Amgen’s Tavneos: The Evidence, the Acquisition, and the Unfinished Patient Question

Original evidence review: October 29, 2026. Procedural update: October 31, 2026.

October 31 update: Settlement approved; patient and regulatory questions remain

On October 29, the U.S. District Court for the Northern District of California granted final approval of the $69 million securities settlement in Homyk v. ChemoCentryx, No. 4:21-cv-03343-JST, and entered judgment. It also approved the allocation plan and attorneys’ fees and litigation expenses. The judgment contains a no-admissions provision; settlement approval is not a judicial finding of clinical-trial fraud. [27,28]

The fee order confirms that the $69 million has been funded into escrow. That is different from eligible investors receiving their payments. The court-authorized administrator still says distribution will follow completion of claims processing. No specific payment date or completed distribution was verified as of this update, and the status of any subsequent appeal remains to be checked. [18,28]

This was the securities settlement hearing, not an FDA withdrawal hearing. This review did not verify an FDA decision granting a hearing, a formal hearing date, or a final U.S. withdrawal order. FDA’s public notice continues to describe interim market availability and recommends that clinicians discuss Tavneos and other available treatments with their patients. [4]

For investors, final approval narrows uncertainty within this securities case. It does not establish the drug’s incremental benefit, resolve its U.S. regulatory status, or compensate patients for injury. No newly verifiable patient-injury case number or compensation judgment, DOJ or SEC enforcement announcement concerning these data issues, or new study sufficient to change the benefit-risk assessment was identified in this update. These are limits of the public-source review, not proof that no individual proceeding, private resolution, or nonpublic investigation exists.

The financial remedy has advanced; the clinical and patient-accountability questions still require separate evidence and decisions. The original analysis below has been updated where its settlement status was superseded.

---

Patients can improve while taking a medicine whose pivotal trial no longer provides reliable evidence of benefit. A company can report growing sales while regulators question whether the product should remain approved. A shareholder lawsuit can reach an approved settlement while the consequences for patients remain unresolved.

Tavneos, the brand name for avacopan, brings these possibilities together. For investors, the central issue is how much confidence the evidence can support, and how that confidence translates into a product’s commercial life, acquisition value, and continuing responsibilities.

Amgen completed its acquisition of ChemoCentryx in October 2022, reporting approximately $3.7 billion in aggregate merger consideration. Tavneos had received FDA approval a year earlier. By 2025, Amgen’s booked Tavneos product sales reached $459 million, up from $283 million in 2024. In the second quarter of 2026, sales increased another 36% year over year to $150 million. Commercial adoption continued as the regulatory dispute developed. [1–3]

On April 27, 2026, the FDA’s Center for Drug Evaluation and Research proposed withdrawing approval, citing inadequate evidence of effectiveness and material untrue statements in the application. The pivotal publication was retracted in June. Amgen disputes CDER’s conclusions and is defending the product with additional analyses. The U.S. public materials reviewed for this note did not establish a final withdrawal order. [4–6]

An investment assessment has to follow that story through to its end: what the trial actually tested, how its conclusions changed, what the newer evidence can establish, and which obligations remain when a financial case is settled.

The clinical promise deserves a fair hearing

ANCA-associated vasculitis can damage the kidneys, lungs, and other organs. Avacopan blocks the C5a receptor, targeting a pathway involved in inflammatory injury. Its appeal also included the possibility of controlling disease with less glucocorticoid exposure. For patients living with the burdens of steroid treatment, that was a consequential proposition. [1]

The proposition required evidence about an entire treatment strategy.

In ADVOCATE, both groups received background immunosuppression with rituximab or cyclophosphamide followed by azathioprine. One group received avacopan for 52 weeks. The other 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 therapy. [7,8]

Lower steroid exposure was therefore partly built into the assigned strategy. It could become a benefit if disease control was maintained and overall treatment harm was reduced. The fact that the protocol used less prednisone could not, by itself, establish that avacopan made that reduction clinically successful.

The control group also continued background treatment. Describing it as having been left untreated after the prednisone taper would misrepresent the comparison. Conversely, its maintenance regimen places limits on how confidently the results can be extended to every treatment combination available today.

This matters commercially because the value proposition depends on a comparison. A medicine’s contribution is the additional benefit, or reduction in harm, achieved relative to a credible alternative. Improvement among people receiving several active treatments does not isolate the contribution of one component.

How a judgment about nine patients changed the trial

ADVOCATE’s sustained-remission endpoint required remission at weeks 26 and 52, no intervening relapse, and restrictions on glucocorticoid treatment for vasculitis near the assessments. Disease activity was measured with the Birmingham Vasculitis Activity Score, or BVAS. Determining whether an abnormality reflected active vasculitis, previous damage, or another cause required clinical interpretation. So did determining why a patient received steroids. [7,8]

A central adjudication committee was intended to apply the rules consistently while remaining unaware of treatment assignments. Yet the integrity of that system depended on how cases reached the committee, as well as on the committee’s own conduct.

According to CDER, the first database lock and unblinded analysis in November 2019 did not demonstrate week-52 superiority. Unblinded sponsor personnel then selected nine patients for readjudication. Five avacopan patients changed from not achieving sustained remission to achieving it. The two-sided P value moved from 0.1025 to 0.0132. CDER concluded that the original analysis and relevant process had not been properly disclosed. [5]

The adjudicator remained blinded. The journal’s retraction also says the two academic authors had been unaware of the post-lock, post-unblinding readjudication. Their lack of knowledge matters when assigning individual responsibility. [5,6]

It does not remove the problem of selective review. People who know the assignments and results can influence an analysis by deciding which cases receive another assessment, even when the person performing that assessment acts independently. The selection can introduce bias before the final clinical judgment is made.

Amgen argues that the underlying clinical data and BVAS scores were unchanged and that the readjudications corrected errors in applying prespecified rules. CDER regards the process as data manipulation accompanied by material misstatements. These are opposing positions in an administrative proceeding; a regulatory finding should not be silently converted into a criminal judgment. Equally, unchanged laboratory records would not establish that the reported endpoint was generated reliably. [5,9]

For investors, the lesson is concrete. The integrity of a trial includes the path from medical records to the analysis that determines whether the study succeeded.

The warning visible in 2021 was different from the later discovery

The FDA’s 2021 advisory committee record already showed that the result was sensitive to assessment methods. Using investigators’ BVAS assessments produced a smaller treatment difference that did not establish week-52 superiority. The agency also questioned the handling of persistent disease activity in the modified adjudication form. [8]

That was a reason to examine robustness. It did not establish that the agency then knew the subsequently disclosed sequence of database locks, unblinding, case selection, and readjudication.

Central review was the prespecified primary method, and disagreement with site investigators can arise for legitimate reasons. Persistent activity can also be difficult to distinguish from residual damage. A sensitivity analysis that fails to reproduce a result is a question to investigate, rather than proof of misconduct.

This distinction is especially relevant to acquisition diligence. Publicly visible concerns about robustness could have prompted closer examination in 2022. They do not establish what an acquirer knew about undisclosed internal events.

A serious diligence review would ask for the complete sequence of database locks and analyses, the reasons for endpoint changes, and the identities and blinding status of those able to initiate them. That is an editorial inference about the information needed to evaluate the asset. The public record does not establish the full scope of Amgen’s actual review or support an allegation that it knowingly acquired manipulated evidence.

A favorable signal survives, with important limits

Amgen commissioned the Duke Clinical Research Institute to conduct an independent, blinded readjudication of the original trial. It revisited the same patient records; it did not enroll a new randomized population. The results cited here are from a September 2026 preprint. [10]

Week-52 sustained remission was 61.4% with avacopan and 52.4% with the prednisone strategy. The adjusted difference was 9.8 percentage points, with a 95% confidence interval of −0.3 to 19.9 percentage points. The analysis met the original −20-percentage-point noninferiority margin but did not establish superiority. [10]

The interval allows a meaningful additional benefit and also includes no additional benefit. Failure to establish superiority does not demonstrate that the drug has no effect. Noninferiority, meanwhile, depends on the acceptability of the margin, the comparator, and the reliability of the study.

The parties also disagree about the approval standard. CDER emphasizes its earlier requirement for week-52 superiority. Amgen argues that the eventual adjunctive indication and the totality of evidence warrant a broader assessment than the earlier steroid-replacement development proposal. A favorable estimate alone does not resolve that dispute. [5,9]

Real-world evidence adds information of a different kind. An Amgen-funded preprint synthesized 71 observational studies, including 53 conference abstracts. Its approximately 93% remission estimate at 12 months came from 12 studies involving 362 patients. These were patients receiving avacopan-containing regimens, usually alongside other effective therapies. [11]

That finding can help describe clinical experience. It cannot establish the drug’s incremental contribution without a reliable comparison. It also cannot be directly compared with the Duke estimate: pooled 12-month remission and ADVOCATE’s specifically defined sustained remission are different outcomes.

A comparative U.S. claims and electronic-record study reported a weighted relapse hazard ratio of 0.81, with a 95% confidence interval of 0.59–1.13. Its direction favored avacopan, but the interval included no difference. Direct BVAS measurements were unavailable, and residual confounding remained a limitation. That analysis is also cited as a preprint. [12]

The evidence can contain a useful signal while leaving uncertainty about its size and regulatory adequacy. Funding should be disclosed and methods examined; company sponsorship alone neither validates nor invalidates the result.

For an investment thesis, the question is what specific evidence would answer the regulator’s objections. A larger collection of favorable observations does not automatically supply that answer.

Serious liver injury changes the stakes of uncertainty

The FDA’s March 2026 safety communication examined reports through October 9, 2024. It identified 76 liver-injury cases with reasonable evidence of a causal association with avacopan, including eight reported deaths and seven biopsy-confirmed cases of vanishing bile duct syndrome. Most reports came from Japan. The syndrome can produce permanent liver damage. [13]

Those figures are time bounded. Later Japanese safety materials reported 20 fatal hepatic cases through April 27, 2026. Amgen says the Japanese reports include cases in which a causal relationship could not be established. Different reporting periods and assessments mean the numbers should not be added together or converted into a treatment mortality rate. [14,15]

A patient can experience remission and serious drug-related injury at the same time. Positive clinical experience therefore does not cancel a safety signal. The existence of a safety signal also does not establish that every exposed patient was harmed.

The implications for value follow from the relationship between benefit and risk. If confidence in benefit falls, the justification for accepting a known risk weakens, even without a further deterioration in safety.

It would also be premature to infer that the trial conduct legally caused any particular injury or death. Product causation, individual harm, and responsibility for the trial require their own evidence. Recognizing those distinctions permits scrutiny without prejudging individual cases.

The acquisition needs an assessment beyond the sales curve

Tavneos’ $459 million in 2025 booked product sales amounted to about 1.25% of Amgen’s $36.751 billion in total revenue, an author calculation. This provides scale; it does not measure all worldwide partner sales or all of the asset’s economics. [2]

The product is a limited part of a diversified company. Its problems do not, on their own, determine the outlook for Amgen as a whole. But group revenue share cannot settle whether the acquisition deployed capital well.

The transaction also included assets beyond Tavneos. Its consideration should not be treated as a purchase price for one year of Tavneos revenue, and revenue should not be treated as profit or acquisition payback. [1]

The more useful question is how the dispute changes the future cash flows and costs that an investor is prepared to underwrite. A model may need to reflect uncertainty about continued authorization, treatment restrictions, additional research, commercial duration, and legal exposure. These are analytical considerations, not quantified losses established by the available record.

Amgen’s second-quarter 2026 filing itself acknowledges uncertainty about whether the FDA will find its additional evidence sufficient. It describes potential consequences ranging from further warnings, monitoring, and use restrictions to withdrawal, with possible effects on sales and legal proceedings. This is disclosed risk, not proof that every listed consequence will occur. [16]

Sales growth during this period demonstrates continued uptake. It does not establish the regulator’s eventual conclusion. A simple extrapolation from the latest quarter would therefore embed an assumption about continued availability that needs to be made explicit.

The securities lawsuit predated the acquisition. Amgen says it became aware of the disputed readjudication history later through that litigation. Awareness of public allegations and knowledge of the underlying internal process are different propositions. Questions about diligence remain legitimate without assuming knowledge that has not been demonstrated. [9,17]

Economic exposure also differs from the direct legal responsibility of a particular parent, subsidiary, officer, or former employee. Ownership alone cannot determine every allocation of liability. Investors need the relevant transaction documents, claims, and rulings before assigning it.

The shareholder settlement has a defined job

The securities litigation began in May 2021. Defendants obtained summary judgment in August 2025; plaintiffs appealed. A subsequent proposed settlement increased from $35 million to $69 million after CDER announced its withdrawal proposal. [17]

On October 29, 2026, the court held the settlement hearing, granted final approval and entered judgment. It also approved the allocation plan and fee-and-expense award. The settlement fund has been placed in escrow, but completed claimant distribution and a specific payment date were not verified as of October 31. The administrator says payments follow completion of claims processing. [18,27,28]

The agreement addresses defined claims connected with purchases of ChemoCentryx common stock and includes no admission of wrongdoing. It is not a patient injury compensation fund. Its approval does not determine whether Tavneos meets the requirements for continued FDA approval. Nor does it produce a comprehensive merits judgment on every disputed aspect of the trial. [19,27]

The FDA process has a separate purpose. The withdrawal statute permits action on grounds including insufficient substantial evidence of effectiveness and material untrue statements, following notice and an opportunity for a hearing. A prior criminal conviction is not a prerequisite. A hearing request must also meet specified evidentiary conditions; requesting one does not guarantee that one will be held. [20]

These distinctions matter when interpreting a favorable legal headline. Resolution of specified securities claims may reduce uncertainty within that case. It does not supply reliable clinical evidence, restore an authorization elsewhere, or calculate compensation for patients.

The patient question belongs inside the investment analysis

The immediate patient problem concerns treatment continuity. Starting a new patient and changing therapy for someone already stable involve different decisions. A managed transition can address the second problem while limiting the first.

The European Commission revoked Tavneos’ authorization on August 4, 2026. In the U.K., the MHRA stopped new starts from September 1 and provided a six-month managed withdrawal for existing patients, with revocation intended for March 1, 2027. Those decisions apply within their own jurisdictions and do not determine the U.S. outcome. [21,22]

The British arrangement illustrates why a regulatory decision needs an operational follow-through. Patients need clinicians to explain the changed evidence, reassess treatment, and arrange alternatives. Continued supply during a transition addresses continuity; it does not amount to compensation for earlier harm.

The harder question concerns patients who may already have suffered injury, incurred costs, or made treatment decisions on an unreliable account of the evidence.

The shareholder settlement does not perform that inquiry for them. Its release concerns defined stock-related claims. It cannot be read as a finding that patient losses have been investigated or remedied, or that public accountability is complete. [19]

Vasculitis International made a useful distinction in its May position statement: responsibility for compromised trial data should be pursued while the treatment’s benefit and risk receive their own assessment. Its concern then included preserving access for patients who believed they benefited. Following the European decision, the organization accepted the regulator’s benefit-risk assessment while continuing to call for accountability and patient support. [23,24]

Patient advocacy therefore cannot be reduced to a demand that a medicine remain available. It can also demand an explanation of what went wrong, practical help during a transition, and responsibility placed with the appropriate parties.

U.S. law firms have advertised investigations of claims involving serious liver injury and deaths. These pages demonstrate that claims are being solicited or evaluated. They do not establish that a lawsuit has been filed, that liability has been found, or that anyone has been paid. The sources reviewed for this note did not establish a completed patient compensation program for Tavneos. That is a limit of this review, rather than a claim that no individual proceeding or private resolution exists. [25]

An effective remedy would require more than a headline amount: an identifiable route for patients, relevant medical records, an assessment of causation and responsibility, and evidence that any award or settlement actually reached eligible people. Which losses are legally compensable would depend on the facts and jurisdiction.

For investors, unresolved patient harm can have implications for litigation, costs, physician confidence, and management credibility. The public record does not support assigning those implications a speculative dollar amount. Patient losses also deserve attention when their financial materiality cannot yet be established.

An investment perspective can acknowledge both points. Unquantified responsibility is still something to examine; ethical importance does not depend on a liability estimate.

What would change the investment case

The next useful information falls into distinct categories.

A U.S. regulatory decision would clarify the product’s legal position and any conditions on continued use. Better comparative evidence could clarify its additional clinical contribution and the patients for whom benefit outweighs harm. Acquisition disclosures or substantive findings could clarify the scope of historical diligence and responsibility. Actual court orders, filed patient claims, and payment records would clarify which legal uncertainties have been resolved.

The approved securities settlement is one milestone within that wider assessment. Amgen has scheduled its next quarterly results for November 3. That gives investors a near-term opportunity to examine how management’s sales outlook and disclosures account for the unresolved evidence and regulatory position. This is an opportunity for clarification, rather than a prediction that the FDA process will conclude by then. [26]

My view is that investors should examine evidence quality as part of the asset itself. A revenue forecast depends on durable permission to market, confidence among prescribers and patients, and a treatment contribution that can withstand scrutiny. If those foundations weaken together, a forecast based primarily on recent sales can become unreliable.

Tavneos also raises a question for how management is evaluated after an acquisition. Defending a potentially useful medicine can be reasonable. The quality of that defense depends on transparent evidence, a credible response to safety findings, and practical attention to patients who face a change in treatment or possible harm.

Approval, acquisition, and settlement each mark a decision. The evidence supporting the medicine, and the obligations to people who relied on it, continue beyond those decisions. A full investment account should follow them there.

---

Sources and evidence notes

[1] Amgen. Acquisition completion, October 20, 2022. Reports approximately $3.7 billion in aggregate merger consideration and identifies additional acquired assets. This note uses the completed-transaction figure rather than the earlier roughly $4 billion headline.

[2] Amgen. 2025 Form 10-K. Tavneos product sales were $423 million U.S. plus $36 million elsewhere in 2025; $256 million plus $27 million in 2024. Total 2025 revenue was $36,751 million. The approximately 1.25% comparison is the author’s calculation, not an estimate of contribution to earnings.

[3] Amgen. Second-quarter 2026 financial results. Reports $150 million in Tavneos sales, up 36% year over year, primarily from volume growth.

[4] FDA. CDER proposes to withdraw approval of Tavneos, April 27, 2026. The public notice describes a proposed withdrawal and interim availability. No final U.S. withdrawal order or formal FDA hearing date was identified in this review.

[5] FDA/CDER. Proposal to withdraw approval and notice of opportunity for a hearing, April 27, 2026, NDA 214487, docket FDA-2026-N-1321. Source for the agency’s findings, the initial analysis, selective readjudication, disclosure concerns, and earlier superiority requirement. Both historical P values cited in the text are two-sided.

[6] Rubin EJ. Retraction of the ADVOCATE article. NEJM, June 29, 2026. Specifies that the two academic authors did not know of the post-lock, post-unblinding readjudication. The retracted paper is historical context, not treated here as reliable current evidence of efficacy.

[7] ADVOCATE. Protocol CL010_168, amendment 4.0, January 18, 2019. Treatment assignments, background therapy, tapering, permitted glucocorticoids, and primary-endpoint definitions.

[8] FDA. Arthritis Advisory Committee transcript, May 6, 2021, especially printed pages 98–99, 145–146, and 275–281. Distinguishes the contemporaneous robustness concerns from the later-disclosed selective readjudication.

[9] Amgen and ChemoCentryx. July 23, 2026 submission supporting summary judgment or a hearing, with company statement. Company arguments and accounts of its knowledge are attributed, not presented as accepted regulatory findings.

[10] Jayne et al. Treatment of ANCA-Associated Vasculitis With Avacopan: Results of the ADVOCATE Trial After Readjudication of the Primary Endpoint. September 16, 2026, preprint. Amgen-sponsored, independently performed DCRI readjudication. The stated superiority P value of 0.0282 is one-sided and does not meet the prespecified 0.025 threshold; the text uses the confidence interval to avoid ambiguity.

[11] Ibiloye et al. Real-world systematic review and meta-analysis. June 17, 2026, Amgen-funded preprint. The 12-month remission estimate comes from a subset of the 71 studies; it is not a randomized estimate of additional benefit.

[12] Dang et al. Comparative effectiveness and safety study. July 22, 2026, preprint. Linked U.S. claims and electronic health records, weighted comparison, and limitations in direct disease-activity measurement.

[13] FDA. Serious liver-injury safety communication, March 31, 2026. Cases reviewed through October 9, 2024. The eight deaths include fatalities among the seven biopsy-confirmed VBDS reports; these are not separate totals to add.

[14] Kissei. Tavneos capsules Q&A and linked Japanese safety information. Reports 20 fatal hepatic cases in Japan through April 27, 2026. The reported estimate of 8,503 users concerns a specified annual period and is not used here as a denominator for cumulative fatal reports.

[15] Amgen. Statement on Japanese serious liver-injury reports, May 2026. The company says the 20 reports include cases for which causality could not be determined. The Japanese and earlier FDA datasets are not assumed to be independent.

[16] Amgen. Second-quarter 2026 Form 10-Q, risk factors concerning Tavneos. Describes uncertainty and potential consequences; a risk disclosure does not establish that each potential proceeding is underway.

[17] Court-authorized ChemoCentryx Securities Litigation FAQ. Procedural history, summary judgment and appeal, and increase in the proposed settlement from $35 million to $69 million.

[18] Court-authorized settlement website and document list. Confirms the October 29, 2026 final approval and related orders. As reviewed October 31, payment remains subject to completion of claims processing; no specific payment date or completed distribution was verified.

[19] Revised Stipulation and Agreement of Settlement, May 15, 2026. Paragraph 1(pp) ties the released plaintiff claims to acquisition of ChemoCentryx common stock; paragraph 38 addresses no admission. This is not a patient injury compensation arrangement.

[20] 21 U.S.C. §355(e), (h) and 21 CFR §12.24. Withdrawal grounds, notice and hearing opportunity, judicial review, and the conditions for granting a hearing. The statute also preserves emergency suspension authority under specified circumstances.

[21] EMA. Tavneos EPAR. Records European Commission revocation on August 4, 2026.

[22] MHRA. Review of Avacopan Vifor, September 1, 2026. New-start restriction, managed transition, guaranteed transitional supply, and intended March 1, 2027, revocation.

[23] Vasculitis International. Patients should not pay for failure in science, May 5, 2026. Cited for its position on accountability and patient access, not as independent proof of effectiveness.

[24] Vasculitis International. Response to European revocation, August 7, 2026. Cited for its acceptance of the regulatory assessment and continuing calls for accountability and transition support.

[25] Lamb law firm investigation notice and Cory Watson investigation page. These are legal-service and case-investigation materials, not court records proving filing, liability, settlement, or payment.

[26] Amgen. Announcement of third-quarter 2026 results webcast, October 28, 2026. Announces results for November 3, 2026, after the U.S. market close. The suggested questions for management are the author’s analysis.

[27] U.S. District Court, Northern District of California. Judgment Approving Class Action Settlement, October 29, 2026, Document 380. Final approval, dismissal of the securities claims, and the no-admissions provision.

[28] Same court. Order Awarding Attorneys’ Fees and Litigation Expenses, October 29, 2026, Document 378. Confirms the $69 million escrow funding. The court-authorized document list also supplies the allocation order.

---

This Editor’s Note presents analysis of public evidence. Opinions and valuation implications are the author’s inferences; regulatory findings, company positions, research results, and judicial actions are identified separately. It provides no personalized investment, medical, or legal advice and makes no recommendation to buy, sell, or hold a security. Procedural status is limited to the stated review date.

© 2026 BioStockEdge. All rights reserved. Reproduction or republication requires permission.