When Trial Data Integrity Unwinds an Approval: The Tavneos ADVOCATE Case

A clinical trial that supported a drug approval four years ago has now helped unwind it. In 2026, Health Canada told doctors to stop starting new patients on Tavneos (avacopan) while it reviews serious concerns about the integrity of the data from the pivotal ADVOCATE trial. In Europe, the regulator went further and moved to revoke the approval outright. The medicine did not change. The data behind it did not change either. What changed is what regulators learned about how that data was handled.

For sponsors, contract research organisations (CROs) and quality teams, the Tavneos case is a study in why data integrity is not a paperwork problem. It is the foundation an approval rests on. This article explains what happened, why Good Clinical Practice and the new ICH E6(R3) guideline matter, and what a sponsor should do now to make sure its own trial data would survive the same scrutiny.

Executive summary

The authorization of Tavneos in Canada in April 2022 rested primarily on the Phase 3 ADVOCATE trial. In 2026, new information raised significant concerns about how the ADVOCATE data was handled before the drug was authorized. The New England Journal of Medicine retracted the ADVOCATE publication on 29 June 2026. Health Canada issued a health professional risk communication advising against starting new patients. The European Medicines Agency reviewed the trial under a formal procedure and concluded that it was conducted in breach of Good Clinical Practice, and the European Union moved to revoke the marketing authorization. The lesson for industry is direct: data integrity failures can reach back years to invalidate an approved product.

What happened with Tavneos and the ADVOCATE trial

Tavneos (avacopan) is authorized in Canada as an add on treatment for adults with severe active ANCA associated vasculitis, a group of rare autoimmune diseases. Its 2022 approval rested mainly on the Phase 3 ADVOCATE trial. In 2026, regulators found significant concerns about the integrity of that trial data. Health Canada is reviewing the approval and has told prescribers not to start new patients, while the European Union has moved to revoke its approval.

Avacopan is a targeted therapy that blocks the complement C5a receptor. It is indicated for the adjunctive treatment of adult patients with severe active anti-neutrophil cytoplasmic autoantibody (ANCA) associated vasculitis, specifically granulomatosis with polyangiitis and microscopic polyangiitis, in combination with standard background therapy that includes glucocorticoids. It is not a monotherapy. The pivotal ADVOCATE trial was a Phase 3 study sponsored by ChemoCentryx that evaluated avacopan in these patients.

On 7 July 2026, Health Canada published a health professional risk communication stating that it was reviewing significant concerns about the data integrity of ADVOCATE, the study that served as the primary efficacy evidence for the authorization. The concerns relate to how the trial data was handled before Tavneos was authorized, which creates uncertainty about the reliability of the efficacy evidence. Health Canada advised healthcare professionals not to initiate treatment in new patients and to review current patients. The publication of the ADVOCATE results was retracted by the New England Journal of Medicine on 29 June 2026.

How a data handling problem becomes an approval problem

A drug approval is a regulator's judgement that a product's benefits outweigh its risks, based on the evidence submitted. If the central evidence turns out to be unreliable, the basis for the approval collapses. That is why a data integrity finding, even years after approval, can lead a regulator to restrict, suspend or revoke a marketing authorization.

The European review made the mechanism explicit. After examining how the study data were handled, the Committee for Medicinal Products for Human Use concluded that the ADVOCATE study was conducted in breach of Good Clinical Practice, and that the data provided at the time of the marketing authorization assessment were incorrect and misleading and could no longer be relied upon. According to the European Medicines Agency, the reliability of the primary result was undermined after unblinded personnel reviewed results, and a re-adjudication of a small number of patients shifted the week 52 analysis from not statistically significant to statistically significant. When the number that demonstrates efficacy depends on choices made by people who could see the treatment assignments, the result can no longer be trusted.

This is the heart of data integrity. It is not about tidy binders. It is about whether an outside reviewer can trace every reported result back to a source, confirm that it was recorded at the time, and confirm that no one changed it in a way that favoured the outcome. When that chain breaks, the conclusion breaks with it.

What data integrity actually requires: ALCOA and beyond

Data integrity means data that can be trusted from creation to reporting. Regulators describe it with the ALCOA principles: attributable, legible, contemporaneous, original and accurate. The expanded ALCOA plus set adds complete, consistent, enduring and available. Trial data that fails any of these tests is both a scientific and a compliance risk.

Principle	What it means	How ADVOCATE type failures show up
Attributable	You can tell who recorded or changed data and when	Changes influenced by unblinded staff with unclear accountability
Legible and enduring	Data is readable and cannot be silently overwritten	A missing or unclear audit trail of adjudication changes
Contemporaneous	Data is recorded at the time of the activity	Results reviewed and reworked after the fact
Original	The first record, or a certified true copy, is kept	Reliance on reworked figures rather than source data
Accurate and complete	Data is correct and nothing is omitted	An analysis that flips outcome after selective re-adjudication

The ADVOCATE concerns map onto these principles. When results are reviewed by people who should have stayed blinded, when adjudication decisions change a primary endpoint, and when a regulator cannot reconstruct how the final numbers were produced, the data fails the tests of being contemporaneous, original and accurate. The scientific consequence is a result that cannot be trusted. The regulatory consequence is an approval that cannot stand.

ICH E6(R3): the new Good Clinical Practice benchmark

ICH E6(R3) is the modernised Good Clinical Practice guideline. It emphasises quality by design, risk-based approaches, clear roles for sponsors and service providers, and strong data governance across the trial data lifecycle. Health Canada adopted it, and it has applied to Canadian clinical trials since 1 April 2026. It raises the bar for exactly the controls that failed in ADVOCATE.

Good Clinical Practice is the international standard for designing, conducting, recording and reporting clinical trials. In Canada, clinical trials of drugs are governed by Part C, Division 5 of the Food and Drug Regulations, supported by Health Canada guidance document GUI-0100, and sponsors must also meet transparency duties through Canada's clinical trials portal. ICH E6(R3), the third revision of the GCP guideline, shifts the focus from documentation for its own sake to building quality into the trial from the design stage and managing the factors that are critical to reliable results and participant safety.

Three E6(R3) themes speak directly to the Tavneos case. First, data governance: sponsors are expected to maintain the integrity of trial data throughout its lifecycle, including clear control over who can access, change and adjudicate data. Second, sponsor oversight: a sponsor remains responsible for the trial even when it delegates work to CROs, laboratories and adjudication committees. Third, blinding and bias control: keeping the right people blinded and documenting every change to the analysis is a core protection against exactly the failure the European regulator described. Our guide to audit trail review explains how inspectors test these controls in practice.

What sponsors and CROs should do now

Treat the Tavneos case as a prompt to test your own trial data against the question a regulator will ask: can you show how every reported result was produced, and prove that no one changed it improperly? Review your blinding, your audit trails, your adjudication procedures and your oversight of vendors. Fix the gaps before an inspection or a data integrity review finds them.

Start with sponsor oversight. Confirm that your quality agreements with CROs, central laboratories and adjudication committees define who is responsible for data integrity, who may access unblinded data, and how changes are documented and approved. A sponsor that cannot describe how its vendors protect blinding and audit trails is carrying a hidden risk. Our articles on sponsor and supplier oversight and on supplier qualification set out how to structure that accountability.

Then examine your data governance and audit trails. Every clinical data system should record who entered or changed a value, when, and why, and that audit trail should be reviewed, not just switched on. Adjudication of endpoints should follow a pre-specified charter, with blinding maintained and every decision traceable. If a change to the analysis moved a result across the line of statistical significance, you must be able to justify it with contemporaneous documentation. When a root cause investigation is needed, it must reach the true cause rather than stopping at a convenient one, a discipline we cover in our guidance on CAPA and root cause.

Finally, prepare for inspection. Health Canada, the FDA and the EMA all inspect for GCP and data integrity, and the same discipline that protects a clinical trial protects a manufacturing record. The investigation standard that FDA applies to a failed batch mirrors the rigour a GCP inspector expects when a trial result is questioned. Building that inspection readiness now, across both clinical and manufacturing quality systems, is the practical takeaway from Tavneos.

Why this reaches beyond one drug

The Tavneos case is part of a broader regulatory trend. Data integrity has become the common thread in enforcement across clinical trials and manufacturing, and regulators are increasingly willing to act on it retrospectively. Any sponsor whose approvals rest on a single pivotal trial should treat that trial's data integrity as a strategic risk, not a closed file.

For companies operating across Canada, the United States and Europe, the alignment of Health Canada, the FDA and the EMA on data integrity means a weakness found by one regulator can quickly become a problem in every market. It also means the controls are portable: an audit trail, a blinding procedure and a vendor oversight framework that satisfy one authority will generally satisfy the others. The investment is made once, and the protection is global.

A closer look at what went wrong in ADVOCATE

The specific failure in ADVOCATE is worth understanding, because it is subtle. The problem was not a fabricated dataset. It was a loss of control over who could see and influence the data, at the exact moment the primary result was being decided.

According to the European regulator, personnel who should have remained blinded reviewed unblinded results, and a re-adjudication of a small number of patients shifted the week 52 primary analysis from not statistically significant to statistically significant. In other words, a result that did not clearly demonstrate efficacy became one that did, after decisions made by people who could see the treatment assignments. For a quality professional, that is the difficult scenario: not obvious fraud, but a governance gap that lets bias enter the data unseen. It is also largely preventable. Strong blinding controls, a pre-specified adjudication charter, a reviewed audit trail and independent sponsor oversight would each have made the change visible and open to challenge. The lesson is that data integrity is protected by process design, not by good intentions.

Clinical data integrity checklist

Use this checklist to test whether your key trials would survive a data integrity review. Keep the completed version in your quality records.

Confirm every pivotal trial has a complete, reviewable audit trail showing who changed each data point and why.
Verify blinding procedures and document every person with access to unblinded data.
Check that endpoint adjudication follows a pre-specified charter with traceable decisions.
Confirm quality agreements with CROs and laboratories assign clear responsibility for data integrity.
Ensure any change that affected a primary analysis is justified with contemporaneous records.
Map your trial data lifecycle and controls against ICH E6(R3) data governance expectations.
Run a mock GCP and data integrity inspection on at least one key study.
Confirm sponsor oversight of vendors is documented, not assumed.

Common mistakes

Treating data integrity as an IT setting rather than a governed process across the data lifecycle.
Switching on an audit trail but never reviewing it.
Delegating a trial to a CRO and assuming oversight transfers with the work.
Allowing unblinded staff to influence analysis or adjudication without documented controls.
Changing an analysis to reach significance without a pre-specified, justified rationale.
Assuming a completed and published trial can never be reopened by a regulator.

Frequently asked questions

Was Tavneos recalled in Canada?

No. Health Canada issued a health professional risk communication and is reviewing the approval. It advised prescribers not to start new patients and to review current patients, but it did not recall the product. Patients are told not to stop treatment without first speaking to their healthcare professional.

What is the ADVOCATE trial?

ADVOCATE was the pivotal Phase 3 trial, sponsored by ChemoCentryx, that supported the approval of avacopan for severe active ANCA associated vasculitis. Its results were the primary efficacy evidence for the authorization, and the New England Journal of Medicine retracted the publication on 29 June 2026.

What does data integrity mean in a clinical trial?

Data integrity means that trial data can be trusted from the moment it is created to the moment it is reported. It is often described by the ALCOA principles: attributable, legible, contemporaneous, original and accurate, extended by complete, consistent, enduring and available. If data fails these tests, the conclusions drawn from it cannot be relied upon.

What is ICH E6(R3) and when does it apply in Canada?

ICH E6(R3) is the current revision of the international Good Clinical Practice guideline. It emphasises quality by design, risk-based approaches, sponsor oversight and data governance. Health Canada adopted it, and it has applied to Canadian clinical trials since 1 April 2026, alongside Part C, Division 5 of the Food and Drug Regulations.

Can a regulator revoke an approval over trial conduct years later?

Yes. If a regulator concludes that the pivotal evidence is unreliable, it can restrict, suspend or revoke a marketing authorization. In the Tavneos case, the European Union moved to revoke the approval after concluding that the ADVOCATE study was conducted in breach of Good Clinical Practice.

How can we prove our own trial data would survive this scrutiny?

Show that every reported result can be traced to source data, that audit trails record and explain every change, that blinding was maintained, and that vendor oversight and adjudication followed documented procedures. A mock GCP and data integrity inspection is the fastest way to find the gaps before a regulator does.

How MFLRC can help

MF License and Regulatory Consultants helps sponsors, CROs and manufacturers build clinical and quality systems that hold up under a data integrity review. We carry out GCP and data integrity gap assessments against ICH E6(R3), review sponsor oversight and vendor qualification frameworks, and strengthen the audit trail review, blinding and adjudication controls that regulators test. Our team also runs mock inspections, builds CAPA and root cause capability, and prepares the standard operating procedures and validation evidence that support a defensible trial data lifecycle.

Whether your risk sits in a pivotal trial, a vendor relationship or a manufacturing record, we help you find and close the gaps before a regulator does. Explore our audit services and quality assurance services, or speak with a senior consultant about your clinical and manufacturing quality systems.

Book a consultation

Conclusion

The Tavneos case shows that an approval is only as strong as the integrity of the data behind it. A trial completed and published years ago was reopened, its central result questioned, and its approval put at risk once regulators understood how the data had been handled. That should focus every sponsor and CRO on a single question: could you show, today, how each of your key results was produced, and prove that no one changed it improperly? The companies that can answer yes have little to fear from the next data integrity review. The ones that cannot should start closing the gap now.