The 9 Patients Who Could Take Tavneos Off the Market

The FDA has proposed an administrative hearing to pull Tavneos from the market — and at the center of its case is a claim that nine patients had their trial outcomes re-adjudicated after the study blind was broken.

Those nine patients, out of 331 in the ADVOCATE trial, are the statistical linchpin. According to the agency's April 27 proposal to withdraw approval, the original endpoint analysis did not support approving the drug. A re-analysis, conducted after unblinded personnel handled the data, did. That re-adjudication was never disclosed to the FDA during its review. Tavneos (avacopan) was approved in October 2021 for ANCA-associated vasculitis, a rare autoimmune disease that inflames small blood vessels and can be fatal if untreated.

Amgen, which acquired ChemoCentryx and now sells Tavneos, is refusing to withdraw the drug. The company's position, stated in its January refusal, is that the re-adjudication was scientifically appropriate and the benefit-risk profile supports continued marketing. It has not explained why unblinded personnel were handling endpoint data or why that process was not disclosed to regulators.

The administrative hearing will determine what happens next. But the structural problem the FDA has put on the record is hard to ignore: a drug approved based on a re-analysis that was never shared with the agency during the review process.

The acquisition adds another layer. Amgen bought ChemoCentryx in October 2022 for $3.7 billion, roughly one year after approval. If the original ADVOCATE analysis did not support approval, that fact was either unknown to Amgen at the time of the deal, or it was known and deemed acceptable. Neither possibility is reassuring.

The nine patients at the center of the FDA's case are not identified in any public document. Their individual re-adjudicated results are not published. What is published are the top-line numbers that made Tavneos look like an advance: in the ADVOCATE trial, week 26 remission was 72.3% for avacopan versus 70.1% for prednisone, meeting criteria for noninferiority, and week 52 sustained remission was 65.7% versus 54.9%. That is the benefit the FDA is asking to sacrifice in order to enforce a principle: a trial's integrity matters as much as its outcome.

Separately, the FDA's March 31 Drug Safety Communication flagged a liver injury signal the agency is also tracking. As of that date, the agency had received reports of 76 cases of drug-induced liver injury, including seven biopsy-confirmed cases of vanishing bile duct syndrome — a rare and potentially irreversible condition in which the bile ducts inside the liver progressively disappear. Three patients in the reported cases died. Most vanishing bile duct syndrome cases occurred in Japan and in patients aged 65 and older, according to Amgen's prescriber letter. The median time from starting the drug to liver injury onset was 46 days, with a range of 33 to 59 days. Amgen has not linked the liver cases to the trial integrity allegations, and the FDA has not stated that the two issues are related.