- Most pivotal studies for cancer drugs recently approved in Europe were randomized controlled trials, widely regarded as a gold standard in research, but new research has found about half of those trials were at high risk of bias.
- A group of Canadian and British researchers analyzed the rigor of clinical testing for new cancer therapies approved by the European Medicines Agency from 2014 to 2016, publishing their findings Wednesday in The BMJ.
- Out of the 54 supporting pivotal studies for 32 cancer drugs approved in that time period, 41 were randomized controlled trials (RCTs). Researchers determined 49% of those RCTs were at high risk of bias, including studies of some prominent therapies such as Bristol-Myers Squibb's Opdivo, AbbVie's Imbruvica and Pfizer's Ibrance.
Balancing patient access to new treatments with the need to conclusively prove drug benefits is a long-standing tension in cancer research. Recently, debate has centered on the rigor of accelerated approvals and the correlation between surrogate measures and patient benefit.
Drugs that have shown effectiveness in certain cancer types on surrogate tests before failing to improve overall survival include Avastin (bevacizumab), Inlyta (axitinib), Afinitor (everolimus) and Tecentriq (atezolizumab), the study's authors wrote in a related blog post for The BMJ.
But their study raises questions about the conduct of RCTs as well. The researchers highlighted three main takeaways from their research.
First, they found RCTs were less prevalent compared to a prior study finding RCTs made up 90% of pivotal studies between 2009 and 2013. In this study for 2014 to 2016, RCTs accounted for only 75% of pivotal trials as more studies featured single-arm designs.
Secondly, a risk-of-bias assessment demonstrated methodological weaknesses in the evidence underpinning many EMA cancer drug approvals. Researchers analyzed several aspects of studies, including the randomization process, missing outcomes data and outcome measurements for bias.
The study authors also found difference in how study limitations were acknowledged between regulatory documents from the EMA and scientific literature. Journal publications were less likely to address these concerns, the researchers found.
In a concurrently-published editorial, two Australian academics said this BMJ study "confirms and extends the existing body of research that raises serious concerns about low standards of evidence supporting new cancer drug approvals."
Positive trial results are not enough, they concluded, stating that assessing the quality of that evidence "is also needed to ensure that these trials accurately estimate treatment effects."