It was a surprise to most people when the FDA issued a Complete Response Letter to BMS and Pfizer regarding apixiban for atrial fibrillation today. The ARISTOTLE Trial demonstrated both reduced stroke and systemic embolus rates and reduced bleeding compared with warfarin. The INR controls were excellent in the warfarin group. The trial has been published in the NEJM.
The contents of a CRL are private between the FDA and the corporate sponsors and I have not seen the document. The press release says that the FDA needs more information about data management and verification.
I have spoken with Chris Granger and Lars Wallentin, the Co-PI’s. While DCRI did not do the data management or monitoring for this trial, we have a copy of the database and our statisticians have performed the independent analyses. Chris and Lars have assured me that all the essential elements of a quality clinical trial are in place and robust. These elements are intended to assure that trials are free of systematic bias and answer clinical relevant questions. They will be stressed in “quality by design” measures being discussed in many policy circles, specifically in clinical trials that are large and intended to measure clinical outcomes in a manner that enables the determination of the balance of risk and benefit.
These measures include: were the right patients included (representative of intended use)?; did the randomization work to assign the right treatments?; did blinding work?; was adherence acceptable?; were the outcomes measured correctly in an unbiased manner?; were significant adverse events measured as planned?; was the analysis plan correct and was it followed?
Perhaps most importantly, the trial had the power to measure clinically meaningful differences in a relevant population, and it used blinding and intention to treat. So often these days we learn a lot about a drug, but don’t have the most important information: compared with current standards, what is the balance of risk and benefit for clinical outcomes?
Hopefully, these issues will be resolved quickly and the story can be told. The FDA,other regulatory authorities, industry and academia are working hard to come up with quality by design methods that support streamlining of trials to get reliable answers with appropriate sample sizes. Current industry standard methods are placing outcome trials out of the price range that makes development of new therapies (or comparative evaluation of existing therapies) feasible. In the past 2 weeks major meetings have occurred in North American and Europe to reach consensus on these issues. Our CTTI consortium is working on vital components to generate evidence to support policy (see www.ctti-clinicaltrials.org).
There is a lot riding on getting this right and clearly explaining what happened and what should be done in the future.
more...