What a disputed trial reveals about acquisition due diligence, regulatory risk, and the limits of a securities settlement.
BioStockEdge | Editor’s Note
Evidence cutoff: October 28, 2026
Company: Amgen (NASDAQ: AMGN)
An acquired drug can have an established market, rising sales, and patients who believe it helps them, while the evidence supporting its approval becomes unreliable. Tavneos puts that uncomfortable possibility at the center of an investment question: how much risk has approval actually removed?
Amgen completed its roughly $4 billion acquisition of ChemoCentryx in October 2022, a year after the FDA approved Tavneos, or avacopan, for adjunctive treatment of severe active ANCA-associated vasculitis. Its appeal included the prospect of controlling disease with less glucocorticoid exposure. By 2025, Amgen’s reported Tavneos product sales reached $459 million, up from $283 million in 2024. Commercial uptake was real. [1,2]
The regulatory position has since deteriorated. In April 2026, the FDA’s Center for Drug Evaluation and Research proposed withdrawing approval, citing insufficient reliable evidence of effectiveness and untrue statements of material fact in the application. The pivotal paper was retracted in June. Amgen disputes the agency’s conclusions and has submitted additional analyses. As of this article’s cutoff, U.S. withdrawal proceedings remain unresolved. [3–5]
For investors, the distinction matters. The issue is broader than whether a new analysis produces a favorable number. It is whether the evidence can still support the product’s legal and commercial life.
A serious product risk, with a defined place in Amgen’s business
The $459 million Amgen booked in Tavneos product sales represented approximately 1.25% of its $36.751 billion in 2025 total revenue, calculated from its annual filing. That denominator includes other revenue, and Amgen’s booked product sales do not represent all sales recorded by overseas partners. This is a measure of scale, not a complete estimate of the asset’s economics. [2]
The distinction prevents two mistakes. Tavneos’ difficulties should not automatically become a thesis about the viability of all of Amgen. Equally, a modest share of group revenue does not make a compromised acquisition immaterial: future sales, the asset’s value, legal exposure, and confidence in the acquisition process can all matter.
The merger proxy also supplies a useful historical benchmark. ChemoCentryx’s disclosed AAV-only forecast put net revenue at $451 million in 2024 and $670 million in 2025. Actual Tavneos product sales were below those amounts. But the forecast and reported sales use different accounting scopes; dividing one by the other would not produce a defensible target-achievement percentage. Nor can revenue alone determine whether an acquisition has earned its purchase price. [6]
The more consequential change is the uncertainty around how long the product can remain available and on what evidentiary basis.
The endpoint was a clinical judgment, not just a laboratory measurement
ADVOCATE compared treatment strategies. Patients received background rituximab or cyclophosphamide followed by azathioprine. One group also received avacopan for 52 weeks; the control group received a protocol-directed prednisone taper, with study-supplied prednisone stopped from week 21. Additional glucocorticoids were permitted in specified circumstances. The rituximab subgroup did not receive subsequent rituximab maintenance treatment. [7,8]
Those details define the comparison. Lower steroid use was partly an assigned feature of the strategy; whether patients could maintain disease control while reducing that exposure was what needed testing.
The week-52 sustained-remission endpoint combined several judgments: remission at weeks 26 and 52, no intervening relapse, and restrictions on glucocorticoid use for vasculitis near the assessments. A clinician therefore had to interpret both disease manifestations and the reason for steroid treatment. [7]
A blinded central committee was intended to apply those rules consistently. Yet the reliability of adjudication depends on the whole workflow, including who selects cases for review.
According to CDER’s withdrawal proposal, an initial database lock and unblinded analysis in November 2019 failed to establish week-52 superiority. Unblinded sponsor personnel subsequently selected nine patients for readjudication. Five avacopan patients changed from not achieving sustained remission to achieving it; the two-sided superiority P value moved from 0.1025 to 0.0132. The original analysis and relevant process were not adequately disclosed to the FDA, CDER concluded. [4]
The adjudicator himself remained blinded. This is important both analytically and when discussing individual responsibility: the public record does not justify treating every academic participant as a knowing collaborator.
But a blinded adjudicator cannot neutralize selection made upstream by people who know treatment assignments and outcomes. The expert’s final decision may be independent while the opportunity to reconsider particular cases is not. Clinical measurements need not be fabricated for endpoint classification to change the reported result.
Amgen argues that the underlying clinical data and BVAS scores were unchanged, that the readjudications corrected application of prespecified rules, and that the process was documented. CDER rejects that characterization and regards it as data manipulation accompanied by material misstatements. A fair account must preserve that disagreement without reducing CDER’s findings to an ordinary difference of clinical opinion. [4,9]
Reanalysis can retain a signal without restoring the original evidence
Amgen commissioned the Duke Clinical Research Institute to conduct an independent blinded readjudication of the original trial. This was a review of the same participants, not a new randomized experiment. The publicly available results cited here are a preprint. [10]
The new week-52 sustained-remission rates were 61.4% with avacopan and 52.4% with the prednisone strategy. The adjusted difference was 9.8 percentage points, with a 95% confidence interval from −0.3 to 19.9 percentage points. The result met the original noninferiority margin of −20 percentage points but did not establish superiority. [10]
That finding is compatible with a clinically relevant incremental benefit. It is also compatible with no incremental benefit. Failure to establish superiority does not prove that the drug has no activity.
Conversely, noninferiority is conditional on the chosen margin, the comparator’s performance, and the study’s reliability. It cannot, by itself, settle whether continued approval is warranted.
CDER says it had required week-52 superiority in its earlier development discussions. Amgen argues that the eventual adjunctive indication and the FDA’s totality-of-evidence approval rationale require a broader assessment than the earlier proposal to replace glucocorticoids. This is a dispute about the regulatory meaning of the evidence as well as the readjudication process. [4,9]
The useful investor question is therefore not simply, “Does any analysis still favor avacopan?” It is, “What evidence would resolve the regulator’s objections, and has that evidence actually been produced?”
Real-world data deserve the same discipline. An Amgen-funded preprint pooled 71 observational studies, including 53 conference abstracts. Its striking 93% remission estimate at 12 months came from 12 studies comprising 362 patients. It was not an estimate of incremental benefit over a randomized control group. [11]
Patients also received effective background therapy, and treatment selection and follow-up can influence observed outcomes. Moreover, pooled 12-month remission is not interchangeable with ADVOCATE’s specifically defined week-52 sustained remission.
A comparative study using linked U.S. claims and electronic health records reported a weighted relapse hazard ratio of 0.81, with a 95% confidence interval of 0.59–1.13. The direction favored avacopan, but no difference remained within the interval; disease activity was not directly measured with BVAS, leaving potential residual confounding. That study is also cited as a preprint. [12]
These are potentially informative findings. They do not automatically repair the pivotal trial’s integrity problems. Robust independent evidence can matter to a totality-of-evidence assessment; an accumulation of favorable observations is not automatically equivalent to such evidence.
Safety changes the cost of uncertainty
The FDA identified 76 postmarketing liver-injury cases with reasonable evidence of a causal association with avacopan in data through October 9, 2024. Eight reported death, and seven involved biopsy-confirmed vanishing bile duct syndrome. Most cases came from Japan. These are selected safety reports, not a denominator-based estimate of the incidence among all treated patients. [13]
A patient can achieve disease remission and still suffer drug-related harm. Benefit signals and serious safety signals need not contradict one another.
The investment implication follows from their interaction: when confidence in benefit weakens, an established risk becomes harder to justify. A safety profile does not have to worsen further for the benefit-risk conclusion to change.
Neither the presence of fatalities nor the trial-process findings establish, without further evidence, that manipulation legally caused any individual death. That question belongs to a different inquiry from whether the drug meets the conditions for continued approval.
A settlement does not close the regulatory or patient questions
The securities litigation began in May 2021, before Amgen’s acquisition. Its existence was public, but public allegations are different from knowledge of the internal readjudication history. Amgen says it learned of that history years later through the litigation. Public information does not establish the full scope of its 2022 diligence or what particular decision-makers knew. The appropriate inference is a diligence question, not an unsupported allegation of knowing acquisition of manipulated data. [9,14]
The district court granted summary judgment for defendants in August 2025; plaintiffs appealed. A proposed settlement was subsequently increased from $35 million to $69 million following CDER’s withdrawal proposal. Its approval hearing is scheduled for October 29, 2026. That is a securities settlement hearing, not an FDA hearing. [14]
The settlement addresses specified stock-purchase claims and contains no admission of wrongdoing. It is not a patient compensation fund. If approved, it may end those securities claims without producing a comprehensive merits judgment on every aspect of the clinical-trial conduct. It cannot determine whether Tavneos may remain marketed. [15]
Patients’ injuries and any further government enforcement have to be considered separately. Patient organizations have called for accountability, and law firms have advertised investigations of liver-injury claims. Neither establishes that a patient case has been filed, adjudicated, or compensated. Investors should not convert the existence of those advertisements into a quantified liability estimate. [18]
The FDA’s withdrawal process does not require a prior criminal conviction. A hearing request also does not guarantee a hearing: the relevant rules require a genuine and substantial factual issue supported by specifically identified reliable evidence. No final U.S. withdrawal order or formal FDA hearing date was identified in the public materials reviewed through October 28. [3,16]
Other jurisdictions have already acted. The European Commission revoked the authorization on August 4, 2026. The U.K. stopped initiation in new patients from September 1 and announced a managed transition for existing patients, with revocation intended for March 1, 2027. Those measures do not establish the outcome or timing of the U.S. process. [17]
Three developments would materially change the investment assessment: an FDA decision clarifying the product’s continuing legal status; evidence that addresses efficacy and safety more convincingly than the disputed record; and actual court filings or enforcement decisions that clarify liability. They should be tracked separately because one does not settle the others.
Tavneos illustrates a limit to a familiar investment shortcut. Approval can reduce development uncertainty, but it cannot make the evidence on which approval rests permanently immune to scrutiny. Revenue demonstrates adoption. An acquisition establishes what a buyer was willing to pay. A securities settlement resolves defined claims. None, on its own, establishes a reliable incremental clinical benefit.
For capital allocation, the most durable question is how much confidence an asset’s evidence can carry—and what happens to its expected value when that confidence has to be rebuilt.
Sources
Amgen acquisition completion, October 20, 2022.
https://www.amgen.com/newsroom/press-releases/2022/10/amgen-successfully-completes-acquisition-of-chemocentryxAmgen 2025 Form 10-K. Tavneos booked product sales: $423m U.S. plus $36m elsewhere; group total revenue: $36,751m. The approximately 1.25% comparison is the author’s calculation.
https://www.sec.gov/Archives/edgar/data/318154/000031815426000010/amgn-20251231.htmFDA, CDER proposes to withdraw approval of Tavneos.
https://www.fda.gov/drugs/drug-alerts-and-statements/cder-proposes-withdraw-approval-tavneosCDER withdrawal proposal and notice of opportunity for a hearing, April 27, 2026.
https://www.fda.gov/media/192160/downloadNEJM retraction, June 29, 2026.
https://www.nejm.org/doi/10.1056/NEJMe2608684ChemoCentryx 2022 merger proxy, AAV-only financial projections. These are disclosed seller forecasts, not Amgen sales commitments.
https://www.sec.gov/Archives/edgar/data/1340652/000119312522244850/d359803ddefm14a.htmADVOCATE protocol, amendment 4.0.
https://cdn.clinicaltrials.gov/large-docs/27/NCT02994927/Prot_000.pdfFDA multidisciplinary review, 2021.
https://www.accessdata.fda.gov/drugsatfda_docs/nda/2021/214487Orig1s000MultidisciplineR.pdfAmgen’s July 23, 2026 hearing submission and accompanying statement. Company arguments are attributed as such.
https://www.amgen.com/tavneos-prescribers
https://www.amgen.com/-/media/Themes/CorporateAffairs/amgen-com/amgen-com/downloads/newsroom/AmgenResponse,-d-,7,-d-,23,-d-,2026_REDACTED.pdfJayne et al. Primary-endpoint readjudication. Preprint, September 16, 2026.
https://www.medrxiv.org/content/10.64898/2026.09.16.26363171v1.fullIbiloye et al. Real-world systematic review and meta-analysis. Amgen-funded preprint, June 17, 2026.
https://www.medrxiv.org/content/10.64898/2026.06.17.26355805v1.fullDang et al. Comparative effectiveness and safety study. Preprint, July 22, 2026.
https://www.medrxiv.org/content/10.64898/2026.07.20.26358270v1.fullFDA liver-injury safety communication, March 31, 2026; reports through October 9, 2024.
https://www.fda.gov/drugs/drug-safety-communications/fda-identifies-cases-serious-liver-injury-patients-taking-tavneos-avacopan-severe-active-antiCourt-authorized securities litigation FAQ and hearing information.
https://www.chemocentryxsecuritieslitigation.com/faqRevised settlement agreement, May 15, 2026.
https://static.blbglaw.com/docs/May%2015,%202026%20-%20Revised%20Stipulation%20of%20Settlement_ChemoCentryx.pdfWithdrawal and review statutes; hearing requirements.
https://www.law.cornell.edu/uscode/text/21/355
https://www.law.cornell.edu/cfr/text/21/12.24EU and U.K. regulatory status.
https://www.ema.europa.eu/en/medicines/human/EPAR/tavneos
https://www.gov.uk/government/news/mhra-concludes-review-of-avacopan-vifor-following-reassessment-of-benefit-risk-balancePublic accountability statements and patient-claim investigations; law-firm pages establish solicitation/investigation, not filed or successful litigation.
https://www.vasculitisint.com/patients-should-not-pay-for-failure-in-science/
https://www.vasculitisint.com/ec-revokes-market-authorisation-tavneos/
https://lamblawoffice.com/tavneos-vasculitis-drug-safety-alert-fda/
https://www.corywatson.com/tavneos-lawsuit/
This article provides evidence-based commentary, not personalized investment or medical advice. It makes no recommendation to buy, sell, or hold a security. Regulatory findings, company positions, and judicial decisions are distinguished throughout. Procedural status is current only to the stated cutoff.
© 2026 BioStockEdge. All rights reserved. Reproduction or republication requires permission.