studies of apitegromab: six in SMA or facioscapulohumeral muscular dystrophy, and one in obesity, the completed EMBRAZE. There is no phase 3 in obesity — none registered, none filed anywhere, none announced, and Scholar Rock's own next step in this direction is preclinical. Anyone who reads today that there is a drug against muscle loss during weight loss is reading about a six-month feasibility study in 87 people whose function measurements showed nothing.

The rest of the class says the same thing

Three independent programmes, one consistent result: blocking myostatin or the receptor it signals through shifts the composition of weight loss towards fat, measurably, and none has yet published evidence of an advantage in strength or physical capability.

Doping: named on one list, absent from another

Apitegromab is prohibited at all times, in and out of competition, under section S4.3 of the 2026 WADA Prohibited List, where it appears by name and first in its line: "Myostatin- or precursor-neutralizing antibodies (e.g. apitegromab, domagrozumab, landogrozumab, stamulumab)". Substances in S4.3 are non-specified, the stricter of the two categories. Our overview of peptides banned in sport has the wider picture.

German criminal law is the other way round: the annex to the German Anti-Doping Act does not name apitegromab, listing only follistatin and its derivatives and stamulumab under substances that alter myostatin function, with a catch-all clause reaching only salts, esters, isomers, complexes and derivatives of those two. That is a statement about the text of a list, not legal advice, and a snapshot — the annex is revised periodically, and the WADA listing makes future inclusion likely. Nothing about the German gap changes the sporting ban.

It is not a grey-market peptide, and it cannot be one

This is where apitegromab separates cleanly from everything else discussed under the heading "myostatin inhibitors" in forums. It is a monoclonal antibody of roughly 150 kDa, grown in mammalian cell culture, whereas chemical peptide synthesis — the process behind the vials sold as research peptides — reaches roughly 5 to 10 kDa. An online seller offering "apitegromab" as a ready-to-inject product has neither manufactured nor tested that molecule; what legitimately exists is laboratory reagent, sold for cell and animal work and labelled as not the clinical drug. Compounding pharmacies cannot make it either, and the FDA's 2026 warning letters to peptide sellers name no myostatin antibody at all — no warning letter, safety alert or import alert anywhere names apitegromab.

Nor is the distinction from follistatin, the substance most often mentioned in the same breath, cosmetic: follistatin is one of the body's own binding proteins and traps activin A, GDF-11 and several bone-forming growth factors alongside myostatin, so it is precisely not selective. For the legal framing, our primer on whether peptides are legal covers it, and what peptides are starts from scratch — apitegromab, for the record, is not one.

Bottom line

Apitegromab's approval as ISEMBYLD on 11 September 2026 is a real first: after two decades of failure, a drug that blocks myostatin has cleared a regulator, for spinal muscular atrophy, on the strength of a functional scale that improved by 2.2 points. It says nothing about weight loss.

For weight loss there is one phase 2 study: 102 people randomised, 87 analysed, 24 weeks, 80% confidence intervals. It found 1.9 kg more lean mass preserved on a DXA scan — a difference the instrument cannot resolve into muscle, organ tissue or water. Fat loss was the same in kilograms, weight loss was slightly lower with the add-on, and eight weeks after treatment ended the gap was down to 0.9 kg. Strength and the ability to stand up from a chair, measured only informally, showed no notable difference. There is no phase 3 and no application anywhere.

The scan moved. Whether anything a person would notice moves with it is the question the next trial, if there is one, will have to answer.

Sources

• FDA, Prescribing Information ISEMBYLD (apitegromab-mstn), BLA 761463
• FDA, BLA Approval Letter, BLA 761463, 11 September 2026
• FDA, FDA Approves First Therapy to Target Muscle Loss in Spinal Muscular Atrophy, 11 September 2026
• EMA, Isembyld — EPAR overview, application withdrawn 13 August 2026 (procedure EMEA/H/C/005909)
• Pratley RE, Denham D, Trivedi R, et al., Nature Medicine 2026;32(7):2673–2678, Apitegromab for lean mass preservation during tirzepatide-induced weight loss: a randomized, double-blind, placebo-controlled phase 2 trial (EMBRAZE, NCT06445075)
• Crawford TO, Servais L, Mercuri E, et al., The Lancet Neurology 2025;24(9):727–739, Safety and efficacy of apitegromab in nonambulatory type 2 or type 3 spinal muscular atrophy (SAPPHIRE)
• Crawford TO, Darras BT, Day JW, et al., Neurology 2024;102(5):e209151, Safety and Efficacy of Apitegromab in SMA Types 2 and 3: the phase 2 TOPAZ study
• Heymsfield SB, Aronne LJ, et al., Nature Medicine 2026, Bimagrumab plus semaglutide alone or in combination for the treatment of obesity: a randomized phase 2 trial (BELIEVE, NCT05616013)
• Look M, et al., Diabetes Obesity and Metabolism 2025, Body composition changes during weight reduction with tirzepatide in SURMOUNT-1
• Scholar Rock, SEC Form 8-K Exhibit 99.1, Scholar Rock Announces FDA Approval of ISEMBYLD (apitegromab-mstn), 14 September 2026
• Scholar Rock, SEC Form 8-K Exhibit 99.1, FDA Issues Complete Response Letter for Apitegromab, solely related to observations at a third-party fill-finish facility, 23 September 2025
• Scholar Rock, SEC Form 8-K Exhibit 99.1, Scholar Rock Reports Positive Phase 2 EMBRAZE Trial Results, 18 June 2025
• ClinicalTrials.gov, all registered apitegromab studies, queried 21 September 2026
• Regeneron, press releases of 2 June 2025 and 17 September 2025, full 26-week COURAGE results for trevogrumab with semaglutide
• World Anti-Doping Agency, The 2026 Prohibited List, section S4.3
• Annex to section 2(3) of the German Anti-Doping Act, BGBl. 2023 I No. 67, part III item 3
• FDA, Warning Letter to NuScience Peptides LLC, 733652, 24 August 2026