Study Finds High Number of European Cancer Drug Trials at Risk for Bias

Half of the cancer drug trials that led to approvals by the European Medicines Agency may have been associated with a high risk for bias.

Half of the cancer drug trials that led to approvals by the European Medicines Agency (EMA) over a 2-year period appeared to be at high risk for bias based on trial design, conduct, or anaylysis, according to an investigation published in BMJ

Most of the risk-creating factors were due to endpoints being based on progression-free survival (PFS) or response rates (RR), but not overall survival (OS).

“The three key findings of this study warrant further discussion,” the investigators wrote. 

In their assessment, the investigators focused on 32 cancer drugs which were approved by the EMA between 2014 and 2016 – including 5 drugs approed for multiple myeloma, 4 indicated for melanoma, and 4 for lung cancer.

Those approvals were based on 54 pivotal studies conducted by experts from the United Kingdom, Canada, and United States. 

Forty-one of the 54 studies (76%) involved randomized-controlled trials. But only 10 of them used OS as a primary endpoint, with most of the others using PFS or RRs instead as the main objective. 

An assessment of the trials focused on 3 core areas: trial design, risk of bias, and the adequacy, completeness, and consistency of reporting across sources. 

Nineteen of the 41 trials had “high risk of bias,” according to the metrics applied by the analysts. Ten of the studies were missing outcome data, and 7 were missing some measurements of the outcome. 

Among the findings:

-      Three of the drugs [cabozantinib (Cabometyx), trametinib (Mekinist), and pegaspargase (Oncaspar)] had high bias risk in 3 of 6 assessed categories. 

-      The EMA’s Committee for Medicinal Products for Human Use (CHMP) expressed concern about the lack of a cobimetinib (Cotellic)-only treatment arms in the trials for that drug. 

-      The trials for nivolumab (Opdivo) did not include some OS statistics – and once relied solely on overall response rate as an endpoint, according to concerns raised by the CHMP. 

-     Ixazomib (ninlaro), a drug for multiple myeloma, was initially granted a marketing authorization by the EMA, which was overturned. The single pivotal trial of the drug showed worsening of results between the initially submitted analysis and the updated analysis. The agent was granted conditional marketing authorization; however, the ultimate approval decision was opposed by 9 members of the CHMP, who signed a divergent opinion.

The investigators found that, when regulators did note deficiencies, they were not included in consequent publications.

“European regulators frequently raised questions about the appropriateness and applicability of the available evidence on new cancer drugs, which were not acknowledged in the scientific literature,” the investigators wrote. 

Overall, further oversight to guide trials toward asking the right questions, and then answering them in publications, is necessary, they added. 

“Regulatory action is needed to ensure drug makers are not only conducting randomized controlled trials – but also focusing on the data and endpoints that proves the treatments are meeting the needs of patients, doctors, and healthcare at large,” the investigators concluded. “Because cancer drugs are responsible for most of the recent increases in pharmaceutical spending across healthcare systems, the evidence base that supports their market entry warrants close scrutiny.”



Naci H, Davis C, Savovic J, et al. Design characteristics, risk of bias, and reporting of randomized controlled trials supporting approvals of cancer drugs by European Medicines Agency, 2014-16: cross sectional analysis. BMJ. 2019;366:l5221. doi:10.1136/bmj.l5221.