Half of FDA drug approvals based on one study

And the pharmaceutical companies get to choose which studies they are.

---

It’s an Introduction to Epidemiology-level caveat that you should never rely on just one study.

However well designed and powered, there’s always room for chance and bias.

So how is it possible that more than half the drugs newly approved by the FDA last year received those approvals based on a single study?

That’s one chilling finding from two new papers published in JAMA Network Open and Health Affairs Scholar, in which the same team reviewed new FDA drug approvals in 2017 and 2022 and how many trials formed the evidence base for those approvals. Using the clinicaltrials.gov website they also looked at the timing of the publication of studies in relation to FDA approvals.

(The Back Page doesn’t know why these two studies by the same researchers applying the same question and method are published in two journals when you’d think they’d be stronger combined, but such are the unfathomable ways of academic publishing.)

If you followed the aducanumab saga of the past few years, you’ll be familiar with the FDA’s accelerated approvals process, in which manufacturers can rely on surrogate endpoints in their studies rather than actual clinical outcomes: amyloid plaque removal rather than slowed cognitive decline, in the case of that dicey Alzheimer’s drug. That means you need to be very sure that the surrogate endpoint is a sound proxy for the outcome you want, which in the case of Alzheimer’s and amyloid is increasingly controversial. 

This relaxing of the rules was part of the 21^st Century Cures Act signed by Barack Obama in 2016, which also allowed treatments for “priority” conditions to be approved with fewer supporting studies and reduced the emphasis on randomised controlled trials.

What’s extraordinary is that when multiple studies are done, as they usually are, the pharmaceutical companies get to choose which ones they submit to the FDA. We get why you’d want a faster process for the sake of patients, that’s obvious – but how does allowing the companies to pick and choose their evidence benefit anyone but the companies?

“Current FDA policy requires pharmaceutical companies to submit 2 completed trials with their approval request. For high-priority conditions, such as some cancers, approvals can be based on a single trial. However, companies are allowed to choose which trials to submit and do not have to disclose additional studies that might have shown less or nonsignificant effectiveness,” the authors write in the study on 2017 approvals, which we’ll call Study 1.

They found that for the 46 drugs approved that year, a mean of 2.2 studies were used for FDA evaluation, when a mean of 25 studies were registered on clinicaltrials.gov.

Nineteen of the 46 (41%) were approved from a single study.

For 33 of them, at least one new set of clinical results was published after approval, even though many of the studies had been completed years earlier. “Although we do not know why these studies were not reported within the required 1-year interval following completion, we speculate that data were not posted because they did not clearly support the case for approval. If this speculation is correct, it would imply that clinical trials used for FDA approval may be a highly selected subset of completed research studies,” the authors write.

In Study 2, they found that of the 37 drugs approved in 2022, 24 (65%) were approved on a single study.

A total of 415 studies were conducted to evaluate those 37 drugs, but only 25% of them have become publicly available.

“New drug approvals in 2022 appeared to be based on fewer studies than before passage of the 21st Century Cures Act,” the authors write. “We believe consumers deserve access to the full range of evidence for the drugs they are considering, not just from the selected studies released to the public.”

In a reassuring – for non-US readers – postscript in Study 1 they say similar agencies such as our very own TGA are not simply following the FDA’s lead. Five drugs that won FDA approval between 2017 and 2020 were rejected in other countries and “health technology agencies in Australia, Canada, or the United Kingdom had declined to endorse 42 of the products because the benefits were uncertain or the costs were excessive”.

Aducanumab, which the FDA approved in June 2021 after a much-criticised process, is one example where the TGA declined to follow the US example.

There can’t be many things worse than languishing from a disease while a treatment grinds its way through a sclerotic bureaucratic machine; but being sold a hastily endorsed, clinically ineffective and hugely expensive drug in your final years is probably one of them.

Send story tips to penny@medicalrepublic.com.au and win our approval.