Amgen told the FDA no — a blinded Duke re-analysis of ADVOCATE met noninferiority at 68.1% versus 67.1% but lost superiority, and the deadline is July 29. We unpack the avacopan withdrawal fight, the ChemoCentryx acquisition that bought it, and the EMA committee already moving against it. Plus Elevar's complete response letter for rivoceranib and camrelizumab in first-line liver cancer over a manufacturing inspection, Summit's ivonescimab survival update ahead of its November decision, Sun Pharma's $11.75 billion Organon approval, Scribe Therapeutics' gene editing IPO, Ipsen's Bylvay failure in biliary atresia, and the FDA peptide panel that voted yes on thin evidence.
Auto-generated from the episode script. Deal names link to their scorecard in the database.
The biggest story in pharma this week wasn't data and it wasn't a deal. It was a company telling the FDA no — in writing, with four days left on the clock.
Amgen wants a hearing on Tavneos. Elevar lost a first-line liver cancer approval to a plant inspection. And a gene editing company with nothing past Phase 1 got an IPO out the door and up 44 percent. Let's take the week apart.
Welcome to The Pharma Closeout Week in Review for Sunday, July 26th. I'm Alex Mercer.
And I'm Maya Patel. A lot happened this week that deserves a second look. Let's get into it.
Story of the week, and it's a fight. Wednesday, July 29th is Amgen's deadline to answer the FDA's formal request that it pull Tavneos off the US market. Amgen's answer went in Thursday. It's no. The company submitted data and analyses backing a request for a hearing, and the statement says flatly that Amgen strongly disagrees with the agency's proposal. So we have an approved rare disease drug whose pivotal trial no longer exists in the literature — and a sponsor refusing to walk.
And the sequencing is what most coverage got backwards. The retraction wasn't the trigger. Per MedPage Today, the agency has been calling for this drug to come off the market since January. The New England Journal pulled ADVOCATE at the end of June — downstream of the data-integrity findings, not the cause of them.
Set the table on the asset. Tavneos — avacopan — oral C5a receptor antagonist, approved in 2021 as adjunctive treatment for adults with severe active ANCA-associated vasculitis. GPA and MPA. And Amgen didn't build it. It bought ChemoCentryx for roughly $3.7 billion in 2022, after the approval was already in hand. One product. One pivotal trial: ADVOCATE, 331 patients, avacopan against a prednisone tapering strategy, everyone on cyclophosphamide or rituximab background.
One pivotal trial. And per the FDA's investigation, sponsor personnel went back in after database lock and unblinding, selected participants for re-adjudication, and moved five avacopan patients from "not in sustained remission" to "sustained remission." Five patients. That reclassification is what produced the statistically significant superiority result the entire product narrative was built on.
Five, out of 331. Which brings us to Thursday, and this is the turn. Amgen had the Duke Clinical Research Institute run a fully blinded re-adjudication of the primary endpoint. Twenty-six-week remission: 68.1 percent on avacopan, 67.1 percent on the prednisone taper. Noninferiority demonstrated. Fifty-two-week sustained remission: 61.4 versus 52.4. Noninferiority met — superiority not met, unlike the original analysis. So an independent read says the drug works. It just doesn't work better.
But here's my issue with "it still works." Noninferiority against a prednisone taper is not the claim that built this product. The nine-point spread at 52 weeks was the steroid-sparing story, and that's what moved avacopan into treatment algorithms. Take superiority out and you have a drug that matches a steroid regimen on remission. And look at what Duke actually re-scored — the primary endpoint. Nothing else. The hepatic safety questions sit entirely outside that review, which is exactly why a key EMA committee has already recommended pulling the European authorization on the combined picture.
That's a fair correction. I'll take it. Where I hold my ground is the statutory test. The question in front of the agency isn't whether the original superiority claim was right — it's whether substantial evidence of effectiveness survives. A blinded third party re-scoring the same trial and landing on noninferiority is a real argument. That's not a press release.
Then the honest read is that this was never an efficacy fight. It's a process fight wearing an efficacy costume. Nobody is disputing 68 versus 67. The agency is asking whether an evidence base it has already characterized as manipulated can be rehabilitated by a re-analysis the sponsor commissioned — however clean the blinding was. That question has nothing to do with vasculitis.
Say more on the hearing route. Why does that matter more than the filing?
Because Amgen isn't asking for informal reconsideration. It's invoking the hearing pathway, which converts an agency decision into an adjudicated record — a docket, submissions, testimony, findings that everyone else gets to cite afterward. That's the piece with legs beyond this one drug.
And Amgen is arguing the wide frame on purpose. Its statement opens on more than 30 million Americans living with a rare disease and an average diagnostic odyssey of five years or more, then notes Tavneos is one of very few options in severe active AAV. That's the benefit-risk case they want the hearing to be about. There's a dealmaking dimension underneath all of this too. Let's bring in Marcus Webb on this one.
Amgen paid roughly $3.7 billion for ChemoCentryx in 2022, and the asset was the deal — one approved rare disease product resting on one pivotal trial. The terms tell you Amgen underwrote regulatory risk and commercial risk. They did not underwrite the possibility that trial conduct itself was compromised. I cannot point to a precedent where a sponsor's own blinded re-adjudication is the central exhibit in a withdrawal proceeding. That absence is the story. And it becomes a priced line item in diligence on every single-pivotal asset that trades from here.
The competitive read: if you're running business development or portfolio strategy and you're buying a single-asset, single-pivotal rare disease company, trial-conduct diligence just stopped being a checkbox. And if you hold a competing program in AAV, Wednesday goes on your calendar — because a withdrawal reopens a market with almost nothing else in it.
The unresolved question isn't whether Amgen gets its hearing. It's what the agency decides "substantial evidence" means when the arithmetic still clears and the conduct doesn't. Every approval on the market resting on a single trial is downstream of that answer.
Pull back and look at the whole week, and the through-line isn't efficacy at all. Story after story turned on the plumbing of the evidence, not on whether the molecule worked.
Elevar is the cleanest example. On July 10th the FDA issued a complete response letter for rivoceranib plus camrelizumab in first-line unresectable HCC — and the cause was deficiencies identified during a cGMP inspection of a manufacturing site listed on the NDA. The Phase 3, CARES-310, put median overall survival at 23.8 months, which the company describes as the longest reported to date in first-line HCC. Published in Lancet Oncology in December. The regimen is already in the 2025 BCLC strategy and in ESMO guidelines.
Guideline-endorsed survival data, stopped by a plant. Nothing about the molecule changed between Thursday and Friday.
And that's the pattern. It recurs all year for Asia-origin assets entering the US — the constraint is the inspection, not the endpoint.
Which takes us to the peptide panel. We said Friday we'd follow whatever moved next there, and it closed out with six of the seven peptides narrowly recommended for broader compounding use — despite what BioPharma Dive characterized as a lack of substantive evidence behind them. And per Fierce Pharma, emideltide, the first peptide to go up against an actual committee vote, failed to make grade.
Which is a strange inversion. A panel voting yes on thin evidence, and no on the one asset that showed up with a dossier. Whatever else that signals, the bar this quarter is being set by the venue, not the data.
Third thread — Summit. We've run the ivonescimab numbers already, so I won't re-litigate them. The point for the week is directional: the survival curve keeps moving the right way, hazard ratio of 0.76 in the Western subgroup of 165 patients, and it still isn't statistically significant. November's decision rests on regulatory precedent rather than a clean endpoint.
And precedent exists — the agency has cleared agents in previously-treated EGFR-mutant lung cancer without significant overall survival before. What Summit is really doing is asking the FDA to be consistent with itself.
On the deal side, the weekend produced no new M&A at all — which is its own data point in a year this active. The one milestone was Organon. Shareholders approved Sun Pharma's acquisition on July 23rd, $11.75 billion enterprise value, Organon becoming a wholly owned subsidiary of Sun Pharmaceutical Holdings USA. We've been tracking that since April, and the remaining questions are closing conditions and antitrust timing, not price. Elsewhere in the roundup: Ipsen's Bylvay failed its biliary atresia trial on Thursday. And Fierce's oral GLP-1 tracker had obesity prescriptions jumping again.
So if you're allocating diligence resources for the back half, the message from this week is that the binding constraint has moved. Molecules are clearing. Manufacturing records, adjudication logs, statistical analysis plans — those are what's failing. And those are the functions nobody staffs properly until one of them costs a launch.
Under the radar this week: Scribe Therapeutics priced an upsized IPO on July 23rd. 8,580,000 shares at $15, gross proceeds of about $128.7 million — and the stock jumped 44 percent on day one. In vivo CRISPR, aimed at cardiometabolic disease.
Two things make that land. Per BioPharma Dive it's the smallest haul among the 14 biotechs that have gone public this year — and it's one of only two that priced without a drug already in Phase 2 or later. So the window isn't just open. It's open again for platform stories that haven't reached Phase 2, and this one priced up rather than down.
I'd read the 44 percent carefully, though. A first-day move like that on a $129 million raise is a small float finding a bid — not a verdict on the biology. And the test here is the one I always come back to: the second-asset test. One in vivo editing program is a program. The platform claim doesn't exist until asset number two reads out.
Either way, if you're a private gene editing company with a 2027 financing plan, your options got repriced this week. The comparable is now a Phase 1 story that cleared.
Looking ahead, the week is front-loaded. Wednesday, July 29th — Amgen's deadline on Tavneos. Thursday, Viatris has an FDA action date. And MannKind's FUROSCIX ReadyFlow autoinjector, subcutaneous furosemide for edema in adults with chronic heart failure and chronic kidney disease, is on the calendar for today.
Wednesday matters twice over, because the Q2 earnings wave lands the same week — Pfizer, Merck, AstraZeneca, Bristol Myers Squibb, AbbVie, GSK, and Amgen itself. Which means management gets asked live what the plan is if the hearing request is denied. And notice something: nobody has put a sourced revenue number on this drug in public all week. That tells you it's small relative to Amgen — and completely beside the point on precedent.
Further out, Novartis flagged a heavy second half on its July 21st call. Pivotal readouts due for pelacarsen, remibrutinib and del-desiran, plus additional ianalumab data and Rhapsido in hidradenitis suppurativa. That's the calendar to build your autumn around.
And one more thing on Wednesday. If Amgen gets its hearing, the first useful signal isn't the outcome — it's the scope. Whether the agency confines the docket to the primary endpoint or opens the hepatic safety record will tell you how far this precedent travels.
That wraps our Week in Review — a week where the evidence base, not the science, decided almost everything. Amgen digging in on Tavneos. Elevar's survival data stopped by an inspection. A peptide panel voting past its own evidence. Scribe getting a gene editing IPO out the door on a Phase 1 story. Wednesday is the one to circle. Follow wherever you listen and you'll start every weekday with this.
I'll be watching whether that hearing request gets acknowledged before the deadline even lands — the timing alone tells you something about how the agency sees this. Have a great week, everyone. Let's go!
The daily Closeout in under 15 minutes — trial readouts, FDA decisions, and the deal math behind them. Free.
Subscribe →