Tavneos Approval Shows FDA Trusts Companies Before Catching Fraud

The FDA is now trying to unsay something it said in 2017. Tavneos, a drug for severe autoimmune inflammation, won approval based on clinical trial data that Ceptera Therapeutics appears to have fabricated.

The regulators want it off the market, but thousands of patients are already taking it. Some are better because they are. Pull the drug and you harm people who believed their government vetted the evidence. Leave it approved and you signal that manufacturing data is a solvable problem if you're patient enough.

In 2004, Merck withdrew Vioxx, its painkiller, after cardiovascular deaths piled up. The remarkable part was not that the risks appeared after approval — that happens — but that Merck had buried the cardiovascular data inside the approval package itself. The FDA's own reviewers had walked past it. Twenty years of reform has followed. Post-market surveillance expanded, adverse event reporting tightened, and the system learned to watch drugs once they were already in bodies.

But the system never learned to trust its own approval decision less. Vioxx teaches a lesson about what happens when a company controls the data pipeline and regulators assume good faith. And Tavneos follows the exact same shape. A manufacturer submits results. Reviewers vet them against pre-existing knowledge and consistency checks, but they depend on the raw data being real.

If it isn't, and if it's constructed with enough technical plausibility, detection becomes retrospective. You notice the problem when the drug is already approved, already prescribed, already working for some people and not others in ways that don't match what the trial suggested they should. The FDA's reforms after Vioxx tightened the rules about what happens next. They did almost nothing about preventing this in the first place, about validating source data before approval, about auditing manufacturing records during the review phase rather than after. The structural gap that let Merck happen, the assumption that a company's trial results are genuine unless proven otherwise, is the same gap that let Tavneos happen. What changes this time is whether regulators will see their own mistake as a signal about their process. Or simply upgrade their post-approval apologies.