Scholar Rock ($SRRK): Why Did the Shares Fall 15% in the Month of an FDA Approval?

01 Bull, base and bear scenarios

These are reading frames, not forecasts, and none of them is a recommendation. Each one is written so that a reader can tell which filed document would confirm or break it.

What all three have in common is that the evidence arrives in quarterly filings rather than in announcements. The approval was the last binary event on this file for some time; what follows is arithmetic.

What Would Falsify This Reading

The reading on this page is that Scholar Rock has won a genuine first — the only muscle-targeted treatment approved in spinal muscular atrophy — and that the market has marked the shares down over the month that contained the approval because the label, the evidence base and the company’s own communication all carry qualifications that a headline does not. Each of the following would damage that reading, and each is checkable against a document rather than against an opinion.

• The fracture signal proves benign in commercial use. SAPPHIRE reported fractures in 9 per cent of patients on the approved 10 mg/kg dose against 2 per cent on placebo, and the label carries a warning, an instruction to use the drug with caution in patients with a history of low bone density or multiple fractures and to weigh the risks and benefits of continuing in patients who fracture, and the FDA has requested that every fracture case, serious or not, be submitted as a 15-day alert report. If a year of post-marketing reporting shows no excess over the trial rate and no change in prescribing, then the most cautious part of this page was over-weighted, and the correct conclusion will be that the agency required close watching of a manageable risk.
• The nominal p-values turn out not to matter. The company reports the pivotal result with an asterisk: a 2.2-point HFMSE improvement at nominal p=0.0121 and a responder analysis at nominal p=0.0125. Nominal means not adjusted for multiplicity. The FDA approved on that package anyway, and if the first years of real-world use reproduce the effect size — the ONYX extension cannot help here, because every patient in it receives 20 mg/kg, twice the approved dose — the statistical caveat will have been a technicality rather than a weakness. If instead the effect shrinks outside a trial population, it was the main thing on the page.
• The label proves usable across its full width. The licence covers adults and children aged 2 and over, while the main efficacy population in SAPPHIRE was 156 patients aged 2 to 12, of whom only 53 received the approved 10 mg/kg dose, all of them non-ambulatory with later-onset Type 2 or Type 3 disease and an HFMSE score between 10 and 45. If payers and prescribers treat the whole labelled population as eligible, the addressable number is far larger than the evidence base. If instead reimbursement settles on the ages that generated the result, the commercial case narrows without any new clinical fact.
• The cash funds the launch without heavy issuance. Liquidity was $492.1 million at June 30, 2026 against first-half operating cash use of $152.4 million, with $200.0 million drawn of a $350 million Blue Owl facility and a covenant requiring unrestricted cash of at least 60 per cent of outstanding principal, which on that balance is $120 million. A fresh $200 million ATM supplement was filed on August 6. If the next two quarterly filings show the share count broadly stable and the ATM barely used, the dilution part of this reading was pessimistic. The share count moved from 108,461,354 to 121,600,396 in the first half of 2026, so the direction of travel is the thing to watch.
• The September 15 correction was an isolated slip. Management told investors on the approval call that the FDA had imposed no post-marketing commitments on bone or reproductive safety, and the company filed an 8-K the next day stating that this was said in error and that a post-marketing requirement and a separate FDA request both exist. One correction, promptly filed, is a company doing the right thing after getting something wrong. A pattern of corrections on safety or commercial disclosure would be a different matter, and it is the kind of thing that only becomes visible over several quarters.

None of these is a prediction. They are the observations that would make the rest of this page wrong, listed so that a reader can check them rather than take the reading on trust.

02 Executive answer: an approval arrived early and the shares still fell

On September 11, 2026 Scholar Rock announced that the FDA had approved ISEMBYLD (apitegromab-mstn) for the treatment of spinal muscular atrophy in adults and children two years of age and older who are currently receiving a survival motor neuron 2 targeted treatment. It is the first and only muscle-targeted treatment licensed in the disease. One point of precision: the 8-K records that on September 11 the company announced that the FDA had approved the product, and the label is dated only “Issued: 09/2026”, so September 11 is the announcement date and the approval letter may carry an earlier one. The tape is consistent with news reaching the market outside trading hours: September 11 closed down 0.5 per cent on 994,400 shares, and the fall came on Monday September 14, down 6.4 per cent to $51.85 on 4,041,900 shares. The announcement arrived nineteen days before the Prescription Drug User Fee Act action date of September 30, and it came with a Rare Pediatric Disease Priority Review Voucher, an asset that can be sold to another company for use on a future application.

For a company whose entire value rested on one regulatory decision, that is as good an outcome as the calendar allowed. And over the month that contained it the shares fell 14.93 per cent, closing at $48.32 on September 30 against a reported twelve-month gain of 29.75 per cent and a gain of 9.69 per cent since the start of the year. On the day the PDUFA clock formally ran out the stock traded 1,285,235 shares, below its typical 1.45 million, and closed up 1.19 per cent. Nothing in the tape suggests a market that was surprised by anything on September 30, because the event had already happened three weeks earlier.

The explanation is not hidden and it is not sentiment. It is in the approval package and in the company’s own filings, and it comes in four parts.

First, the strength of the evidence is qualified by the company itself. The pivotal SAPPHIRE study met its primary endpoint with a 2.2-point improvement on the Hammersmith Functional Motor Scale-Expanded against placebo, and 34.2 per cent of treated patients gained at least three points against 13.5 per cent on placebo. Both results are reported with an asterisk marking them as nominal p-values — 0.0121 and 0.0125 respectively — meaning they are not adjusted for multiple comparisons. The label states that the main efficacy population was 156 patients aged 2 to 12, and that within it 53 patients received at least one dose of the recommended 10 mg/kg and 50 received placebo — the 103-patient comparison on which the primary result rests. Fifty-three patients is the size of the evidence base at the licensed dose, inside a trial that enrolled 188 patients aged 2 to 21 across nine countries. Every one of them was non-ambulatory at baseline, with later-onset Type 2 or Type 3 disease and an HFMSE score between 10 and 45 at screening, on nusinersen for at least ten months or risdiplam for at least six. The approved indication carries none of those conditions.

Second, the label carries a bone-safety warning. Fractures occurred in 9 per cent of patients on the approved 10 mg/kg dose against 2 per cent on placebo. The prescribing information carries this in Warnings and Precautions: ISEMBYLD may increase the risk of fractures, including serious fractures; some fracture events did not have a clear precipitating event such as a fall; prescribers are told to use the drug with caution in patients with a history of low bone density or multiple fractures and, in patients who experience fractures, to weigh the risks and benefits of continuing. In the trial no patient discontinued treatment because of a fracture. The patient-facing Important Safety Information in the approval release puts the same points more plainly, including that a provider may consider stopping treatment after a fracture and that patients should disclose a history of low bone density or fractures before starting.

Third, the FDA attached two post-marketing demands to that risk — and the company told investors it had not. On the September 14 approval call management stated that the agency required no post-marketing commitments regarding bone or reproductive safety. On September 15 Scholar Rock filed a Form 8-K stating that this was said in error, and setting out what the FDA actually did. The filing distinguishes the two: the agency outlined a post-marketing requirement that the company establish a pregnancy safety study to collect and analyse pregnancy complications and birth outcomes in exposed pregnancies, and separately requested that it submit all serious and non-serious domestic and foreign fracture cases as 15-day alert reports. The error fell on the single most sensitive point of the label.

Fourth, there is no revenue yet and the spending has changed shape. The last reported quarter ended on June 30, 2026, ten and a half weeks before the approval, so the trailing revenue field at the September 30 reading is zero against a trailing net result of −$408.59 million. Liquidity was $492.1 million. Within the first-half expense line, general and administrative costs rose 29 per cent year on year to $100.9 million while research and development was flat at $110.0 million — the commercial build arriving in the accounts before the product did.

So the answer to the question in the title is that the market is not pricing the approval, which it had already priced. It is pricing the distance between a licence and a business: a monthly intravenous infusion added on top of an existing therapy, sold into a label that warns about fractures, by a company with 289 employees, no commercial history, $492 million of liquidity and a $5.88 billion equity value. Nineteen per cent of the float is short at a ratio of 13.4 days, and that position did not close on the approval.

None of that is a judgment on whether the shares are worth their price. It is a statement of what has to be proved next, and of where the proof will appear: in quarterly product revenue, in European regulatory progress on the application resubmitted (announced October 8, 2026), and in the post-marketing fracture reporting that the FDA has now required by name.

03 What exactly was approved, and for whom

Precision about the label matters here more than usual, because the licensed population, the evidence population and the commercially addressable population are three different things on this file.

ISEMBYLD is apitegromab-mstn, a fully human monoclonal IgG4 antibody that binds promyostatin and latent myostatin and inhibits the activation of myostatin, blocking myostatin signalling. Myostatin is a negative regulator of muscle mass, and the company states that its inhibition results in increased muscle mass and strength; what SAPPHIRE demonstrated, though, is a change in motor function on the Hammersmith scale, not a measured gain in muscle mass or strength. The clinical logic in spinal muscular atrophy is that existing therapies address the motor neuron and not the muscle: nusinersen and risdiplam are designed to prevent motor neuron loss, and ISEMBYLD is the first approved treatment to target the muscular component directly.

The approved indication is the treatment of spinal muscular atrophy in adults and paediatric patients two years of age and older who are currently receiving a survival motor neuron 2 targeted treatment. Two features of that sentence do the work. It is an add-on licence, not a monotherapy licence: a patient has to be on an SMN2-targeted treatment already. And it is age-broad, covering adults as well as children from two years upwards.

The recommended dose is 10 mg/kg by intravenous infusion. In SAPPHIRE the infusions were given once every four weeks for approximately one year. The approval release states that, based on eligibility, infusions can be given at convenient locations including hospital, home or an infusion centre, which matters commercially because a monthly hospital visit and a monthly home infusion are different propositions for a family and for a payer.

Alongside the approval the FDA awarded Scholar Rock a Rare Pediatric Disease Priority Review Voucher. The release states that it may be used to obtain priority review for a future marketing application; vouchers of this kind are transferable under the Federal Food, Drug, and Cosmetic Act, which makes it an asset with a market price independent of anything ISEMBYLD does commercially, though the release itself does not say so. The company had previously disclosed Rare Pediatric Disease designation for apitegromab and had warned in its risk factors that a marketing application, if approved, might not meet the eligibility criteria for a voucher. It did.

What was not approved is equally worth stating. The approval is United States only. There is no approval in Europe: the original marketing authorisation application, validated in March 2025, was withdrawn in August 2026 and resubmitted (announced October 8, 2026), with a CHMP opinion anticipated in the first half of 2027. The subcutaneous formulation of apitegromab is not approved; a Phase 1 study in healthy volunteers has been completed and the company describes planned FDA and EMA regulatory engagements on it following potential approvals. Apitegromab is not approved in facioscapulohumeral muscular dystrophy, where it is in Phase 2. And the approval says nothing about price: the approval release describes payer engagement and a patient support programme but discloses no list price, and Merlintrader has found no public disclosure of a wholesale acquisition cost.

Approval release, September 11, 2026

04 The regulatory path: a rejection caused by someone else’s factory

The twenty-three months between the pivotal data and the approval are the most useful thing on this file for judging management, because the delay had nothing to do with the drug.

What this sequence says, and what it does not

It says that the company’s clinical package survived a full review cycle, a rejection and a resubmission without the agency raising a question about efficacy or safety as grounds for the Complete Response Letter. That is a meaningful piece of evidence about the asset, and it is the reason the second review ran short: by March 2026 the only open question was whether the product could be reliably filled and finished.

It also says that Scholar Rock lost roughly a year to a supplier. A Form 483 in July 2025, an Official Action Indicated classification, a CRL in September, a Warning Letter to the facility in November: that is a contractor deteriorating under inspection while the sponsor’s application sat behind it. The company’s response — a Type A meeting with the facility in the room, then a resubmission with two sites rather than one — is the correct response, and it is the single clearest piece of execution evidence available on this management team.

What the sequence does not establish is anything about commercial performance. A company can be good at regulatory recovery and poor at selling, and the skills have almost nothing in common. It also does not mean the manufacturing question is closed: the product is now licensed with two fill-finish facilities, and the quality record of contract manufacturers is a live operational risk for any biologic, not a box that was ticked in March.

Form 10-Q, quarter ended June 30, 2026 — regulatory history

05 SAPPHIRE, and what a nominal p-value means

SAPPHIRE is registered as NCT05156320. It was a global, multi-national, randomised, double-blind, placebo-controlled Phase 3 trial in 188 patients across nine countries, all with a diagnosis of 5q SMA and aged 2 to 21 years. Within the main efficacy population aged 2 to 12, patients were randomised one to one to one to receive apitegromab at 20 mg/kg, which is twice the recommended dose, apitegromab at 10 mg/kg, which is the recommended dose, or placebo, by intravenous infusion once every four weeks for approximately one year. The registry record is explicit that the exploratory subpopulation aged 13 to 21 was randomised between apitegromab 20 mg/kg and placebo only: there was no 10 mg/kg arm above the age of twelve, so no patient older than twelve received the approved dose in SAPPHIRE. Every patient in the trial was non-ambulatory at baseline, with later-onset Type 2 or Type 3 SMA diagnosed before starting an SMN upregulator, an HFMSE score between 10 and 45 at screening, and at least ten months of nusinersen or six months of risdiplam behind them. Every patient in the trial was already receiving an approved SMN2-targeted treatment, either nusinersen or risdiplam.

The headline results, as the company reports them:

• A 2.2-point improvement on the Hammersmith Functional Motor Scale-Expanded at one year in patients receiving 10 mg/kg plus an SMN2-targeted treatment, compared with patients receiving the SMN2-targeted treatment alone. Nominal p=0.0121. The label states that the main efficacy population was the 156 patients aged 2 to 12, within which 53 received at least one dose of 10 mg/kg and 50 received placebo — the 103-patient comparison the approval release describes.
• 34.2 per cent of treated patients showed an increase of at least three points on that scale, against 13.5 per cent of placebo patients, an odds ratio of 3.8. Nominal p=0.0125.
• A safety database of more than 500 individuals across all apitegromab studies globally, some on treatment for more than seven years.
• Ninety-eight per cent of SAPPHIRE participants elected to continue into the ONYX long-term extension study.

The asterisk

Both efficacy figures in the approval release carry an asterisk, and the footnote reads “nominal p-value”. This is the company being accurate, and it deserves to be explained rather than skipped.

A nominal p-value is the raw probability calculated for a single comparison, without any adjustment for the fact that several comparisons were made. SAPPHIRE tested two doses against placebo, and it reported results in a main efficacy population narrower than the enrolled population. When a trial makes multiple comparisons, the chance of at least one of them crossing the conventional 0.05 threshold by chance alone rises above 5 per cent, and the statistical convention is to control for that — by allocating the available error rate across the comparisons in a pre-specified order, for example. A result described as nominal has not had that control applied, or the sponsor is not claiming it.

What follows from that, honestly stated, is narrow. It does not mean the result is false. It does not mean the FDA was lax: the agency saw the full dataset, including the 20 mg/kg arm, the safety database and the long-term extension, and it licensed the 10 mg/kg dose with a broad age label. It does mean that a reader should not treat p=0.0121 as the same object as a multiplicity-controlled p=0.0121, and that the durability of the effect is a question real-world use will answer rather than one SAPPHIRE has closed. The ONYX extension will not answer it for the licensed dose, because every patient in ONYX receives 20 mg/kg.

The gap between the evidence population and the label

The more consequential number is 53: the patients who received the approved dose, all of them aged 2 to 12, all non-ambulatory, all with an HFMSE score between 10 and 45. The trial enrolled up to 21 years, but only at twice the approved dose above twelve. The label covers adults and children from two years upwards, with no requirement that a patient be non-ambulatory, have Type 2 or Type 3 disease, or sit in any HFMSE band. Cure SMA’s president, quoted in the approval release, explicitly welcomed the FDA’s decision as one that “supports access for a broad population within the SMA community” — a fair reading of a licence that is wider than the population in which the primary result was generated.

The label does state the basis for the extension. Section 8.4 says the indication is supported by the adequate and well-controlled study in patients 2 to 12 years of age together with latent myostatin levels in patients two years and older and pharmacokinetic modelling comparing patients 2 to 12 with patients 13 and older. Section 8.5 adds that clinical studies did not include patients 65 years of age and older, so the licence is not simply open-ended at the top: above 65 the label declares an absence of data.

For a patient advocacy organisation that breadth is the point: adults with SMA have been waiting for something that improves motor function. For anyone modelling the commercial case it introduces a question no clinical document answers, which is whether payers will reimburse across the full label or settle on the population that produced the evidence. That question will be answered by coverage policies, not by trial data, and coverage policies are published by payers rather than by Scholar Rock.

ClinicalTrials.gov — SAPPHIRE (NCT05156320)

06 The fracture signal, the labelled warning and the FDA’s two post-marketing demands

This is the section that most deserves to be read slowly, because the numbers are small, the language is specific, and the FDA has attached obligations to it by name.

Three things about that warning matter commercially and not just clinically.

The first is that some fracture events had no clear precipitating event. A fracture that occurs without trauma is a different clinical event from a fracture caused by an accident, and the label is explicit that both occurred. In a population with reduced mobility and, in many cases, pre-existing low bone density, distinguishing a drug-related fracture from a disease-related one in an individual patient will be difficult, and the label does not pretend otherwise.

The second is the instruction to weigh the risks and benefits of continuing. A label that puts the continuation decision in front of a prescriber after each fracture introduces a point of attrition into any commercial model that assumes patients stay on therapy — although the label also records that in the trial no patient actually discontinued because of a fracture. Nine per cent over a mean exposure of 49 weeks is the figure at the licensed dose, and it was measured in a controlled study with monitoring that commercial practice will not replicate exactly, in 53 patients.

The third is that the mechanism makes the signal plausible rather than surprising. Myostatin inhibition increases muscle mass; the interaction between rapidly changing muscle loading and bone in a population with low baseline bone density is precisely where a sceptical reviewer would look. The FDA appears to have looked there, which is what the post-marketing demands indicate.

What the FDA required

Two obligations are documented in the Form 8-K filed on September 15, 2026, and the filing is careful to distinguish them: one is a post-marketing requirement, the other a separate request.

• A pregnancy safety study. The FDA outlined a post-marketing requirement that the company establish a study to collect and analyse information on pregnancy complications and birth outcomes for exposed pregnancies. The label documents why. Section 8.1 reports that intravenous apitegromab-mstn given to rats throughout pregnancy and lactation produced delayed sexual maturation and altered neurobehavioral function in offspring of both sexes at all doses and impaired reproductive function in male offspring at the highest dose, with no no-effect dose identified and the lowest dose tested associated with maternal exposures below those in humans at the recommended dose. Section 8.3 states that animal data suggest adverse effects on male and/or female reproductive function, and the juvenile animal data in section 8.4 report decreased corpora lutea, increased preimplantation loss and fewer implantation sites and live fetuses, again with no no-effect dose identified. The patient-facing safety information adds that it is not known whether ISEMBYLD will harm an unborn baby or whether it passes into breast milk, and tells patients to inform their provider immediately if they become pregnant during treatment.
• Fifteen-day alert reports for every fracture — a request, not a requirement. The 8-K distinguishes the two, and this is the second. The FDA requested that the company submit all serious and non-serious domestic and/or foreign cases of fractures as 15-day alert reports. The scope of that request is the notable part: not serious fractures, not United States fractures, but all of them, worldwide, on an expedited clock normally reserved for serious unexpected events.

The company states that it intends to conduct the study and submit the reports in compliance with the agency’s request. Both obligations are costs and both are information: a fracture reporting stream of that breadth means the post-marketing safety picture will accumulate quickly, and it will accumulate where the FDA can see it before anyone else can.

Form 8-K, September 15, 2026 — post-marketing requirement and request

07 September 15: the company files a correction against its own management

On Monday September 14, 2026 at 8:00 a.m. Eastern Time Scholar Rock hosted a live audio webcast to discuss the approval. During the question-and-answer portion of that call, in the company’s own words from the subsequent filing, “the Company’s management stated in error that the FDA did not require any post-marketing safety commitments regarding bone or reproductive safety of ISEMBYLD.”

The following day the company filed a Form 8-K setting out the requirement and the request described in the previous section. The filing is signed by the General Counsel and Corporate Secretary.

It is worth being fair about what this is and is not. It is not a restatement, a trading suspension or an enforcement matter. A company that discovers an incorrect statement on an investor call and files a correction within a day is doing the thing a well-run company is supposed to do, and the filing is unambiguous rather than hedged: it says management was in error and then states the facts. The correction itself is evidence of a functioning legal and compliance function.

What it is, equally, is a misstatement on the single most sensitive point of the approval, made to investors, on the day the company had chosen to explain the approval. Of all the things on which management could have been wrong on that call — launch timing, payer conversations, manufacturing capacity, the voucher — it was wrong about whether the regulator had imposed obligations on bone safety, in a product whose label warns about fractures at 9 per cent against 2 per cent.

Two readings are available, and the evidence does not yet allow a choice between them. One is that a question came from the floor, an executive answered from memory about a document finalised days earlier, and the company corrected it immediately: a process failure of the smallest kind. The other is that a company preparing its first commercial launch described its own label to investors more favourably than the label reads, and was corrected only because the discrepancy was checkable. The difference between those two readings is a pattern, and a pattern needs more than one data point.

For a reader the practical consequence is narrow and concrete: on this file, the primary documents are worth more than the call. The label, the approval release and the 8-K say what they say. Where this page quotes a commercial or safety fact, it is sourced to a filing rather than to management commentary, and that is a deliberate choice following September 15.

08 The launch: what has to happen for a licence to become revenue

The approval release describes a launch that was already in motion on the day of approval. Product was stated to be available to ship “in the coming days”. Scholar Rock Supports, a dedicated patient support team, was described as available immediately, offering help with understanding insurance coverage, financial assistance for eligible patients, and disease and treatment education, as well as help navigating site-of-care options and infusion logistics. The company said it was working closely with top commercial and government payers to establish reliable, broad access.

That is the right set of activities and it is also, at this stage, a list of intentions rather than results. What a reader can check against documents, and what has to be true for the commercial case to work, breaks into four pieces.

Access

ISEMBYLD is an add-on to an existing therapy. A patient is already on nusinersen, which is given by intrathecal injection, or risdiplam, which is taken orally. Adding a monthly intravenous infusion to either of those requires a payer to agree to fund two treatments for one disease. Merlintrader has found no public disclosure of a list price, which means the single most important commercial variable is not yet public. Coverage policies from large commercial payers and from state Medicaid programmes are the documents that will settle this, and they are published by the payers.

Site of care

The release states that infusions can be given at hospital, at home or at an infusion centre, based on eligibility. That flexibility is commercially significant: a monthly hospital appointment for a child with a mobility impairment is a real barrier, and home infusion removes much of it. The qualifier “based on eligibility” is doing work, because home infusion eligibility is determined by payers and by clinical judgement rather than by the label.

The prescriber base

SMA is managed by a small number of specialised neuromuscular centres. That concentration cuts both ways: a commercial team of modest size can reach most of the prescribers, which suits a company with 289 employees, and the same concentration means that a handful of centres forming a cautious view on the fracture warning would be visible in the numbers quickly.

The size of the opportunity

The company states that approximately 35,000 SMA patients globally have been treated with an SMN-targeted treatment. That figure is the anchor for any sizing exercise and it deserves three caveats, all of which follow from the label rather than from scepticism. It is global, and the approval is United States only. It is cumulative — patients who have been treated — rather than a count of patients on therapy at any one date. And the label requires that a patient be currently receiving an SMN2-targeted treatment, so patients who discontinued are outside it.

09 Europe: an application withdrawn and resubmitted

Scholar Rock submitted a marketing authorisation application for apitegromab in SMA to the European Medicines Agency and received validation in March 2025. In the Form 10-Q for the quarter ended June 30, 2026 the company states that if apitegromab is approved by the FDA and/or the European Commission it expects to initiate a commercial product launch in the applicable jurisdictions upon approval. That application did not reach a decision: in August 2026 the company informed the CHMP that it was withdrawing it, and it then resubmitted, as announced on October 8, 2026 and described below.

Two observations follow. The first is that the original European file stayed open for about seventeen months before it was withdrawn, and the United States file was rejected in September 2025 over a third-party fill-finish facility that subsequently received a Warning Letter. The company discloses the link between the two directly, and in strong terms. The 10-Q states that the Complete Response Letter could impact the current review process of the MAA with the EMA and could result in a delay or a resubmission; that under the EMA’s mutual recognition agreement with the FDA the company believes satisfactory resolution of the facility’s Official Action Indicated classification will be necessary for MAA approval; and that in the most adverse scenario GMP non-compliance at the facility could result in a negative CHMP opinion and refusal of the MAA by the European Commission. The second-quarter results release of August 6, 2026 names the site: the MAA includes the Catalent Indiana fill-finish facility, the EMA is awaiting the FDA’s inspection classification of it, the company is engaging with the EMA on next steps including the potential addition of its second fill-finish facility to the MAA, and it plans to give updated guidance on CHMP timing once aligned with the agency.

The second is that Europe is where the mobility of the opportunity lies. The approximately 35,000 patients globally treated with an SMN-targeted therapy are not concentrated in the United States, and SMA newborn screening programmes and reimbursement for nusinersen and risdiplam are established across much of Western Europe. An approval there would roughly change the shape of the addressable population rather than adding to it at the margin.

Update, October 8, 2026. A related outcome materialised: the application was withdrawn rather than refused, and then resubmitted. Scholar Rock’s release of October 8 states that in August 2026 the company informed the CHMP that it was withdrawing the previous application because of a lack of GMP compliance at the Catalent Indiana fill-finish facility, and that it has now resubmitted the application, in line with EMA guidance, with its alternate fill-finish facility in place of Catalent Indiana: the same U.S. site that supported the September 2026 FDA approval of ISEMBYLD, which the company says has a record of recent successful FDA and EMA inspections. The company anticipates a CHMP opinion in the first half of 2027 and says vials are already held at a third-party provider awaiting packaging and labelling for European commercial use upon approval. The resubmission removes the dependency on the facility that caused the problem; it does not shorten the review that has to run again, and the timing is the company’s expectation rather than an agency commitment. Company release

On Japan, Merlintrader has found no public disclosure of a marketing application. What they do disclose is patent litigation: two Japanese patents, JP6823167 and JP7344337, are the subject of invalidation trials, and three European patents, EP2981822, EP3816625 and EP4358995, are subject to opposition proceedings. The company notes that such proceedings are expensive and could result in revocation or amendment of patents in ways that no longer cover its product candidates. That is boilerplate risk-factor language around a specific and non-boilerplate list of patent numbers, and intellectual property on the lead asset is worth watching for a company whose entire value now rests on one marketed product.

Form 10-Q, quarter ended June 30, 2026 — EMA application and patent proceedings

10 The pipeline beyond ISEMBYLD, and the list the company wants to partner away

Scholar Rock describes itself as building a world-leading anti-myostatin pipeline, and the Form 10-Q for the second quarter of 2026 divides what it owns into things it is developing and things it is trying to find partners for. That division is the most informative thing about the pipeline, because it tells a reader where management is spending.

What the company is developing itself

• ISEMBYLD in SMA, now approved, with the ONYX long-term extension study continuing. Ninety-eight per cent of participants treated in SAPPHIRE elected to enter ONYX, but every patient in ONYX receives apitegromab 20 mg/kg every four weeks for 104 weeks — twice the approved dosage — so the extension is the main source of multi-year data on apitegromab, not on the licensed dose.
• OPAL (NCT07047144, cited in the company’s August 6, 2026 results release), a Phase 2 trial of apitegromab in SMA patients under two years of age, initiated in the third quarter of 2025, with a population that includes patients treated with SMN1-targeted gene therapy. This is the age extension below the current label.
• FORGE, a Phase 2 trial of apitegromab in facioscapulohumeral muscular dystrophy, initiated in the second quarter of 2026: approximately 60 patients randomised one to one to apitegromab 10 mg/kg intravenously or placebo every four weeks for 52 weeks, with the change in lean muscle volume at twelve months as the primary endpoint. This is the indication extension, and it is the test of whether muscle-targeted myostatin inhibition generalises beyond SMA.
• A subcutaneous formulation of apitegromab. A Phase 1 study in healthy volunteers has been completed and demonstrated favourable bioavailability and a pharmacodynamic profile comparable to intravenous administration. Further development is ongoing, including planned FDA and EMA regulatory engagements following potential approvals. Commercially this is the most interesting item in the list: it would turn a monthly infusion into an injection, and most of the access friction described in the launch section is a consequence of the infusion.
• SRK-439, a subcutaneously administered fully human anti-pro/latent myostatin antibody with high inhibitory potency while maintaining selectivity towards myostatin, in development for rare, severe and debilitating neuromuscular diseases. A Phase 1 study in healthy volunteers is under way with topline data anticipated in the second half of 2026. That is the nearest dated milestone on the whole file and the window is open now.

What the company is seeking partners for

The 10-Q states that as the company focuses its strategy on rare neuromuscular diseases it is currently seeking partnerships for its additional programmes. These are:

• SRK-181, a Phase 2-ready inhibitor of latent TGFβ1 for solid tumours resistant to anti-PD-(L)1 antibody therapy.
• SRK-373, a highly selective inhibitor of the latent TGFβ1 isoform with selective activity in the fibrotic extracellular matrix, in preclinical development for fibrotic diseases.
• SRK-256, an inhibitor of RGMc, or hemojuvelin, in preclinical development for iron-restricted anaemias.
• Myostatin inhibition combined with GLP-1 weight loss, following the positive Phase 2 EMBRAZE study, which the company describes as demonstrating proof of concept in the ability of apitegromab to drive statistically significant preservation of lean mass during tirzepatide-induced weight loss.

EMBRAZE deserves a note of its own, because it is the item on this list with the largest theoretical value and the least company commitment. Lean-mass preservation during GLP-1 weight loss is one of the most contested commercial questions in pharmaceuticals, and Scholar Rock has a positive Phase 2 proof of concept in it. The company is not developing it; it is looking for a partner. A reader can take that two ways — focus, or an admission that the company cannot fund a trial of the size that indication would require — and the filings do not settle which. What they do settle is that no partnership has been announced, and that nothing in the current valuation needs to be attributed to it.

One further detail from the 10-Q is worth recording for the way it was disclosed rather than for what it says: the company notes that apitegromab was evaluated both in SAPPHIRE in Type 2 and Type 3 SMA and in EMBRAZE in combination with tirzepatide in obesity, and that topline data from each have been released, with apitegromab not interacting in any new ways despite the different patient populations and disease indications. That is a safety-consistency statement across two very different populations, and it is the sort of cross-reading the company is entitled to make while a reader should still treat each trial’s safety profile as its own.

11 Financial position at June 30, 2026: funded, and spending like a company about to launch

All the figures in this section come from the Form 10-Q for the quarter ended June 30, 2026, filed on August 6, 2026. They predate the approval by ten and a half weeks, which is why none of them contains product revenue.

Where the money went in the first half

The interesting movement in that table is not the total, it is the mix. Research and development was flat year on year over the half, at $110.0 million against $111.1 million, and actually fell in the second quarter. General and administrative rose 29 per cent over the half, from $78.1 million to $100.9 million, and in the second quarter it reached $50.7 million against research and development of $58.2 million. The two lines have almost converged.

The company explains this directly in its own management discussion: it expects general and administrative expense to continue to be substantial as it invests in building the infrastructure to support the commercialisation of apitegromab. That is the correct thing for a company five weeks away from an approval to be doing, and it is also a commitment that does not reverse. A commercial organisation built before a launch is a fixed cost from the quarter it is hired, and it will be in the accounts whether or not the revenue arrives on schedule.

Net loss for the first half was $215.4 million, against $184.8 million in the first half of 2025. Cash used in operating activities was $152.4 million. Of the $63.0 million gap between the loss and the cash out, $37.9 million is equity-based compensation — a real cost to shareholders expressed as dilution rather than as cash — $17.0 million is changes in operating assets and liabilities, mostly a $14.9 million increase in accrued expenses not yet paid, and the remaining $8.1 million is other non-cash items, including the $3.3 million loss on extinguishment of the Oxford debt. Only the first of those three is a permanent non-cash cost; the second is a timing difference that reverses when the invoices are settled.

Runway, stated carefully

The 10-Q makes two separate statements, and both are worth having. In the notes the company asserts that its existing cash, cash equivalents and marketable securities as of June 30, 2026 will be sufficient to fund current operations through at least a period of one year after the date the financial statements are issued — the standard accounting assertion, and a floor. In the management discussion it goes further: based on its current operating model it expects those resources to fund operating expenses and capital expenditure requirements into the second half of 2027, while adding that additional capital will be required to complete clinical development and commercialisation.

The arithmetic a reader can do independently runs past that: $492.1 million of liquidity against first-half operating cash use of $152.4 million is about nineteen months at that rate, which reaches the end of January 2028. The company’s own estimate is the shorter of the two, which is a reason to prefer it. Three things explain the difference. Operating cash use will rise, because a commercial organisation is being paid for a full year rather than a partial one. It will also be offset, for the first time, by product revenue of an unknown amount. And $120 million of the liquidity is encumbered in effect rather than in form, because of the covenant described in the next section.

Form 10-Q, quarter ended June 30, 2026

12 The Blue Owl facility, and the covenant that locks up cash

In February 2026 Scholar Rock entered into a loan agreement with Blue Owl, replacing the Oxford loan agreement — originally a 2020 agreement with Oxford Finance and Silicon Valley Bank, amended and restated with Oxford alone in February 2025 — which was terminated on February 27, 2026. The terms as disclosed in the second-quarter 10-Q:

• Size. Up to $350.0 million of term loans, of which $100.0 million was received as an initial term loan in February 2026 and a further $100.0 million as the next available tranche on March 31, 2026, bringing the outstanding balance to $200.0 million at June 30, 2026. The balance sheet confirms it: long-term debt of $196.2 million against $99.7 million at December 31, 2025, with proceeds from the debt facility of $197.7 million in the cash flow statement for the half.
• What is left, and on whose option. The remaining $150.0 million under the $350 million plafond is a delayed draw term loan commitment the company may borrow at its sole option, available from the date of FDA approval for apitegromab in SMA until the earliest of full usage, termination or September 30, 2027, subject to customary conditions including the minimum cash covenant. That window opened with the approval. Separately and in addition, an uncommitted incremental facility of up to $200.0 million exists but requires Blue Owl’s consent, which makes that one an option held by the lender rather than by the borrower.
• Cost. Three-month Term SOFR, subject to a 1.00 per cent per annum floor, plus 5.00 per cent; the agreement also provides a base-rate option, with the base rate not less than 2.00 per cent, plus 4.00 per cent. Interest is payable quarterly, on any prepayment date and at maturity.
• Maturity. The earlier of February 26, 2032 and the date the loans otherwise become due and payable in full. The company may prepay at any time, in whole or in part, at its option.
• Covenant. The company must maintain a minimum unrestricted cash balance of at least 60 per cent of the aggregate outstanding principal amount, unless certain milestones are met and maintained.

The covenant is the term that changes how the balance sheet should be read. On the $200.0 million drawn at June 30, 2026 it requires at least $120 million of unrestricted cash to stay in place — roughly a quarter of the $492.1 million of liquidity, which is no longer a rounding item. Read forward it tightens further: if the company drew the remaining $150 million delayed draw and reached the full $350 million, the same covenant would require $210 million of unrestricted cash to sit untouched, which means the incremental borrowing capacity delivers substantially less incremental spendable cash than its headline suggests. Debt capacity on this facility is not the same thing as free cash, and the escape clause is milestone-based rather than discretionary.

The rate deserves a sentence too. Term SOFR plus 500 basis points on a 2032 maturity is venture-style pricing, appropriate for a company that had no approved product when the facility was signed in February 2026 and was waiting on a resubmitted BLA. It is an expensive but non-dilutive way to fund a launch, and the sequence is explicit in the filing rather than a matter of inference. The February initial term loan was used to repay all outstanding Oxford obligations, totalling $103.7 million, including a $2.0 million final payment and a $1.0 million prepayment fee, at a $3.3 million loss on extinguishment recognised in the half. It brought in no new money: it refinanced existing debt. The new money arrived with the $100 million tranche drawn on March 31, 2026.

Total debt to equity stood at 0.80 in third-party data as of September 30, 2026, with long-term debt to equity at 0.78 — consistent with the $196.2 million of long-term debt against $253.8 million of equity, and itself the confirmation that the drawn balance is $200 million rather than $100 million. The current ratio was 7.02. This is not a stretched balance sheet. It is a balance sheet where the leverage is small in absolute terms, expensive per dollar, and attached to a cash-retention requirement that scales with any further drawing.

13 The ATM and the dilution arithmetic

Scholar Rock has funded itself with equity for its whole life, and the second-quarter filings make the mechanics unusually easy to follow. The company has a sales agreement with Jefferies LLC, dated November 14, 2022, under which it may offer and sell shares at its sole discretion through Jefferies as sales agent — an at-the-market programme. Authorised share capital is 300,000,000 shares.

Read in order, those rows tell a clear story. In the first half of 2026 the company added about 13.1 million shares, a 12.1 per cent increase. Roughly 3.5 million of those came from ATM sales raising $160.8 million net; 6.8 million came from the exercise of pre-funded warrants, which brings in little or no cash because the price was paid when the warrants were issued; the remainder came from option exercises and the vesting of restricted and performance stock units. So the share count rose for three distinct reasons, only one of which put meaningful money in the bank.

The August 6 prospectus supplement is the forward-looking item. It replaces and supersedes the November 2025 supplement in its entirety, and the offering under the prior supplement has been terminated. It covers up to $200 million of common stock, filed under the shelf registration statement on Form S-3ASR, File No. 333-282530, which was filed on October 7, 2024 and became automatically effective on filing. The filing also discloses that approximately $257.7 million of net proceeds has been raised under the programme in total to that date.

Two points of precision, because this is the kind of place where a page can mislead without stating anything false. The $200 million is the amount registered and available under the new supplement, not a plan to sell it; an ATM is an option the company holds, used at its discretion. And the timing is worth noting without over-reading: the supplement was filed on August 6, 2026, five weeks before the approval, at a point when the company had a PDUFA date of September 30 and a resubmitted application. Putting capacity in place before a binary event, rather than after it, is ordinary treasury practice.

What a reader should watch is simple and will be visible in the next two quarterly filings: the share count, and the ATM proceeds line. A company that has just been approved and has $492 million of liquidity does not need to sell shares immediately. A company funding a commercial organisation, an infusion product launch, two Phase 2 trials, a long-term extension, a subcutaneous programme and a Phase 1 for SRK-439, while paying Term SOFR plus 500 basis points on $200 million of debt, will need capital again at some point. The question is at what price and in what quantity, and the answer appears in the filings rather than in guidance.

Form 8-K, August 6, 2026 — ATM prospectus supplement

14 Ownership, short interest and what the September 30 session actually showed

The August ownership amendments report Invus Public Equities at 9,575,013 shares (8.0%) and Redmile Group at 4,311,092 (3.6%), using their dated filing bases. Invus Schedule 13G/A; Redmile Schedule 13G/A.

The market figures in this section are third-party data as of September 30, 2026, not filed numbers. They are reported with that status because none of them can be reconstructed from an SEC document at that date.

The institutional figure above 100 per cent

A reported institutional ownership of 101.68 per cent cannot be a share of the register, and it is worth saying why rather than quietly repeating it. Figures of this kind are built by summing institutional positions from quarterly 13F filings and dividing by a share count taken at a different date. When the share count used as the denominator is older than the positions in the numerator — and this company’s share count rose 12.1 per cent in six months — the quotient can exceed 100 per cent. The honest statement is that institutional ownership is high, which the direction of the field supports, and that the precise percentage is an artefact of mismatched dates.

The short position is the most informative field

Nineteen point three per cent of the float, 19.41 million shares, at a short ratio of 13.40, meaning thirteen days at the 1.45 million shares a day that market data showed as typical on September 30, 2026. That position existed through an FDA approval and is still there. Short interest of that size against a specialist biotechnology company with a newly approved product is a structural fact about this file rather than a day’s sentiment, and it cuts in both directions. It means that good commercial news has a mechanical amplifier behind it, because thirteen days of volume cannot be covered quickly. It also means that a significant pool of capital has looked at the same approval, the same label and the same cash position as everyone else and reached the opposite conclusion, and that pool is not small or inattentive.

The consensus target of $69.12 sits on the other side of the same question: 43 per cent above the close, with an aggregated recommendation score of 1.11, which on the data source’s one-to-five scale means the covering analysts are almost uniformly positive. A 19 per cent short position and an almost uniformly positive sell-side aggregate describing the same security is the clearest possible statement that the commercial outcome is genuinely unresolved. Neither number is adopted on this page.

15 Insiders after the approval: two transfers, and neither one is a sale

Katie Peng’s October 1 Form 144 lists a proposed sale of 1,006 shares with an aggregate estimated market value of $48,610 under a plan adopted March 20, 2026. A Form 144 is notice of an intended sale; it does not establish that a sale was executed. Form 144.

Two Forms 4 were filed in the days after the approval, and both are easy to misread. They are not sales.

• David Hallal, Chairman and Chief Executive Officer. On September 21, 2026 he transferred 36,652 shares of common stock to The Hallal Family Irrevocable Trust-2021 for no consideration. Transaction code G, a bona fide gift, at a price of $0.00. Following the transaction his reported direct holding is 1,602,382, which the filing states consists of 161,878 shares of common stock, 315,504 restricted stock units subject to time-based vesting and 1,125,000 units subject to performance- and time-based vesting. The same filing reports indirect holdings of 36,652 shares through the 2021 trust and 205,000 shares through The Hallal Family Irrevocable Trust-2012, of which his spouse is a trustee and for which he disclaims beneficial ownership.
• Keith Woods, Chief Operating Officer. On September 24, 2026 he transferred 48,034 shares to the RKW Irrevocable Trust for no consideration, with his spouse serving as trustee. Transaction code G, again at $0.00. Following the transaction his reported direct holding is 572,021, which the filing states consists of 0 shares of common stock, 122,021 time-vesting restricted stock units and 450,000 units subject to performance- and time-based vesting, with the 48,034 shares now reported indirectly through the trust.

Merlintrader has found no filing showing transaction code S, for an open-market sale, or code P, for a purchase. The data source’s insider-transactions figure reads −0.38 per cent over its own window, consistent with transfers out of direct holdings rather than with disposals into the market.

Two things are worth drawing out, carefully, because the temptation to over-read insider filings is strong and usually wrong.

The first is that estate-planning transfers to irrevocable family trusts ten and thirteen days after the first approval in a company’s history are exactly what one would expect, and they say nothing about conviction in either direction. A gift is not a sale; the economic exposure moves within a family rather than out of it, and the timing follows a liquidity event because that is when such planning becomes worth doing.

The second is a detail in the footnotes that is more interesting than the transactions, and it has to be read with the indirect holdings in view. After his gift the chief operating officer holds no shares of common stock directly, and 48,034 indirectly through the RKW Irrevocable Trust, against 572,021 unvested units, 450,000 of which depend on performance conditions as well as time. The chief executive reports 161,878 shares held directly plus 241,652 held indirectly — 36,652 through The Hallal Family Irrevocable Trust-2021 and 205,000 through The Hallal Family Irrevocable Trust-2012 — against 1,440,504 unvested units. So the shares that moved did not leave the family, and beyond them the great majority of each man’s reported economics is future vesting rather than stock held outright. That aligns them with the share price over a multi-year horizon and is worth knowing when reading anything they say about a launch.

EDGAR — Scholar Rock Forms 4

16 Management and execution: what the record actually shows

Scholar Rock is led by David L. Hallal, who is Chairman and Chief Executive Officer, with Keith Woods as Chief Operating Officer. The company had 289 employees according to third-party data as of September 30, 2026 and is headquartered at 301 Binney Street, Cambridge, Massachusetts.

On execution the available evidence splits cleanly into two parts, and conflating them is the most common mistake made about companies at this stage.

Regulatory execution: demonstrated

Between October 2024 and September 2026 this team took a positive Phase 3 result, filed it, received a Complete Response Letter caused by a third-party fill-finish facility that was subsequently issued a Warning Letter, held an in-person Type A meeting with the agency with the facility’s representatives present, qualified a second fill-finish site, resubmitted, and obtained an approval nineteen days before the action date with a label broader than the evidence population. That is a hard thing to do and they did it. A reader assessing this management team on the question “can they handle the FDA and a failing supplier” has a documented answer.

Commercial execution: unproven by definition

The company has never sold a product. The commercial organisation has been built during 2026 — visible in general and administrative expense rising 29 per cent over the half to $100.9 million, nearly level with research and development — but it has no track record, because there was nothing to sell before mid-September. Nothing in the regulatory recovery described above predicts whether a specialist neuromuscular sales organisation can obtain payer coverage for a monthly infusion added on top of an existing therapy. These are unrelated competencies.

The communication record, and the one blemish on it

The September 15 Form 8-K is the only disclosure failure in the period reviewed, and it is worth placing precisely. Management told investors on the approval call that the FDA had required no post-marketing commitments on bone or reproductive safety, which was wrong — the agency had outlined a post-marketing requirement for a pregnancy safety study and separately requested 15-day alert reports on fractures — and the company corrected it in a filing the next day. The speed and clarity of the correction is a point in favour of the control environment. The substance of the error — understating the regulator’s requirements on the label’s central safety concern, to investors, on the company’s set-piece approval call — is a point against the care with which commercial and safety messages are prepared.

One correction is not a pattern. It does change how management commentary should be weighted relative to documents on this file, and the practical consequence is stated in the section above: where this page makes a safety or commercial claim, it is sourced to a filing.

Incentives

The Forms 4 filed in September disclose that the chief executive holds 161,878 shares of common stock directly and 241,652 indirectly through two family trusts, against 1,440,504 unvested restricted and performance stock units, and that the chief operating officer holds no shares directly and 48,034 indirectly through a family trust, against 572,021 unvested units, of which 450,000 carry performance conditions. Whatever one concludes about the launch, the people running it are paid in a currency that only becomes valuable if the share price is higher in several years than it is now.

17 Competitive position: complementary rather than competing, which is the point

ISEMBYLD does not compete with the drugs that dominate SMA treatment. It requires them. The label covers only patients who are currently receiving a survival motor neuron 2 targeted treatment, which in practice means nusinersen or risdiplam, the two agents used in SAPPHIRE. Every patient who takes ISEMBYLD is, by the terms of the licence, also taking something else.

That is commercially unusual and worth thinking through in both directions.

In favour: there is no head-to-head competition to lose, no formulary displacement to negotiate, and no need to persuade a physician to stop an established therapy. The clinical argument is additive and mechanistically distinct — SMN-targeted treatments are designed to prevent motor neuron loss and do not directly address muscle, as the company puts it, and ISEMBYLD is the first approved treatment to target the muscle. The existing prescriber base is already identified, already treating these patients, and already managing a specialty therapy. And the installed base is substantial: approximately 35,000 SMA patients globally have been treated with an SMN-targeted therapy.

Against: an add-on has to be paid for in addition to an expensive therapy rather than instead of one. Payers assess incremental cost against incremental benefit, and the incremental benefit on offer is a 2.2-point improvement on the Hammersmith scale, generated at nominal significance in 53 non-ambulatory children who received the approved dose, in a label that extends to adults and carries no mobility, subtype or severity restriction. The bar for an add-on is not the same as the bar for a first therapy, and no list price has been disclosed against which to judge it.

The second competitive question is about the field rather than the drug. Myostatin inhibition has a long history of clinical disappointment — the company’s own chief executive referred in the approval release to “decades of failed industry-wide efforts to unlock the potential of myostatin inhibition”. Scholar Rock has now produced the first approval in that class, which is a genuine scientific achievement and also an invitation: a validated mechanism attracts capital and competitors, and the company’s platform claim rests on being first and selective rather than on being alone. FORGE in facioscapulohumeral muscular dystrophy and the EMBRAZE result in lean-mass preservation during GLP-1 weight loss are both attempts to extend the franchise, and the second of those is in the most contested commercial area in the industry, where Scholar Rock is seeking a partner rather than developing.

18 Retail sentiment

Two things have to be said about this block for it to be useful rather than decorative.

The first is that the tagged-message split returned by the platform at this reading was degenerate — it reports every tagged message on one side — which on a sample this small carries no information and is therefore not reproduced here. The score of 30 out of 100 with a bearish label is the platform’s headline reading and the only number worth quoting. A watcher count of 2,373 is modest, which is itself informative: this is not a retail-crowded ticker.

The second is that a bearish retail reading on a stock that has just been approved by the FDA, with 19.3 per cent of its float short, is consistent rather than contrarian. Retail discussion and the short base are pointing the same way, and the sell-side aggregate is pointing the other. That is what an unresolved commercial question looks like from the outside.

These are comments and scores produced by retail traders who are not professional analysts. They are not clinical evidence, they are not a regulatory fact and they do not establish anything about demand for ISEMBYLD. They are included because a reader who looks at this ticker anywhere else will see them, and because the direction of the reading is a fact about the market even when its content is not a fact about the company.

StockTwits · $SRRK

19 Catalyst map after the approval

The approval was the last binary event on this file for some time. What follows is a mixture of one dated milestone, several undated regulatory files and a recurring stream of quarterly evidence. Each row states what the source says and what it cannot say.

The structural point is that this file has shifted from event-driven to evidence-driven. For two years the question was whether the FDA would approve. From now until a European decision, the questions are answered in quarterly filings and in post-marketing reports, which arrive on a schedule nobody can trade around.

20 Red flags

These are the items a sceptical reader should hold against the rest of the page. Each is drawn from a filing or a label, not from commentary.

• Fractures at 9 per cent against 2 per cent, in the label, with a continuation decision put to the prescriber. The prescribing information warns of an increased risk of fractures including serious fractures, notes that some events had no clear precipitating event, and tells prescribers to use the drug with caution in patients with a history of low bone density or multiple fractures and to weigh the risks and benefits of continuing in patients who fracture. The rate is 5 of 53 patients at the approved dose against 1 of 53 on placebo, a difference of four events, and in the open-label extension at twice the dose 22 patients (9 per cent) fractured, including five serious femoral events. The FDA has requested that every fracture case worldwide, serious or not, be submitted as a 15-day alert report. This is the central risk on the product and it is not a theoretical one.
• Management misstated the FDA’s demands to investors. On the September 14 approval call management said the FDA had imposed no post-marketing commitments on bone or reproductive safety. The company filed a correction the next day, on September 15. The error landed on the most sensitive point of the label.
• The pivotal p-values are nominal and the evidence population is much narrower than the label. A 2.2-point HFMSE improvement at nominal p=0.0121, generated in 53 patients who received the approved 10 mg/kg dose against 50 on placebo, all of them aged 2 to 12, all non-ambulatory, all with later-onset Type 2 or Type 3 disease and an HFMSE score between 10 and 45 — supporting a licence that covers adults and children from two years upwards with none of those conditions attached. No patient above the age of twelve received the approved dose in SAPPHIRE, and the label states that clinical studies did not include patients 65 and older.
• No product revenue and no disclosed price. The trailing revenue field at the September 30, 2026 reading is zero against a trailing net result of −$408.59 million, and Merlintrader has found no public disclosure of a list price or wholesale acquisition cost for ISEMBYLD. The single most important commercial variable is not yet public.
• A fixed commercial cost base arriving before the revenue. General and administrative expense rose 29 per cent over the first half to $100.9 million, against research and development of $110.0 million, and the company says it expects that line to remain substantial. A commercial organisation is a cost from the quarter it is hired, whether or not sales follow on schedule.
• Dilution with a pattern. Shares outstanding rose from 108,461,354 to 121,600,396 in six months. A new $200 million ATM prospectus supplement was filed on August 6, 2026, and roughly $257.7 million has been raised under the programme to date.
• Debt capacity that is not free cash. The Blue Owl facility requires unrestricted cash of at least 60 per cent of outstanding principal unless milestones are met and maintained, priced at three-month Term SOFR with a 1.00 per cent floor plus 5.00 per cent. On the $200.0 million drawn at June 30, 2026 that is $120 million of the $492.1 million of liquidity. The remaining $150 million under the plafond is a delayed draw at the company’s option, open from the approval until September 30, 2027; the further $200 million incremental tranche is uncommitted and needs the lender’s consent.
• Manufacturing history. The first BLA was rejected in September 2025 because of a third-party fill-finish facility that had received a Form 483 in July, was classified Official Action Indicated, and received a Warning Letter in November. The product is now licensed with two fill-finish sites, which mitigates concentration but does not remove contract-manufacturing risk from a biologic.
• Patent proceedings on the platform. Three European patents are subject to opposition proceedings and two Japanese patents to invalidation trials, by number, in a company whose value now rests on one marketed product.
• A 19.3 per cent short position that survived the approval. 19.41 million shares at a ratio of 13.40 days. A large, informed pool of capital has reached the opposite conclusion from the sell-side aggregate of $69.12.
• Everything above the cash line is the launch. Book value is $2.09 per share and cash $4.04 per share, against a close of $48.32 and a price to book of 23.15. There is no revenue base to fall back on if uptake disappoints.

21 Merlintrader bottom line

Scholar Rock has done the hard thing. It took a mechanism with decades of industry failure behind it, ran a Phase 3 that worked, survived a Complete Response Letter caused by a supplier’s factory rather than by its own drug, qualified a second fill-finish site, resubmitted, and received an approval nineteen days before the action date with a label considerably broader than the population that generated its primary result. On top of that it received a Rare Pediatric Disease Priority Review Voucher, which is a saleable asset with no connection to how ISEMBYLD performs commercially. That is a company that executes on regulatory matters, and the record is documented rather than asserted.

What it has not done is sell anything. The last reported quarter closed ten and a half weeks before the approval, so there is no revenue line yet, no disclosed price, and no reimbursement evidence. The organisation built to do the selling is already in the accounts — general and administrative expense up 29 per cent over the half to $100.9 million, nearly level with research and development — which means the cost of the launch is a fact while the benefit is a forecast. Liquidity of $492.1 million against first-half operating cash use of $152.4 million buys time — the company itself expects it to last into the second half of 2027 — and a $200 million ATM supplement, a $150 million delayed draw the company may take at its own option until September 30, 2027 and a further $200 million incremental facility subject to the lender’s consent buy more, at a price in dilution or interest.

The label is the recurring theme here, because it is where the commercial case and the clinical case meet. A 2.2-point improvement at nominal significance in 53 non-ambulatory children who received the approved dose, extended by the licence to adults and to patients with none of the trial’s restrictions, sold as a monthly infusion on top of an existing therapy, with a warning that fractures occurred in 9 per cent against 2 per cent and an instruction to weigh the risks and benefits of continuing after one. Each of those clauses is a small friction, and launches are decided by the accumulation of small frictions. The FDA evidently thought the benefit justified the risk and the breadth, which is the judgement that matters most and has already been made. Whether payers and prescribers reach the same conclusion across the whole label is the open question, and it is not a question any amount of reading can answer in advance.

Two facts frame how unresolved this is. Nineteen point three per cent of the float is sold short at thirteen days of volume to cover, and the position did not close on the approval. The data source’s consensus target is $69.12 with an almost uniformly positive recommendation score. Those two groups are looking at the same label, the same cash balance and the same correction filed on September 15, and they disagree completely.

What a reader can do is watch the right documents rather than the right headlines. The first partial quarter of ISEMBYLD revenue establishes that the product ships and is paid for. The first full quarter establishes the shape. A European decision on the application resubmitted (announced October 8, 2026), with a CHMP opinion anticipated in the first half of 2027, would change the size of the opportunity. SRK-439’s Phase 1 topline is due in the second half of 2026 and tells you about a subcutaneous route, not about a disease. And the post-marketing fracture reports, which the FDA asked for by name and sees before anyone else, are the stream most likely to settle the argument in either direction.

This is editorial research for informational purposes. It is not advice, not a recommendation, and it takes no view on whether the shares are worth their price.